WEBVTT 00:00:00.000 --> 00:00:03.446 align:middle line:90% [SQUEAKING] [RUSTLING] 00:00:03.446 --> 00:00:04.919 align:middle line:90% [CLICKING] 00:00:04.919 --> 00:00:09.375 align:middle line:90% 00:00:09.375 --> 00:00:11.000 align:middle line:84% HAZHIR RAHMANDAD: Today, we are focused 00:00:11.000 --> 00:00:15.400 align:middle line:84% on growth of a professional field and modeling that. 00:00:15.400 --> 00:00:20.220 align:middle line:84% So I see fewer models submitted so far. 00:00:20.220 --> 00:00:26.120 align:middle line:84% So we'll see how many will show up. 00:00:26.120 --> 00:00:28.860 align:middle line:90% So now we have 2, 3, 4, 5, 6. 00:00:28.860 --> 00:00:31.640 align:middle line:90% We have seven right now. 00:00:31.640 --> 00:00:34.700 align:middle line:90% So we could go through these. 00:00:34.700 --> 00:00:37.280 align:middle line:84% But before we get started, maybe just 00:00:37.280 --> 00:00:40.520 align:middle line:84% a little bit of, what were the common challenges you were 00:00:40.520 --> 00:00:48.200 align:middle line:84% facing, things that you could have gotten help for, 00:00:48.200 --> 00:00:51.100 align:middle line:84% or when you were thinking about it, were challenging. 00:00:51.100 --> 00:00:54.260 align:middle line:84% Anything that comes to your mind, feel free to share. 00:00:54.260 --> 00:00:58.660 align:middle line:84% So anybody wants to start us on that? 00:00:58.660 --> 00:01:12.250 align:middle line:90% 00:01:12.250 --> 00:01:14.730 align:middle line:84% AUDIENCE: I have done the conceptualization 00:01:14.730 --> 00:01:21.970 align:middle line:84% with handling quality, the needs of quality, 00:01:21.970 --> 00:01:27.310 align:middle line:90% and how the job's being done. 00:01:27.310 --> 00:01:29.870 align:middle line:84% What would the quality of job be in that? 00:01:29.870 --> 00:01:32.650 align:middle line:90% You have the mix of people? 00:01:32.650 --> 00:01:35.930 align:middle line:84% And in that regard, I also had the problem of allocation 00:01:35.930 --> 00:01:39.290 align:middle line:90% of the tasks, yes. 00:01:39.290 --> 00:01:41.050 align:middle line:90% I was thinking that-- 00:01:41.050 --> 00:01:44.010 align:middle line:84% because, for example, we have 10 experts, 00:01:44.010 --> 00:01:47.132 align:middle line:84% 10 novice, but we have five tasks. 00:01:47.132 --> 00:01:48.090 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 00:01:48.090 --> 00:01:49.998 align:middle line:84% AUDIENCE: So how do you overcome? 00:01:49.998 --> 00:01:56.092 align:middle line:90% 00:01:56.092 --> 00:01:58.300 align:middle line:84% HAZHIR RAHMANDAD: Those are very important questions, 00:01:58.300 --> 00:02:02.260 align:middle line:84% and they actually may even point to important policy lever points 00:02:02.260 --> 00:02:07.180 align:middle line:84% because the allocation is not determined fully. 00:02:07.180 --> 00:02:10.660 align:middle line:84% It depends on the underlying matching technology that 00:02:10.660 --> 00:02:12.920 align:middle line:90% is being used in the space. 00:02:12.920 --> 00:02:15.140 align:middle line:84% I'm calling it technology, not that, necessarily, it 00:02:15.140 --> 00:02:17.500 align:middle line:84% is a technology, but whatever is the process, yeah. 00:02:17.500 --> 00:02:20.220 align:middle line:90% 00:02:20.220 --> 00:02:22.232 align:middle line:90% Other observations? 00:02:22.232 --> 00:02:23.940 align:middle line:84% AUDIENCE: Just on that, I couldn't really 00:02:23.940 --> 00:02:25.460 align:middle line:90% figure out if it was like-- 00:02:25.460 --> 00:02:26.660 align:middle line:90% if there was a different-- 00:02:26.660 --> 00:02:28.660 align:middle line:84% I ended up not doing this, but I couldn't really 00:02:28.660 --> 00:02:30.620 align:middle line:84% figure out if there was a different demand 00:02:30.620 --> 00:02:34.380 align:middle line:84% for professionals versus rookies or if there 00:02:34.380 --> 00:02:40.060 align:middle line:84% was a way to sort of make that part of it. 00:02:40.060 --> 00:02:42.200 align:middle line:84% HAZHIR RAHMANDAD: So is demand homogeneous? 00:02:42.200 --> 00:02:46.700 align:middle line:90% 00:02:46.700 --> 00:02:48.920 align:middle line:84% Yeah, so that's an interesting observation. 00:02:48.920 --> 00:02:51.820 align:middle line:84% So you could actually think about demand 00:02:51.820 --> 00:02:59.710 align:middle line:84% to have different dimensions of size, of complexity, of margin. 00:02:59.710 --> 00:03:02.290 align:middle line:84% And if the quality of work has been going down, 00:03:02.290 --> 00:03:07.130 align:middle line:84% for example, your size of not only the number of projects, 00:03:07.130 --> 00:03:10.310 align:middle line:84% but size per project or the margin per project 00:03:10.310 --> 00:03:13.950 align:middle line:84% may also shrink, so that's an interesting idea. 00:03:13.950 --> 00:03:15.530 align:middle line:84% It wasn't in the problem definition, 00:03:15.530 --> 00:03:17.190 align:middle line:84% but that's a good call, Jake, that it 00:03:17.190 --> 00:03:21.750 align:middle line:84% could be a feature you could incorporate. 00:03:21.750 --> 00:03:25.310 align:middle line:84% AUDIENCE: There are two questions that I never really 00:03:25.310 --> 00:03:25.890 align:middle line:90% resolved. 00:03:25.890 --> 00:03:33.550 align:middle line:84% One is how much to explicitly account for alternatives. 00:03:33.550 --> 00:03:37.210 align:middle line:84% So someone chooses not to go with the new field, 00:03:37.210 --> 00:03:40.910 align:middle line:90% they're choosing an alternative. 00:03:40.910 --> 00:03:46.030 align:middle line:84% And that has implications for the recognition 00:03:46.030 --> 00:03:47.330 align:middle line:90% accrued to the alternative. 00:03:47.330 --> 00:03:50.030 align:middle line:90% 00:03:50.030 --> 00:03:51.990 align:middle line:84% So accounting for that, I chose not 00:03:51.990 --> 00:03:54.670 align:middle line:84% to go with thinking about the quality of the alternatives. 00:03:54.670 --> 00:03:55.990 align:middle line:90% It became too much. 00:03:55.990 --> 00:03:58.710 align:middle line:84% I'm not even sure I needed recognition. 00:03:58.710 --> 00:04:03.510 align:middle line:84% So that piece, and there was one other. 00:04:03.510 --> 00:04:06.230 align:middle line:90% 00:04:06.230 --> 00:04:09.590 align:middle line:84% Oh, thinking about the elements of-- 00:04:09.590 --> 00:04:11.950 align:middle line:84% the determinants of service quality, 00:04:11.950 --> 00:04:19.990 align:middle line:84% and in particular, how to respond 00:04:19.990 --> 00:04:23.910 align:middle line:84% to excess demand and the implications 00:04:23.910 --> 00:04:25.790 align:middle line:90% of that on learning. 00:04:25.790 --> 00:04:29.310 align:middle line:84% If it's an overburdened system, rookies probably 00:04:29.310 --> 00:04:31.370 align:middle line:84% aren't learning as much as they could. 00:04:31.370 --> 00:04:34.010 align:middle line:84% I chose not to model all of that, 00:04:34.010 --> 00:04:35.970 align:middle line:84% but it was in the back of my head. 00:04:35.970 --> 00:04:39.670 align:middle line:90% 00:04:39.670 --> 00:04:42.830 align:middle line:84% AUDIENCE: I was wondering a little bit about the utility. 00:04:42.830 --> 00:04:47.990 align:middle line:84% And if it was financial benefits, this would-- 00:04:47.990 --> 00:04:50.110 align:middle line:84% what's behind the utility for the service? 00:04:50.110 --> 00:04:55.200 align:middle line:84% Because that is such a big driver for the whole thing. 00:04:55.200 --> 00:04:56.920 align:middle line:84% It could be such a strong reinforcing. 00:04:56.920 --> 00:05:00.280 align:middle line:84% If there's financial benefits for somebody 00:05:00.280 --> 00:05:05.880 align:middle line:84% to use this service, it's going to increase the demand so much. 00:05:05.880 --> 00:05:09.640 align:middle line:90% So I was thinking how to-- 00:05:09.640 --> 00:05:11.360 align:middle line:84% and then it became so complicated 00:05:11.360 --> 00:05:14.200 align:middle line:84% because it can be-- because it's an open area. 00:05:14.200 --> 00:05:18.560 align:middle line:84% So it's like, OK, it's hard not to model a certain-- 00:05:18.560 --> 00:05:23.040 align:middle line:84% specific field when you think about such things, as what need 00:05:23.040 --> 00:05:25.840 align:middle line:84% is this service fulfilling for people? 00:05:25.840 --> 00:05:28.363 align:middle line:84% And how much do they benefit from that? 00:05:28.363 --> 00:05:29.280 align:middle line:90% HAZHIR RAHMANDAD: Yep. 00:05:29.280 --> 00:05:33.320 align:middle line:90% 00:05:33.320 --> 00:05:37.080 align:middle line:84% So the quality or utility impact on future demand 00:05:37.080 --> 00:05:41.080 align:middle line:84% is a pretty big part of this model. 00:05:41.080 --> 00:05:42.680 align:middle line:90% Very good. 00:05:42.680 --> 00:05:46.080 align:middle line:84% Anybody else wants to share what they 00:05:46.080 --> 00:05:51.030 align:middle line:84% might have found challenging or interesting, formulation issues? 00:05:51.030 --> 00:05:53.770 align:middle line:90% 00:05:53.770 --> 00:05:59.170 align:middle line:84% AUDIENCE: I want to have novices going through training, taking 00:05:59.170 --> 00:06:02.130 align:middle line:90% some labor out of experts too. 00:06:02.130 --> 00:06:05.330 align:middle line:84% But then I tried to do that, but then the allocation 00:06:05.330 --> 00:06:07.270 align:middle line:90% became too much. 00:06:07.270 --> 00:06:09.570 align:middle line:84% I was barely handling the allocation 00:06:09.570 --> 00:06:13.450 align:middle line:84% of time of experts and novices as it was. 00:06:13.450 --> 00:06:18.450 align:middle line:84% And right now, my trainees can decide to work, or train, 00:06:18.450 --> 00:06:20.190 align:middle line:90% or will not have anything to do. 00:06:20.190 --> 00:06:24.810 align:middle line:84% But experts are not involved in training novices. 00:06:24.810 --> 00:06:28.530 align:middle line:84% And then that seems logical, but I couldn't do it. 00:06:28.530 --> 00:06:32.570 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so kind of on-the-job training 00:06:32.570 --> 00:06:35.070 align:middle line:84% versus explicit expert allocation. 00:06:35.070 --> 00:06:42.890 align:middle line:90% 00:06:42.890 --> 00:06:48.520 align:middle line:84% Yeah, and this is most important and non-trivial, as you say. 00:06:48.520 --> 00:06:51.700 align:middle line:84% Allocation problems are generally hard, 00:06:51.700 --> 00:06:55.000 align:middle line:84% so having good formulations for those. 00:06:55.000 --> 00:06:57.700 align:middle line:90% Oh, we have two Alban's today. 00:06:57.700 --> 00:06:59.620 align:middle line:90% [CHUCKLES] 00:06:59.620 --> 00:07:02.280 align:middle line:90% 00:07:02.280 --> 00:07:05.380 align:middle line:90% Any other thoughts? 00:07:05.380 --> 00:07:07.860 align:middle line:84% AUDIENCE: Another thing that I didn't end up 00:07:07.860 --> 00:07:11.900 align:middle line:84% tackling-- cause I thought it was trickier-- was allocation 00:07:11.900 --> 00:07:18.940 align:middle line:84% of when both rookies versus pros try to leave the market 00:07:18.940 --> 00:07:21.180 align:middle line:84% because there isn't enough demand, 00:07:21.180 --> 00:07:22.880 align:middle line:90% generally, to make it appealing. 00:07:22.880 --> 00:07:31.580 align:middle line:90% 00:07:31.580 --> 00:07:35.620 align:middle line:84% HAZHIR RAHMANDAD: Yep, that's also pretty important. 00:07:35.620 --> 00:07:37.780 align:middle line:84% And when you leave the market, it's 00:07:37.780 --> 00:07:41.740 align:middle line:84% kind of-- if you are a rookie, you may go completely 00:07:41.740 --> 00:07:43.000 align:middle line:90% and you may never come back. 00:07:43.000 --> 00:07:44.420 align:middle line:84% When you're a pro, the chances are 00:07:44.420 --> 00:07:47.032 align:middle line:84% you go into a different mode of operation. 00:07:47.032 --> 00:07:48.490 align:middle line:84% But if there is demand coming back, 00:07:48.490 --> 00:07:51.010 align:middle line:84% you may come back to the field as well. 00:07:51.010 --> 00:07:53.310 align:middle line:84% So that's actually an interesting additional wrinkle 00:07:53.310 --> 00:07:54.870 align:middle line:90% you could consider. 00:07:54.870 --> 00:07:57.272 align:middle line:84% I think these are pretty good thoughts. 00:07:57.272 --> 00:07:58.730 align:middle line:84% All of them are important problems. 00:07:58.730 --> 00:08:00.990 align:middle line:84% So let's now dive in and see how you have been 00:08:00.990 --> 00:08:03.050 align:middle line:90% handling those in your models. 00:08:03.050 --> 00:08:09.030 align:middle line:84% And what are some of the solutions you 00:08:09.030 --> 00:08:10.650 align:middle line:90% have for dealing with these? 00:08:10.650 --> 00:08:18.150 align:middle line:84% And so let's open all of them for-- 00:08:18.150 --> 00:08:20.730 align:middle line:84% so I think we have the models open. 00:08:20.730 --> 00:08:24.858 align:middle line:84% So let's do a quick walkthrough in terms of the structures 00:08:24.858 --> 00:08:25.650 align:middle line:90% that you have used. 00:08:25.650 --> 00:08:27.470 align:middle line:84% And then we can do the deep dives. 00:08:27.470 --> 00:08:30.290 align:middle line:90% So this one is Jake's model. 00:08:30.290 --> 00:08:39.950 align:middle line:84% And we have here the rookies, veterans, and their capacity. 00:08:39.950 --> 00:08:44.240 align:middle line:84% So, Jake, can you tell me what you have in mind by capacity 00:08:44.240 --> 00:08:44.780 align:middle line:90% here? 00:08:44.780 --> 00:08:46.440 align:middle line:90% Is it about their skills? 00:08:46.440 --> 00:08:49.340 align:middle line:90% Or is it about their-- 00:08:49.340 --> 00:08:52.080 align:middle line:90% what's the idea? 00:08:52.080 --> 00:08:56.660 align:middle line:84% AUDIENCE: I didn't do as much in the way of skill. 00:08:56.660 --> 00:09:02.880 align:middle line:84% I think you could, in theory, do a sort of utility 00:09:02.880 --> 00:09:05.060 align:middle line:90% of each, a rookie and a vet. 00:09:05.060 --> 00:09:09.880 align:middle line:84% But this is literally just the amount of work 00:09:09.880 --> 00:09:12.890 align:middle line:90% that an employee can do in the-- 00:09:12.890 --> 00:09:14.640 align:middle line:84% HAZHIR RAHMANDAD: Ah, I see, I see, I see. 00:09:14.640 --> 00:09:17.350 align:middle line:84% So why do you need stocks for that? 00:09:17.350 --> 00:09:18.600 align:middle line:90% I'm just trying to understand. 00:09:18.600 --> 00:09:22.040 align:middle line:84% Why is it just the number of rookies multiplied hours 00:09:22.040 --> 00:09:24.220 align:middle line:90% per week or something? 00:09:24.220 --> 00:09:31.920 align:middle line:84% AUDIENCE: It is, but I modeled it this way. 00:09:31.920 --> 00:09:33.880 align:middle line:84% HAZHIR RAHMANDAD: We can come back to it. 00:09:33.880 --> 00:09:37.182 align:middle line:84% I just was trying to understand what the idea is. 00:09:37.182 --> 00:09:38.140 align:middle line:90% We can come back to it. 00:09:38.140 --> 00:09:41.280 align:middle line:84% So then you have the stock and flow for demand. 00:09:41.280 --> 00:09:44.680 align:middle line:84% This is potential demand, current active demand. 00:09:44.680 --> 00:09:47.760 align:middle line:84% That is unmet current active demand. 00:09:47.760 --> 00:09:50.680 align:middle line:90% That is met. 00:09:50.680 --> 00:09:53.180 align:middle line:90% It has kind of a start. 00:09:53.180 --> 00:09:56.480 align:middle line:84% So, basically, these are projects that are happening, 00:09:56.480 --> 00:09:59.640 align:middle line:84% projects that are kind of in the queue, 00:09:59.640 --> 00:10:03.320 align:middle line:84% and projects that could materialize, but are not 00:10:03.320 --> 00:10:04.840 align:middle line:90% yet active. 00:10:04.840 --> 00:10:07.420 align:middle line:90% And you have cycling through. 00:10:07.420 --> 00:10:10.220 align:middle line:84% So maybe, rather than projects, think of them as clients. 00:10:10.220 --> 00:10:13.480 align:middle line:90% 00:10:13.480 --> 00:10:18.000 align:middle line:84% And then you have tracking of the quality of work 00:10:18.000 --> 00:10:21.742 align:middle line:84% through success versus failed work. 00:10:21.742 --> 00:10:23.200 align:middle line:84% We should come back to that and see 00:10:23.200 --> 00:10:25.360 align:middle line:90% how exactly it is formulated. 00:10:25.360 --> 00:10:27.200 align:middle line:84% And then you have reputation, and I 00:10:27.200 --> 00:10:30.200 align:middle line:84% assume reputation then has an impact on recruiting 00:10:30.200 --> 00:10:33.200 align:middle line:90% new potential actors. 00:10:33.200 --> 00:10:36.580 align:middle line:84% I don't see people leaving the field. 00:10:36.580 --> 00:10:40.850 align:middle line:84% So in this sense, once you go from potential demand 00:10:40.850 --> 00:10:45.050 align:middle line:84% to the right-hand side-- well, unless this market expansion is 00:10:45.050 --> 00:10:49.198 align:middle line:84% also potentially market shrinking, that is. 00:10:49.198 --> 00:10:50.990 align:middle line:84% AUDIENCE: Yeah, I think it can go two ways. 00:10:50.990 --> 00:10:52.532 align:middle line:84% HAZHIR RAHMANDAD: It can go two ways. 00:10:52.532 --> 00:10:53.870 align:middle line:90% So then you can have both sides. 00:10:53.870 --> 00:10:55.890 align:middle line:90% Makes sense. 00:10:55.890 --> 00:10:58.230 align:middle line:90% Let's look at the next one. 00:10:58.230 --> 00:11:02.650 align:middle line:90% So this is Jason's model. 00:11:02.650 --> 00:11:05.630 align:middle line:90% And we have recognition. 00:11:05.630 --> 00:11:08.490 align:middle line:90% You have active projects. 00:11:08.490 --> 00:11:13.410 align:middle line:84% You have another addressable market, 00:11:13.410 --> 00:11:16.990 align:middle line:84% where the small market that then turn into active projects. 00:11:16.990 --> 00:11:22.270 align:middle line:84% This is kind of similar to Jake's three big categories. 00:11:22.270 --> 00:11:25.570 align:middle line:84% Although, there might be some differences. 00:11:25.570 --> 00:11:29.770 align:middle line:84% You are trying to model something about the alternatives 00:11:29.770 --> 00:11:32.630 align:middle line:84% and how they are getting recognition. 00:11:32.630 --> 00:11:36.190 align:middle line:84% So you're not keeping it as a fixed construct outside. 00:11:36.190 --> 00:11:38.620 align:middle line:84% But maybe in competition, you may 00:11:38.620 --> 00:11:40.380 align:middle line:84% get more market share or less market share 00:11:40.380 --> 00:11:42.780 align:middle line:90% and that may have an impact. 00:11:42.780 --> 00:11:45.280 align:middle line:84% Then you have rookie experience the structure. 00:11:45.280 --> 00:11:49.100 align:middle line:84% This seems to be coming out of Sterman's standard formulations 00:11:49.100 --> 00:11:54.980 align:middle line:84% there with some effect of experience. 00:11:54.980 --> 00:11:57.343 align:middle line:84% AUDIENCE: Yeah, this is just a non-conserved copula. 00:11:57.343 --> 00:11:58.260 align:middle line:90% HAZHIR RAHMANDAD: Yes. 00:11:58.260 --> 00:11:59.802 align:middle line:84% AUDIENCE: Which is also from Sterman. 00:11:59.802 --> 00:12:01.218 align:middle line:84% HAZHIR RAHMANDAD: Yeah, but you're 00:12:01.218 --> 00:12:02.820 align:middle line:84% using the co-flow of total experience 00:12:02.820 --> 00:12:08.100 align:middle line:84% to match the experienced employees or the rookies? 00:12:08.100 --> 00:12:09.360 align:middle line:90% AUDIENCE: Only experienced. 00:12:09.360 --> 00:12:10.360 align:middle line:90% HAZHIR RAHMANDAD: I see. 00:12:10.360 --> 00:12:10.920 align:middle line:90% Yeah, got it. 00:12:10.920 --> 00:12:11.620 align:middle line:90% AUDIENCE: Yeah. 00:12:11.620 --> 00:12:12.995 align:middle line:84% HAZHIR RAHMANDAD: So you actually 00:12:12.995 --> 00:12:15.320 align:middle line:84% are creating different grades of experience. 00:12:15.320 --> 00:12:17.800 align:middle line:84% So this is actually an important thing to have in mind. 00:12:17.800 --> 00:12:20.820 align:middle line:84% So in typical models of skills, you either 00:12:20.820 --> 00:12:24.760 align:middle line:84% break down the population into different levels of skills, 00:12:24.760 --> 00:12:27.220 align:middle line:84% so it is more of an aging chain, like the experience 00:12:27.220 --> 00:12:29.340 align:middle line:90% versus rookies structure here. 00:12:29.340 --> 00:12:31.980 align:middle line:84% Or you could have a co-flow structure that says, 00:12:31.980 --> 00:12:35.080 align:middle line:84% you have a population of workers and they are building skills. 00:12:35.080 --> 00:12:38.130 align:middle line:84% And that is the co-flow that goes with the population. 00:12:38.130 --> 00:12:42.570 align:middle line:84% Now, it is technically fine to also combine the two here, 00:12:42.570 --> 00:12:44.090 align:middle line:90% just as Jason is doing here. 00:12:44.090 --> 00:12:47.310 align:middle line:84% But it is not as common, and because kind of, 00:12:47.310 --> 00:12:49.070 align:middle line:90% they're doing the same thing. 00:12:49.070 --> 00:12:51.830 align:middle line:84% So it might be a little bit of an overkill to do both of them 00:12:51.830 --> 00:12:52.330 align:middle line:90% together. 00:12:52.330 --> 00:12:54.530 align:middle line:84% But, technically, there is nothing wrong with that. 00:12:54.530 --> 00:12:55.670 align:middle line:90% So that's OK. 00:12:55.670 --> 00:12:59.070 align:middle line:84% And here it basically creates different shades 00:12:59.070 --> 00:13:01.690 align:middle line:84% of experienced employees, so to speak. 00:13:01.690 --> 00:13:04.610 align:middle line:84% So that's what might be happening here. 00:13:04.610 --> 00:13:06.310 align:middle line:84% AUDIENCE: One thing-- we can come back 00:13:06.310 --> 00:13:08.390 align:middle line:84% to it-- that I like was thinking about 00:13:08.390 --> 00:13:12.870 align:middle line:84% is I have utilization kind of adjusting the experience 00:13:12.870 --> 00:13:13.790 align:middle line:90% accrual. 00:13:13.790 --> 00:13:17.430 align:middle line:84% The idea was, if there isn't a ton of demand, 00:13:17.430 --> 00:13:19.590 align:middle line:90% you're not learning as much. 00:13:19.590 --> 00:13:21.432 align:middle line:90% Your experience is valuable. 00:13:21.432 --> 00:13:22.390 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 00:13:22.390 --> 00:13:23.650 align:middle line:90% AUDIENCE: And then it saturates. 00:13:23.650 --> 00:13:24.270 align:middle line:90% I'm not sure of that. 00:13:24.270 --> 00:13:25.270 align:middle line:84% HAZHIR RAHMANDAD: So we should come back to it 00:13:25.270 --> 00:13:27.190 align:middle line:84% because those are important formulations 00:13:27.190 --> 00:13:28.850 align:middle line:90% and could become tricky. 00:13:28.850 --> 00:13:29.610 align:middle line:90% It's not obvious. 00:13:29.610 --> 00:13:32.050 align:middle line:84% So I definitely want to look at it. 00:13:32.050 --> 00:13:33.900 align:middle line:90% Yeah, definitely. 00:13:33.900 --> 00:13:40.000 align:middle line:84% So this is Gerardo's model with customers 00:13:40.000 --> 00:13:43.960 align:middle line:84% that are in the potential, interested, current, 00:13:43.960 --> 00:13:47.160 align:middle line:90% and disappointed groups. 00:13:47.160 --> 00:13:52.120 align:middle line:84% So this structure makes sense with the kind of caveat 00:13:52.120 --> 00:13:57.300 align:middle line:84% that not everybody is going to be disappointed first. 00:13:57.300 --> 00:13:59.300 align:middle line:84% So you may have current-based customers. 00:13:59.300 --> 00:14:02.640 align:middle line:84% Some of them go back into interested or something. 00:14:02.640 --> 00:14:04.480 align:middle line:90% That's possible. 00:14:04.480 --> 00:14:07.000 align:middle line:84% Then you have projects that are defined based 00:14:07.000 --> 00:14:10.360 align:middle line:90% on the active customer base. 00:14:10.360 --> 00:14:12.900 align:middle line:84% Then you have your practitioners. 00:14:12.900 --> 00:14:18.200 align:middle line:84% And with the rookie experience structure, that makes sense. 00:14:18.200 --> 00:14:21.200 align:middle line:84% Then we have the reputation of the field, which 00:14:21.200 --> 00:14:24.360 align:middle line:84% comes from some measure of current quality 00:14:24.360 --> 00:14:27.780 align:middle line:84% that's coming from our rookies, and total work, and so on. 00:14:27.780 --> 00:14:31.380 align:middle line:90% That's pretty clear. 00:14:31.380 --> 00:14:33.090 align:middle line:90% Thank you. 00:14:33.090 --> 00:14:37.490 align:middle line:90% Next one, this is Hassan's. 00:14:37.490 --> 00:14:40.090 align:middle line:90% 00:14:40.090 --> 00:14:42.910 align:middle line:90% So we have novices. 00:14:42.910 --> 00:14:46.090 align:middle line:90% 00:14:46.090 --> 00:14:50.570 align:middle line:84% And Hassan is kind of putting a stock for how many people are 00:14:50.570 --> 00:14:54.090 align:middle line:84% even possible to be potentially joining the field if I'm-- 00:14:54.090 --> 00:14:55.570 align:middle line:84% AUDIENCE: To be available, what are 00:14:55.570 --> 00:14:59.428 align:middle line:84% the ones who have joined, but are not working as of yet? 00:14:59.428 --> 00:15:00.470 align:middle line:90% HAZHIR RAHMANDAD: Oh, OK. 00:15:00.470 --> 00:15:05.690 align:middle line:84% So right now, you are actually capturing 00:15:05.690 --> 00:15:12.490 align:middle line:84% the allocation of projects to people using stocks. 00:15:12.490 --> 00:15:13.570 align:middle line:90% That's an interesting-- 00:15:13.570 --> 00:15:15.010 align:middle line:90% AUDIENCE: There's benefit. 00:15:15.010 --> 00:15:16.510 align:middle line:84% HAZHIR RAHMANDAD: --way of doing it. 00:15:16.510 --> 00:15:19.930 align:middle line:90% So it is fairly operational. 00:15:19.930 --> 00:15:25.070 align:middle line:84% But also, it creates this time constant challenge. 00:15:25.070 --> 00:15:27.090 align:middle line:84% That is, there are dynamics of working 00:15:27.090 --> 00:15:31.610 align:middle line:84% on projects that are in the order of weeks and months. 00:15:31.610 --> 00:15:33.650 align:middle line:84% And then there are dynamics of the field that 00:15:33.650 --> 00:15:35.730 align:middle line:90% are in the orders of years. 00:15:35.730 --> 00:15:39.830 align:middle line:84% And you then need to capture both of them consistently. 00:15:39.830 --> 00:15:41.110 align:middle line:90% It's quite doable. 00:15:41.110 --> 00:15:43.090 align:middle line:84% It just adds a lot of complexity that 00:15:43.090 --> 00:15:45.570 align:middle line:90% may not serve you that much. 00:15:45.570 --> 00:15:48.610 align:middle line:84% That is, the short-term dynamic, the faster dynamics 00:15:48.610 --> 00:15:51.770 align:middle line:84% of allocation, you could assume they are almost 00:15:51.770 --> 00:15:53.650 align:middle line:90% happening instantly. 00:15:53.650 --> 00:15:56.370 align:middle line:84% Based on the current demand and current available-- 00:15:56.370 --> 00:15:58.050 align:middle line:84% current total supply, you just get 00:15:58.050 --> 00:16:00.750 align:middle line:84% a fraction that are allocated and you're done with that. 00:16:00.750 --> 00:16:03.490 align:middle line:84% You don't necessarily need to capture 00:16:03.490 --> 00:16:07.770 align:middle line:84% the movement between projects for this bigger picture 00:16:07.770 --> 00:16:10.330 align:middle line:84% question of how a field evolves over time. 00:16:10.330 --> 00:16:13.830 align:middle line:84% So that's a little bit of these tastings. 00:16:13.830 --> 00:16:15.250 align:middle line:90% But one of the-- 00:16:15.250 --> 00:16:19.110 align:middle line:84% I would suggest going with simpler alternative here, 00:16:19.110 --> 00:16:23.890 align:middle line:84% but the logic is fine if we can come back to it. 00:16:23.890 --> 00:16:28.140 align:middle line:84% And then you have potential demand, odd address demand that 00:16:28.140 --> 00:16:30.820 align:middle line:84% turns into ongoing projects, and then 00:16:30.820 --> 00:16:32.620 align:middle line:84% turns into recently developed projects 00:16:32.620 --> 00:16:34.900 align:middle line:84% that field-- that kind of informs 00:16:34.900 --> 00:16:37.300 align:middle line:90% the recognition of the field. 00:16:37.300 --> 00:16:41.860 align:middle line:84% And then you have the reputation as a separate factor 00:16:41.860 --> 00:16:47.100 align:middle line:84% that comes from experts versus non-experts. 00:16:47.100 --> 00:16:49.620 align:middle line:84% And you have some opportunity, that you should come back 00:16:49.620 --> 00:16:53.860 align:middle line:90% and see how exactly it works. 00:16:53.860 --> 00:17:00.580 align:middle line:84% Next one is Sercin's So this starts 00:17:00.580 --> 00:17:05.180 align:middle line:84% with clients on the top with potential, those 00:17:05.180 --> 00:17:09.099 align:middle line:84% who know about the field, those who are being served 00:17:09.099 --> 00:17:12.280 align:middle line:90% by novices or by experts. 00:17:12.280 --> 00:17:16.380 align:middle line:84% So you're kind of dividing them at the time of getting 00:17:16.380 --> 00:17:18.859 align:middle line:84% the project assigned, whether they get served 00:17:18.859 --> 00:17:21.280 align:middle line:90% by novices or by the experts. 00:17:21.280 --> 00:17:25.220 align:middle line:84% So it's a little bit close to what 00:17:25.220 --> 00:17:31.830 align:middle line:84% Hassan was doing in looking at the explicit allocation right 00:17:31.830 --> 00:17:36.430 align:middle line:84% at the point-- at the project level. 00:17:36.430 --> 00:17:38.630 align:middle line:84% And then you have the same idea here 00:17:38.630 --> 00:17:40.990 align:middle line:84% for the novices and experts serving 00:17:40.990 --> 00:17:46.330 align:middle line:84% the client, so very similar comment to Hassan's model. 00:17:46.330 --> 00:17:50.070 align:middle line:84% And then we have perceived potential clients 00:17:50.070 --> 00:17:55.830 align:middle line:84% per practitioner that may have an effect on getting people 00:17:55.830 --> 00:17:57.550 align:middle line:90% into the field. 00:17:57.550 --> 00:17:58.950 align:middle line:90% It's good. 00:17:58.950 --> 00:18:05.110 align:middle line:84% We should look at some of these later. 00:18:05.110 --> 00:18:08.430 align:middle line:90% This one is Alban's. 00:18:08.430 --> 00:18:14.470 align:middle line:84% So we have not awarded potential clients, 00:18:14.470 --> 00:18:17.990 align:middle line:84% awarded potential clients, active clients. 00:18:17.990 --> 00:18:20.870 align:middle line:84% And active clients are co-flowed with projects 00:18:20.870 --> 00:18:24.440 align:middle line:90% and completed projects. 00:18:24.440 --> 00:18:28.780 align:middle line:84% It's a little bit tricky to get this co-flow to work correctly, 00:18:28.780 --> 00:18:31.020 align:middle line:84% so we have to come back to it and see how it works. 00:18:31.020 --> 00:18:34.040 align:middle line:90% 00:18:34.040 --> 00:18:38.280 align:middle line:84% Then we have the reputation, experts and novices. 00:18:38.280 --> 00:18:42.080 align:middle line:84% You have the demand feedback on growth of the field. 00:18:42.080 --> 00:18:44.120 align:middle line:90% This is great. 00:18:44.120 --> 00:18:46.780 align:middle line:90% Let's go to the next one. 00:18:46.780 --> 00:18:49.800 align:middle line:90% 00:18:49.800 --> 00:18:53.716 align:middle line:90% Ross's model has? 00:18:53.716 --> 00:18:54.760 align:middle line:90% AUDIENCE: Certificate. 00:18:54.760 --> 00:18:55.677 align:middle line:90% HAZHIR RAHMANDAD: Hmm? 00:18:55.677 --> 00:18:56.840 align:middle line:90% AUDIENCE: Certificate. 00:18:56.840 --> 00:18:58.320 align:middle line:90% HAZHIR RAHMANDAD: Certificate. 00:18:58.320 --> 00:19:02.280 align:middle line:90% Yes, it's an interesting idea. 00:19:02.280 --> 00:19:05.640 align:middle line:84% Let's see how it is operationalized later. 00:19:05.640 --> 00:19:11.580 align:middle line:84% So fulfilled demand, potential demand, current demand. 00:19:11.580 --> 00:19:17.940 align:middle line:90% 00:19:17.940 --> 00:19:21.080 align:middle line:84% So, Ross, by fulfilled demand, you 00:19:21.080 --> 00:19:23.230 align:middle line:84% mean demand that has been completed, 00:19:23.230 --> 00:19:24.950 align:middle line:84% but it is in the memory of the people? 00:19:24.950 --> 00:19:29.010 align:middle line:90% So what is project depreciation? 00:19:29.010 --> 00:19:31.810 align:middle line:84% AUDIENCE: I guess it's time the project becomes obsolete 00:19:31.810 --> 00:19:36.130 align:middle line:84% and people become inactive [INAUDIBLE]. 00:19:36.130 --> 00:19:37.630 align:middle line:90% HAZHIR RAHMANDAD: I see. 00:19:37.630 --> 00:19:39.970 align:middle line:84% So you're basically thinking of a constant pool 00:19:39.970 --> 00:19:41.790 align:middle line:90% of potential demand. 00:19:41.790 --> 00:19:44.370 align:middle line:84% And after a while, they come back and-- 00:19:44.370 --> 00:19:46.970 align:middle line:90% yep, got it. 00:19:46.970 --> 00:19:51.810 align:middle line:84% Then you have reputation, quality 00:19:51.810 --> 00:19:59.330 align:middle line:84% of the demand that has been fulfilled, 00:19:59.330 --> 00:20:04.310 align:middle line:84% novices and experts, recognition and reputation. 00:20:04.310 --> 00:20:07.970 align:middle line:84% So these are coming from the size of the project 00:20:07.970 --> 00:20:11.170 align:middle line:90% and the quality of projects. 00:20:11.170 --> 00:20:12.730 align:middle line:90% Cool. 00:20:12.730 --> 00:20:13.250 align:middle line:90% Good. 00:20:13.250 --> 00:20:15.890 align:middle line:84% So I think, actually, we have a little bit more 00:20:15.890 --> 00:20:18.590 align:middle line:84% heterogeneity today in terms of the structures that are used. 00:20:18.590 --> 00:20:22.020 align:middle line:84% So this is kind of very nice to go into some of them 00:20:22.020 --> 00:20:26.480 align:middle line:84% in more depth and see what we can learn from each. 00:20:26.480 --> 00:20:28.980 align:middle line:84% So maybe we can start with Ross's today. 00:20:28.980 --> 00:20:34.560 align:middle line:90% So here, we have-- 00:20:34.560 --> 00:20:39.060 align:middle line:84% let's start from the kind of choice of the customers. 00:20:39.060 --> 00:20:41.660 align:middle line:84% So, basically, this formulation of customers, this 00:20:41.660 --> 00:20:44.600 align:middle line:84% is a conserved set of stocks and flows. 00:20:44.600 --> 00:20:48.220 align:middle line:84% So you have a net total population who may ever 00:20:48.220 --> 00:20:50.400 align:middle line:90% be interested in this field. 00:20:50.400 --> 00:20:53.060 align:middle line:84% They are just cycling through different stages 00:20:53.060 --> 00:20:58.940 align:middle line:84% of being potential demand, interested in demand, 00:20:58.940 --> 00:21:01.680 align:middle line:90% or being served right now. 00:21:01.680 --> 00:21:04.800 align:middle line:84% So demand is kind of the projects that are being served, 00:21:04.800 --> 00:21:07.140 align:middle line:90% I assume. 00:21:07.140 --> 00:21:08.820 align:middle line:90% Those projects get completed. 00:21:08.820 --> 00:21:13.040 align:middle line:84% And these people, after a while, again, would need some projects. 00:21:13.040 --> 00:21:15.400 align:middle line:84% So what is the unit you're using here? 00:21:15.400 --> 00:21:18.032 align:middle line:90% Is it kind of in the-- 00:21:18.032 --> 00:21:19.650 align:middle line:90% it's in the project units. 00:21:19.650 --> 00:21:33.630 align:middle line:90% 00:21:33.630 --> 00:21:38.050 align:middle line:90% The project is-- it's OK. 00:21:38.050 --> 00:21:39.590 align:middle line:90% I think project is fine. 00:21:39.590 --> 00:21:41.730 align:middle line:84% I would think of them more as clients, 00:21:41.730 --> 00:21:44.310 align:middle line:84% but it probably is fine to just think 00:21:44.310 --> 00:21:47.710 align:middle line:84% of each client to have one or n projects at any time, 00:21:47.710 --> 00:21:48.690 align:middle line:90% at any point in time. 00:21:48.690 --> 00:21:51.510 align:middle line:84% Just kind of the concept of a client that 00:21:51.510 --> 00:21:59.590 align:middle line:84% is looking for, basically, a professional, a client 00:21:59.590 --> 00:22:01.390 align:middle line:84% that is being served by a professional, 00:22:01.390 --> 00:22:04.150 align:middle line:84% and a client that doesn't currently 00:22:04.150 --> 00:22:09.790 align:middle line:84% have a job because it takes time to have another project. 00:22:09.790 --> 00:22:13.130 align:middle line:84% That is a little bit easier to imagine than, 00:22:13.130 --> 00:22:15.530 align:middle line:84% for example, a project that has been fulfilled 00:22:15.530 --> 00:22:19.170 align:middle line:84% and then will return into demand. 00:22:19.170 --> 00:22:23.770 align:middle line:84% So it's just relabeling the unit and the variable names. 00:22:23.770 --> 00:22:27.250 align:middle line:84% Perhaps no impact in the underlying math of the model, 00:22:27.250 --> 00:22:31.390 align:middle line:84% but it's a little bit cleaner that way. 00:22:31.390 --> 00:22:36.910 align:middle line:84% But also, let's remember that once it is a fixed pool, 00:22:36.910 --> 00:22:41.930 align:middle line:84% you don't have much opportunity for expanding the field. 00:22:41.930 --> 00:22:46.230 align:middle line:84% Now, there is some opportunity because, here, the project field 00:22:46.230 --> 00:22:48.830 align:middle line:84% awareness controls what is happening 00:22:48.830 --> 00:22:50.810 align:middle line:90% with the projects coming in. 00:22:50.810 --> 00:22:56.270 align:middle line:84% So it's basically the project potential demand 00:22:56.270 --> 00:23:00.510 align:middle line:84% multiplied by some fractional rate that gets materialized 00:23:00.510 --> 00:23:03.470 align:middle line:84% from all of the people who are-- who 00:23:03.470 --> 00:23:05.430 align:middle line:84% have something that can be served 00:23:05.430 --> 00:23:09.790 align:middle line:90% by this type of capability. 00:23:09.790 --> 00:23:13.750 align:middle line:84% A fraction will join, and that has a base fraction multiplied 00:23:13.750 --> 00:23:18.360 align:middle line:84% by some field attractiveness or impact of attractiveness 00:23:18.360 --> 00:23:18.940 align:middle line:90% and growth. 00:23:18.940 --> 00:23:25.640 align:middle line:84% Now, that baseline fractional order rate of 5% per month-- 00:23:25.640 --> 00:23:28.980 align:middle line:84% and the unit, you could just write it as 1 over months. 00:23:28.980 --> 00:23:31.380 align:middle line:90% You don't need DMNL over months. 00:23:31.380 --> 00:23:34.720 align:middle line:84% That's perfectly fine, but it's just easier. 00:23:34.720 --> 00:23:37.960 align:middle line:84% That kind of tells me you need to be careful with field 00:23:37.960 --> 00:23:39.840 align:middle line:90% attractiveness. 00:23:39.840 --> 00:23:42.560 align:middle line:84% If you're thinking of baseline to be maximum, and then 00:23:42.560 --> 00:23:46.160 align:middle line:84% field attractiveness being between 0 and 1 rate, 00:23:46.160 --> 00:23:47.880 align:middle line:84% then it works fine because then there 00:23:47.880 --> 00:23:52.240 align:middle line:84% is no chance that this number becomes too large. 00:23:52.240 --> 00:23:54.920 align:middle line:84% And that's probably how you have formulated it, 00:23:54.920 --> 00:23:56.660 align:middle line:90% not having gone through it. 00:23:56.660 --> 00:23:59.160 align:middle line:84% But when I'm looking at it, I'm saying, 00:23:59.160 --> 00:24:01.760 align:middle line:84% I like this baseline fractional order, then, 00:24:01.760 --> 00:24:06.160 align:middle line:84% to be maximum fractional order rate, 00:24:06.160 --> 00:24:09.880 align:middle line:84% so that I know by looking at the variable names 00:24:09.880 --> 00:24:15.130 align:middle line:84% that field attractiveness this is formulated to be maxed out 00:24:15.130 --> 00:24:17.810 align:middle line:90% at 1 or something like that. 00:24:17.810 --> 00:24:19.070 align:middle line:90% Just taste. 00:24:19.070 --> 00:24:22.570 align:middle line:84% There's nothing mathematical here. 00:24:22.570 --> 00:24:24.650 align:middle line:84% We'll come back to the field attractiveness. 00:24:24.650 --> 00:24:29.410 align:middle line:84% Then the demand that is materialized 00:24:29.410 --> 00:24:31.930 align:middle line:90% needs to be completed. 00:24:31.930 --> 00:24:37.210 align:middle line:84% And that completion rate comes from a maximum completion 00:24:37.210 --> 00:24:43.390 align:middle line:84% rate and an average project completion rate. 00:24:43.390 --> 00:24:46.410 align:middle line:84% So it is demand divided by time to complete 00:24:46.410 --> 00:24:48.350 align:middle line:84% or maximum project completion rate. 00:24:48.350 --> 00:24:50.170 align:middle line:84% So I assume this maximum project completion 00:24:50.170 --> 00:24:54.230 align:middle line:84% rate is coming from the capacity to complete projects. 00:24:54.230 --> 00:24:59.130 align:middle line:90% 00:24:59.130 --> 00:25:03.010 align:middle line:90% And this is a little bit-- 00:25:03.010 --> 00:25:05.850 align:middle line:84% again, variable name-wise, I would 00:25:05.850 --> 00:25:10.580 align:middle line:84% use minimum time to complete or something along those lines 00:25:10.580 --> 00:25:14.200 align:middle line:84% to highlight that this is just for extreme conditions, 00:25:14.200 --> 00:25:17.340 align:middle line:84% not the typical time to complete the project because, 00:25:17.340 --> 00:25:20.160 align:middle line:84% if it is the typical time to complete the project, 00:25:20.160 --> 00:25:23.660 align:middle line:84% it should come from the pathway that 00:25:23.660 --> 00:25:27.260 align:middle line:84% accounts for the labor availability 00:25:27.260 --> 00:25:30.300 align:middle line:84% because that is the normal binding process. 00:25:30.300 --> 00:25:34.540 align:middle line:84% So for this flow, I want it to be binding 90% of the time 00:25:34.540 --> 00:25:37.620 align:middle line:84% by availability of labor consultants. 00:25:37.620 --> 00:25:39.560 align:middle line:90% That's what determines the rate. 00:25:39.560 --> 00:25:41.460 align:middle line:84% But just to be robust, I don't want 00:25:41.460 --> 00:25:44.620 align:middle line:84% it to be completing projects with infinite speed 00:25:44.620 --> 00:25:48.940 align:middle line:84% and/or with the speed that is infeasible given the demand. 00:25:48.940 --> 00:25:52.320 align:middle line:84% So that makes a cleaner formulation, 00:25:52.320 --> 00:25:54.300 align:middle line:84% which means this would be minimum time 00:25:54.300 --> 00:25:56.840 align:middle line:84% to complete projects in the way I think about it. 00:25:56.840 --> 00:26:00.340 align:middle line:84% But now you have put it at 12 months, which is not actually 00:26:00.340 --> 00:26:01.160 align:middle line:90% the minimum time. 00:26:01.160 --> 00:26:04.780 align:middle line:84% It is closer to what the typical time might be. 00:26:04.780 --> 00:26:08.367 align:middle line:84% So can you tell how you were thinking about it? 00:26:08.367 --> 00:26:09.950 align:middle line:84% AUDIENCE: Yeah, I guess I didn't think 00:26:09.950 --> 00:26:11.450 align:middle line:90% about what you just explained. 00:26:11.450 --> 00:26:14.680 align:middle line:90% That's a good point. 00:26:14.680 --> 00:26:15.930 align:middle line:90% HAZHIR RAHMANDAD: That's fine. 00:26:15.930 --> 00:26:18.870 align:middle line:84% So that's kind of whether you are 00:26:18.870 --> 00:26:22.470 align:middle line:84% using this time as the bounding constraint 00:26:22.470 --> 00:26:25.950 align:middle line:84% for extreme conditions or as the main driver. 00:26:25.950 --> 00:26:27.770 align:middle line:84% These are two different use cases. 00:26:27.770 --> 00:26:30.270 align:middle line:84% So if you want to use it as main driver, 00:26:30.270 --> 00:26:33.950 align:middle line:84% that would create the desired completion rate. 00:26:33.950 --> 00:26:37.670 align:middle line:84% Then desired completion rate can be compared, for example, 00:26:37.670 --> 00:26:40.870 align:middle line:84% with the resources available, with the consultants available. 00:26:40.870 --> 00:26:47.390 align:middle line:84% And that can give an actual practical completion rate 00:26:47.390 --> 00:26:50.190 align:middle line:90% that is bounded by both sides. 00:26:50.190 --> 00:26:52.310 align:middle line:84% So that would be the way to do it rather 00:26:52.310 --> 00:26:54.750 align:middle line:90% than doing the minimum here. 00:26:54.750 --> 00:26:57.990 align:middle line:84% That's the alternative way to think about it. 00:26:57.990 --> 00:27:00.610 align:middle line:84% It works, probably works just fine. 00:27:00.610 --> 00:27:03.910 align:middle line:84% I'm not too worried about it not working 00:27:03.910 --> 00:27:07.500 align:middle line:84% in terms of breaking some extreme condition criteria, 00:27:07.500 --> 00:27:09.740 align:middle line:84% but it would have other consequences. 00:27:09.740 --> 00:27:11.480 align:middle line:90% We'll get back to this maximum. 00:27:11.480 --> 00:27:15.020 align:middle line:84% Projecting, then you have a higher order delay for projects, 00:27:15.020 --> 00:27:16.560 align:middle line:90% yes, a third order delay. 00:27:16.560 --> 00:27:18.800 align:middle line:84% This is just how long it takes for people 00:27:18.800 --> 00:27:24.600 align:middle line:84% to come up with a new project that they want to have. 00:27:24.600 --> 00:27:30.320 align:middle line:84% Now, keep in mind that, here, you have a 20-month delay, 00:27:30.320 --> 00:27:36.720 align:middle line:84% minimum, for people to actually realize the next project 00:27:36.720 --> 00:27:43.600 align:middle line:84% or bring up the next project because of that 5% baseline 00:27:43.600 --> 00:27:45.240 align:middle line:90% fractional order rate. 00:27:45.240 --> 00:27:50.140 align:middle line:84% So that's OK, but keep in mind that you have two delays here. 00:27:50.140 --> 00:27:52.720 align:middle line:84% One is that time-to-depreciate delay. 00:27:52.720 --> 00:27:54.480 align:middle line:90% That is 36 months. 00:27:54.480 --> 00:27:56.860 align:middle line:84% And then you have another 20 months delay. 00:27:56.860 --> 00:28:01.040 align:middle line:84% So in average, you have 56 months 00:28:01.040 --> 00:28:08.400 align:middle line:84% of delay between two projects from the time you actually 00:28:08.400 --> 00:28:12.720 align:middle line:84% finish one project until you're bringing in another project. 00:28:12.720 --> 00:28:15.580 align:middle line:84% In your back of the envelope, any calculation you're doing, 00:28:15.580 --> 00:28:18.560 align:middle line:84% you should have both of them in place. 00:28:18.560 --> 00:28:20.540 align:middle line:84% Again, nothing problematic there, 00:28:20.540 --> 00:28:25.640 align:middle line:84% just to think about the implications of the structure 00:28:25.640 --> 00:28:26.500 align:middle line:90% you have put here. 00:28:26.500 --> 00:28:29.900 align:middle line:84% So that's the core for the project side. 00:28:29.900 --> 00:28:33.140 align:middle line:84% Now, let's look at the novices and experts. 00:28:33.140 --> 00:28:35.180 align:middle line:84% So we have people entering the field. 00:28:35.180 --> 00:28:36.420 align:middle line:90% It depends on awareness. 00:28:36.420 --> 00:28:42.400 align:middle line:90% It depends on fulfillment ratio. 00:28:42.400 --> 00:28:49.760 align:middle line:84% It depends on maximum novice entering rate. 00:28:49.760 --> 00:28:53.860 align:middle line:84% So let's actually get a feel for how this is happening. 00:28:53.860 --> 00:28:56.520 align:middle line:84% So maximum novel-- novice entering rate 00:28:56.520 --> 00:28:58.060 align:middle line:90% is actually in person per month. 00:28:58.060 --> 00:29:03.850 align:middle line:84% So this is telling something about capacity of cold field 00:29:03.850 --> 00:29:08.010 align:middle line:84% for attracting new people, the maximum capacity. 00:29:08.010 --> 00:29:10.130 align:middle line:84% Then you have this attractiveness 00:29:10.130 --> 00:29:13.850 align:middle line:84% that feeds into people coming into the field. 00:29:13.850 --> 00:29:17.970 align:middle line:84% And then demand fulfillment ratio 00:29:17.970 --> 00:29:24.030 align:middle line:84% is the ratio of fulfilled demand to maximum demand. 00:29:24.030 --> 00:29:26.390 align:middle line:84% And maximum demand is coming from-- 00:29:26.390 --> 00:29:30.850 align:middle line:90% 00:29:30.850 --> 00:29:34.850 align:middle line:90% oh, some ad hoc number. 00:29:34.850 --> 00:29:38.010 align:middle line:84% So I would be worried about having an ad hoc number 00:29:38.010 --> 00:29:40.330 align:middle line:90% as the maximum demand. 00:29:40.330 --> 00:29:41.730 align:middle line:90% AUDIENCE: Some initial. 00:29:41.730 --> 00:29:43.268 align:middle line:90% I guess the total demand is. 00:29:43.268 --> 00:29:44.310 align:middle line:90% HAZHIR RAHMANDAD: Ah, OK. 00:29:44.310 --> 00:29:48.650 align:middle line:84% This is basically capturing the total number of all of the-- 00:29:48.650 --> 00:29:51.128 align:middle line:84% AUDIENCE: Yeah, that's the initial value for the potential. 00:29:51.128 --> 00:29:52.170 align:middle line:90% HAZHIR RAHMANDAD: Got it. 00:29:52.170 --> 00:29:55.072 align:middle line:84% So, basically, you start with maximum demand in this stock. 00:29:55.072 --> 00:29:57.530 align:middle line:84% And because this is conserved, then you put everything else 00:29:57.530 --> 00:30:00.780 align:middle line:90% at 0 at the beginning. 00:30:00.780 --> 00:30:05.900 align:middle line:84% This is the total number in all of these stocks. 00:30:05.900 --> 00:30:10.980 align:middle line:84% So the demand fulfillment ratio is a term 00:30:10.980 --> 00:30:15.140 align:middle line:84% that gives me the feeling that 1 means we are serving 00:30:15.140 --> 00:30:17.860 align:middle line:90% anybody who wants to be served. 00:30:17.860 --> 00:30:21.540 align:middle line:84% That's the feeling I get from the term. 00:30:21.540 --> 00:30:23.160 align:middle line:84% The way you have formulated here, 00:30:23.160 --> 00:30:25.080 align:middle line:90% it can never get close to 1. 00:30:25.080 --> 00:30:28.100 align:middle line:84% It just can't because the other two stocks 00:30:28.100 --> 00:30:33.120 align:middle line:84% will always have a big chunk of folks who are cycling through. 00:30:33.120 --> 00:30:35.580 align:middle line:84% Even if you have project completion rates that 00:30:35.580 --> 00:30:38.680 align:middle line:84% are happening as fast as technically possible, 00:30:38.680 --> 00:30:40.740 align:middle line:90% it's like zero time. 00:30:40.740 --> 00:30:45.620 align:middle line:84% You still would have 36 versus 20 ratio 00:30:45.620 --> 00:30:47.580 align:middle line:84% of how many people are in the fulfilled demand 00:30:47.580 --> 00:30:50.020 align:middle line:90% versus potential demand. 00:30:50.020 --> 00:30:54.660 align:middle line:84% So this demand fulfillment ratio-- 00:30:54.660 --> 00:30:58.120 align:middle line:84% let's see if my computer comes back. 00:30:58.120 --> 00:31:00.870 align:middle line:90% 00:31:00.870 --> 00:31:03.390 align:middle line:84% Yes, so this demand fulfillment ratio 00:31:03.390 --> 00:31:10.010 align:middle line:84% can never exceed 36 divided by 56 or some ratio like that. 00:31:10.010 --> 00:31:14.350 align:middle line:84% So that puts some kind of an ad hoc normalization 00:31:14.350 --> 00:31:18.270 align:middle line:84% or maximum level for this demand fulfillment ratio 00:31:18.270 --> 00:31:22.830 align:middle line:84% that makes tracking of what will happen next a little bit harder 00:31:22.830 --> 00:31:23.330 align:middle line:90% for me. 00:31:23.330 --> 00:31:29.510 align:middle line:84% So, again, something that is not wrong, but does not-- 00:31:29.510 --> 00:31:31.390 align:middle line:84% the name of the variable doesn't correspond 00:31:31.390 --> 00:31:37.150 align:middle line:84% to what I read from it when I'm looking at it. 00:31:37.150 --> 00:31:40.510 align:middle line:90% But that's aside. 00:31:40.510 --> 00:31:45.910 align:middle line:84% So 1 minus this demand fulfillment ratio 00:31:45.910 --> 00:31:50.050 align:middle line:90% is giving a signal of-- 00:31:50.050 --> 00:31:55.510 align:middle line:84% AUDIENCE: I was trying to have some excess demand, some proxy 00:31:55.510 --> 00:31:57.198 align:middle line:90% for [INAUDIBLE]. 00:31:57.198 --> 00:31:58.740 align:middle line:84% HAZHIR RAHMANDAD: Some excess demand. 00:31:58.740 --> 00:32:00.000 align:middle line:90% Can you explain? 00:32:00.000 --> 00:32:03.360 align:middle line:84% AUDIENCE: Something like, if there is excess demand, 00:32:03.360 --> 00:32:04.620 align:middle line:90% then novices enter. 00:32:04.620 --> 00:32:05.620 align:middle line:90% HAZHIR RAHMANDAD: I see. 00:32:05.620 --> 00:32:09.320 align:middle line:84% So you're saying novices are entering based on excess demand, 00:32:09.320 --> 00:32:13.040 align:middle line:84% or based on some signal that demand is not being served 00:32:13.040 --> 00:32:15.545 align:middle line:90% as well as it could have been? 00:32:15.545 --> 00:32:17.420 align:middle line:84% AUDIENCE: Yeah, that's what I'm trying to do. 00:32:17.420 --> 00:32:18.440 align:middle line:90% But [INAUDIBLE]. 00:32:18.440 --> 00:32:20.690 align:middle line:84% HAZHIR RAHMANDAD: It may not be doing exactly what you 00:32:20.690 --> 00:32:23.520 align:middle line:84% want because the reason this demand fulfillment ratio might 00:32:23.520 --> 00:32:26.920 align:middle line:84% be small could be that the reputation of the field 00:32:26.920 --> 00:32:31.680 align:middle line:84% is pretty terrible, and people are not coming to the field 00:32:31.680 --> 00:32:36.620 align:middle line:90% to be served. 00:32:36.620 --> 00:32:39.440 align:middle line:90% 00:32:39.440 --> 00:32:43.240 align:middle line:84% Or it could be because there is a lot of demand, 00:32:43.240 --> 00:32:45.080 align:middle line:84% but you don't have enough supply. 00:32:45.080 --> 00:32:49.720 align:middle line:84% And, therefore, a lot of people are stuck in the demand stock. 00:32:49.720 --> 00:32:53.320 align:middle line:84% So you can have an inflated demand stock, 00:32:53.320 --> 00:32:56.610 align:middle line:84% which is a signal for these people to-- for the novices 00:32:56.610 --> 00:32:57.850 align:middle line:90% to join the field. 00:32:57.850 --> 00:33:01.650 align:middle line:84% Or you may have an infinite kind of inflated potential demand 00:33:01.650 --> 00:33:06.730 align:middle line:84% stock, which is a signal not to come to this field in some ways 00:33:06.730 --> 00:33:10.090 align:middle line:84% because this is not a reputable field. 00:33:10.090 --> 00:33:12.990 align:middle line:84% So both of those are coded the same way here. 00:33:12.990 --> 00:33:14.610 align:middle line:84% And, therefore, this would not do 00:33:14.610 --> 00:33:18.110 align:middle line:84% what you want to do because it's not clear which of these-- 00:33:18.110 --> 00:33:20.570 align:middle line:84% AUDIENCE: But would it fill the attractiveness, 00:33:20.570 --> 00:33:22.660 align:middle line:90% take care of this? 00:33:22.660 --> 00:33:24.410 align:middle line:84% HAZHIR RAHMANDAD: So field attractiveness, 00:33:24.410 --> 00:33:25.990 align:middle line:90% where does it take care of it? 00:33:25.990 --> 00:33:30.470 align:middle line:84% So it comes into here as a separate factor. 00:33:30.470 --> 00:33:33.610 align:middle line:90% 00:33:33.610 --> 00:33:35.210 align:middle line:84% So the way I would think about it 00:33:35.210 --> 00:33:38.770 align:middle line:84% is, what is visible to a novice in this case? 00:33:38.770 --> 00:33:41.130 align:middle line:84% What are the signals that they may see? 00:33:41.130 --> 00:33:43.330 align:middle line:84% So the signals they may see is that this 00:33:43.330 --> 00:33:47.070 align:middle line:90% is a booming consulting space. 00:33:47.070 --> 00:33:48.450 align:middle line:90% What does booming means? 00:33:48.450 --> 00:33:52.770 align:middle line:84% That there is a queue of people who want 00:33:52.770 --> 00:33:56.610 align:middle line:90% projects to be completed and-- 00:33:56.610 --> 00:33:59.350 align:middle line:84% so if I join, there will be projects for me. 00:33:59.350 --> 00:34:03.490 align:middle line:84% That's probably the simple, easy, fairly straightforward 00:34:03.490 --> 00:34:04.770 align:middle line:90% signal. 00:34:04.770 --> 00:34:08.170 align:middle line:84% You can do that by just looking at the delivery delay 00:34:08.170 --> 00:34:11.590 align:middle line:84% compared to kind of desired delivery delay, 00:34:11.590 --> 00:34:13.929 align:middle line:90% or something along those lines. 00:34:13.929 --> 00:34:17.010 align:middle line:84% So as long as you capture how fast projects are getting 00:34:17.010 --> 00:34:21.130 align:middle line:84% completed compared to how fast, typically, clients 00:34:21.130 --> 00:34:23.650 align:middle line:84% want them to be completed, that will give you 00:34:23.650 --> 00:34:27.810 align:middle line:84% the signal for whether this is an attractive space 00:34:27.810 --> 00:34:29.179 align:middle line:90% to join for that. 00:34:29.179 --> 00:34:29.929 align:middle line:90% AUDIENCE: Backlog. 00:34:29.929 --> 00:34:31.590 align:middle line:84% HAZHIR RAHMANDAD: Yeah, some version of a backlog. 00:34:31.590 --> 00:34:32.965 align:middle line:84% And you already have the backlog. 00:34:32.965 --> 00:34:34.350 align:middle line:90% So demand is actually a backlog. 00:34:34.350 --> 00:34:35.570 align:middle line:90% This is being served. 00:34:35.570 --> 00:34:39.810 align:middle line:84% But delivery delay compared to the desired delivery delay 00:34:39.810 --> 00:34:42.665 align:middle line:90% is the core idea. 00:34:42.665 --> 00:34:44.290 align:middle line:84% AUDIENCE: Does it make sense to compare 00:34:44.290 --> 00:34:48.810 align:middle line:84% the backlog of the supply, kind of the capacity? 00:34:48.810 --> 00:34:52.520 align:middle line:84% HAZHIR RAHMANDAD: Compare the backlog to capacity. 00:34:52.520 --> 00:34:54.977 align:middle line:84% So backlog to capacity is the delivery delay. 00:34:54.977 --> 00:34:56.060 align:middle line:90% AUDIENCE: Oh, yeah, right, 00:34:56.060 --> 00:34:58.018 align:middle line:84% HAZHIR RAHMANDAD: That's by definition exactly. 00:34:58.018 --> 00:34:59.180 align:middle line:90% AUDIENCE: Thank you. 00:34:59.180 --> 00:35:02.100 align:middle line:90% But is it capacity of-- 00:35:02.100 --> 00:35:05.640 align:middle line:84% HAZHIR RAHMANDAD: So think about capacity as projects per month. 00:35:05.640 --> 00:35:06.460 align:middle line:90% AUDIENCE: Yeah. 00:35:06.460 --> 00:35:08.760 align:middle line:84% HAZHIR RAHMANDAD: Think about backlog as projects. 00:35:08.760 --> 00:35:10.900 align:middle line:84% So project divided by project per month 00:35:10.900 --> 00:35:14.420 align:middle line:84% is, in months, the delivery delay, average delivery delay. 00:35:14.420 --> 00:35:16.560 align:middle line:84% Then you have the desired delivery delay. 00:35:16.560 --> 00:35:19.380 align:middle line:84% And the balance of those two would 00:35:19.380 --> 00:35:22.720 align:middle line:84% be one way to get this signal for novices entering. 00:35:22.720 --> 00:35:26.940 align:middle line:84% AUDIENCE: So is this kind of analogous to the pressure 00:35:26.940 --> 00:35:28.443 align:middle line:90% to expand capacity in market? 00:35:28.443 --> 00:35:29.860 align:middle line:84% HAZHIR RAHMANDAD: Yes, it would be 00:35:29.860 --> 00:35:32.720 align:middle line:84% analogous to pressure to expand capacity. 00:35:32.720 --> 00:35:36.100 align:middle line:90% 00:35:36.100 --> 00:35:39.260 align:middle line:84% Yes, I think-- and this is actually a decent signal. 00:35:39.260 --> 00:35:42.540 align:middle line:84% In these cases, you don't have very clean signal 00:35:42.540 --> 00:35:47.100 align:middle line:84% whether this is a good field other than that, OK, consultants 00:35:47.100 --> 00:35:49.430 align:middle line:90% in this space are pretty busy. 00:35:49.430 --> 00:35:50.990 align:middle line:84% There are a lot of jobs for them, 00:35:50.990 --> 00:35:55.363 align:middle line:84% or now there is not much job in this space. 00:35:55.363 --> 00:35:56.530 align:middle line:90% AUDIENCE: I have a question. 00:35:56.530 --> 00:35:59.450 align:middle line:84% Shouldn't it also be tied to reputation? 00:35:59.450 --> 00:36:02.770 align:middle line:84% Because, well, what if reputation is low 00:36:02.770 --> 00:36:07.222 align:middle line:84% and then you are looking at the very short period? 00:36:07.222 --> 00:36:08.430 align:middle line:90% HAZHIR RAHMANDAD: Absolutely. 00:36:08.430 --> 00:36:12.690 align:middle line:84% So that's a very good kind of second input into the novice. 00:36:12.690 --> 00:36:15.910 align:middle line:84% So is this a busy field that has good reputation 00:36:15.910 --> 00:36:18.970 align:middle line:84% versus a busy field that doesn't have a good reputation? 00:36:18.970 --> 00:36:21.790 align:middle line:84% Makes a difference because that is informative about the longer 00:36:21.790 --> 00:36:22.910 align:middle line:90% term trends. 00:36:22.910 --> 00:36:24.890 align:middle line:84% Yes, it might be busy, but it is a fad. 00:36:24.890 --> 00:36:27.590 align:middle line:84% And everybody who kind of has had experience with this says, 00:36:27.590 --> 00:36:29.030 align:middle line:90% don't go there. 00:36:29.030 --> 00:36:31.290 align:middle line:90% So then you can add that factor. 00:36:31.290 --> 00:36:36.930 align:middle line:84% Absolutely, Maria, that does come in as a nice addition. 00:36:36.930 --> 00:36:38.950 align:middle line:84% So the field attractiveness, if it 00:36:38.950 --> 00:36:41.390 align:middle line:84% is capturing something about quality, 00:36:41.390 --> 00:36:43.950 align:middle line:90% is a good input into this. 00:36:43.950 --> 00:36:50.120 align:middle line:84% Again, I'm also need to look at the formulations for field 00:36:50.120 --> 00:36:52.720 align:middle line:84% attractiveness to see whether that is already capturing 00:36:52.720 --> 00:36:55.240 align:middle line:84% the right non-linearity there, or you need 00:36:55.240 --> 00:36:57.060 align:middle line:84% some additional transformation here, 00:36:57.060 --> 00:36:59.800 align:middle line:84% but that-- because you're using it in two different places, just 00:36:59.800 --> 00:37:01.060 align:middle line:90% as a multiplier. 00:37:01.060 --> 00:37:04.640 align:middle line:84% So you have it as a multiplier into entering, 00:37:04.640 --> 00:37:07.280 align:middle line:84% and you have field attractiveness as a multiplier 00:37:07.280 --> 00:37:08.880 align:middle line:90% into projects ordered. 00:37:08.880 --> 00:37:13.620 align:middle line:84% And these are two different use cases of an underlying concept. 00:37:13.620 --> 00:37:17.180 align:middle line:84% One is impacting the novice's decision to enter a field. 00:37:17.180 --> 00:37:22.720 align:middle line:84% The other is impacting project owners to start new projects. 00:37:22.720 --> 00:37:25.440 align:middle line:84% So the functions for decisions of these two groups 00:37:25.440 --> 00:37:27.280 align:middle line:90% are different. 00:37:27.280 --> 00:37:28.840 align:middle line:84% You don't have degrees of freedom 00:37:28.840 --> 00:37:30.880 align:middle line:84% to separate these two when you just 00:37:30.880 --> 00:37:32.880 align:middle line:90% have both of them multiplied. 00:37:32.880 --> 00:37:35.880 align:middle line:84% So I would probably pick one of them, 00:37:35.880 --> 00:37:39.160 align:middle line:84% anchor to that so that my project attractiveness, or field 00:37:39.160 --> 00:37:44.160 align:middle line:84% attractiveness, is anchored, for example, to project owners. 00:37:44.160 --> 00:37:47.370 align:middle line:84% That is, its ups and downs are consistent with what 00:37:47.370 --> 00:37:51.250 align:middle line:84% a project owner's weight on attractiveness is. 00:37:51.250 --> 00:37:54.850 align:middle line:84% So that I don't need a free parameter for the project owner. 00:37:54.850 --> 00:38:00.010 align:middle line:84% But when I come to novices, I will add a new parameter 00:38:00.010 --> 00:38:06.090 align:middle line:84% to regulate the strength of the impact of this perception 00:38:06.090 --> 00:38:09.170 align:middle line:84% on novice behavior because the two are 00:38:09.170 --> 00:38:10.430 align:middle line:90% using different functions. 00:38:10.430 --> 00:38:12.730 align:middle line:84% These are different decisions, so you can't just 00:38:12.730 --> 00:38:14.390 align:middle line:90% randomly use the same thing. 00:38:14.390 --> 00:38:15.310 align:middle line:90% Is that clear? 00:38:15.310 --> 00:38:19.630 align:middle line:84% This is kind of potentially important general idea. 00:38:19.630 --> 00:38:25.426 align:middle line:90% 00:38:25.426 --> 00:38:28.330 align:middle line:84% And, folks, if you're online, have an idea, 00:38:28.330 --> 00:38:31.810 align:middle line:84% please just jump in, or have a question because I'm not seeing 00:38:31.810 --> 00:38:33.710 align:middle line:90% your screen in front of me. 00:38:33.710 --> 00:38:36.110 align:middle line:84% So if you raise your hand, I may not see it. 00:38:36.110 --> 00:38:39.930 align:middle line:90% So please just jump in. 00:38:39.930 --> 00:38:41.450 align:middle line:90% So people enter. 00:38:41.450 --> 00:38:42.510 align:middle line:90% They are novice. 00:38:42.510 --> 00:38:44.600 align:middle line:90% There is a training process. 00:38:44.600 --> 00:38:49.560 align:middle line:84% The training process can be bounded by two factors. 00:38:49.560 --> 00:38:56.660 align:middle line:84% So you have the minimum of the novice, training per expert. 00:38:56.660 --> 00:39:01.720 align:middle line:84% So you divide all of that by time to training. 00:39:01.720 --> 00:39:11.940 align:middle line:90% 00:39:11.940 --> 00:39:14.060 align:middle line:84% So this is an interesting formulation. 00:39:14.060 --> 00:39:16.420 align:middle line:90% I think it is fine. 00:39:16.420 --> 00:39:17.880 align:middle line:90% There is nothing wrong here. 00:39:17.880 --> 00:39:19.800 align:middle line:84% So you have two different pathways. 00:39:19.800 --> 00:39:24.200 align:middle line:84% One is that novices take some time to be trained. 00:39:24.200 --> 00:39:27.060 align:middle line:84% So it is novice divided by time to training. 00:39:27.060 --> 00:39:28.980 align:middle line:90% That's one constraint. 00:39:28.980 --> 00:39:31.220 align:middle line:84% If you don't have people, you can train them. 00:39:31.220 --> 00:39:35.620 align:middle line:84% The other constraint is the expert ability or expert 00:39:35.620 --> 00:39:38.380 align:middle line:90% capacity for training novices. 00:39:38.380 --> 00:39:41.270 align:middle line:84% And you get that by experts multiplied 00:39:41.270 --> 00:39:44.330 align:middle line:84% by a fraction of experts that are dedicated to training. 00:39:44.330 --> 00:39:45.950 align:middle line:84% That is, what fraction of each expert 00:39:45.950 --> 00:39:47.870 align:middle line:90% might be dedicated to training? 00:39:47.870 --> 00:39:50.070 align:middle line:90% That's kind of sets a maximum. 00:39:50.070 --> 00:39:57.330 align:middle line:84% And training per expert, that is trainee per expert. 00:39:57.330 --> 00:40:03.590 align:middle line:84% That is just translating, how many trainee equivalents 00:40:03.590 --> 00:40:07.330 align:middle line:90% are going for each expert? 00:40:07.330 --> 00:40:10.490 align:middle line:84% I don't know if you need the trainee per expert. 00:40:10.490 --> 00:40:13.150 align:middle line:90% 00:40:13.150 --> 00:40:16.510 align:middle line:90% Yeah, probably you need it. 00:40:16.510 --> 00:40:21.070 align:middle line:84% AUDIENCE: This is weird because, in the book, 00:40:21.070 --> 00:40:23.450 align:middle line:84% I think you have units of experts, 00:40:23.450 --> 00:40:25.270 align:middle line:90% which has always bothered me. 00:40:25.270 --> 00:40:31.390 align:middle line:84% And Russ is doing a correction to give you units of trainees, 00:40:31.390 --> 00:40:33.775 align:middle line:90% which makes more sense. 00:40:33.775 --> 00:40:35.150 align:middle line:84% HAZHIR RAHMANDAD: Yes, so I think 00:40:35.150 --> 00:40:38.470 align:middle line:84% you can have it because, ultimately, this trainee 00:40:38.470 --> 00:40:41.670 align:middle line:84% per expert is multiplied by fraction of experts dedicated 00:40:41.670 --> 00:40:42.610 align:middle line:90% to training. 00:40:42.610 --> 00:40:44.927 align:middle line:90% AUDIENCE: Yeah. 00:40:44.927 --> 00:40:46.510 align:middle line:84% HAZHIR RAHMANDAD: These two parameters 00:40:46.510 --> 00:40:48.890 align:middle line:84% are fully collinear with each other. 00:40:48.890 --> 00:40:51.710 align:middle line:84% I don't think they are used anywhere else, 00:40:51.710 --> 00:40:54.410 align:middle line:84% so which means you have, really, one degree of freedom. 00:40:54.410 --> 00:41:00.150 align:middle line:84% This is like SIR model with contact rate and infection rate. 00:41:00.150 --> 00:41:01.130 align:middle line:90% You can have them. 00:41:01.130 --> 00:41:02.385 align:middle line:90% Conceptually, there is-- 00:41:02.385 --> 00:41:03.510 align:middle line:90% AUDIENCE: It gives clarity. 00:41:03.510 --> 00:41:04.968 align:middle line:84% HAZHIR RAHMANDAD: It gives clarity, 00:41:04.968 --> 00:41:08.630 align:middle line:84% but it is not doing anything for you mathematically. 00:41:08.630 --> 00:41:09.930 align:middle line:90% There's nothing wrong here. 00:41:09.930 --> 00:41:13.790 align:middle line:90% 00:41:13.790 --> 00:41:17.070 align:middle line:84% Yes, there is kind of a little bit of a taste issue, 00:41:17.070 --> 00:41:21.290 align:middle line:84% that you have this minimum of the stock things, stock things. 00:41:21.290 --> 00:41:25.110 align:middle line:84% And then you divide the total thing by this time to training. 00:41:25.110 --> 00:41:28.010 align:middle line:84% And people get confused with this formulation. 00:41:28.010 --> 00:41:30.130 align:middle line:84% When I opened, it I said, well, what is happening? 00:41:30.130 --> 00:41:32.910 align:middle line:84% It's the minimum of novice and something. 00:41:32.910 --> 00:41:34.590 align:middle line:84% And the reason is that I normally think 00:41:34.590 --> 00:41:36.530 align:middle line:90% of two different processes. 00:41:36.530 --> 00:41:39.960 align:middle line:84% One is that novices get trained because they are kind 00:41:39.960 --> 00:41:42.040 align:middle line:90% of spending time on the job. 00:41:42.040 --> 00:41:44.540 align:middle line:84% That happens regardless of getting training. 00:41:44.540 --> 00:41:48.120 align:middle line:84% So that is novice divided by time to training. 00:41:48.120 --> 00:41:49.780 align:middle line:90% That is coming from novices. 00:41:49.780 --> 00:41:50.840 align:middle line:90% That's one path. 00:41:50.840 --> 00:41:53.140 align:middle line:84% And then the other path, I would say, OK, 00:41:53.140 --> 00:41:58.080 align:middle line:84% each expert, if they spend full time training people, 00:41:58.080 --> 00:42:02.680 align:middle line:84% they will be able to train half a novice per month 00:42:02.680 --> 00:42:05.720 align:middle line:84% to become an expert or something, some number, 00:42:05.720 --> 00:42:06.980 align:middle line:90% if that is the full time. 00:42:06.980 --> 00:42:10.520 align:middle line:84% So that is my trainee per expert. 00:42:10.520 --> 00:42:13.160 align:middle line:84% Then I'll multiply that by the fraction of time 00:42:13.160 --> 00:42:17.240 align:middle line:84% that experts are going to invest in training, 00:42:17.240 --> 00:42:22.120 align:middle line:84% and that becomes a second channel. 00:42:22.120 --> 00:42:27.340 align:middle line:84% And that will take care of the training rate for me. 00:42:27.340 --> 00:42:28.680 align:middle line:90% AUDIENCE: Like a separate flow? 00:42:28.680 --> 00:42:29.763 align:middle line:90% HAZHIR RAHMANDAD: Exactly. 00:42:29.763 --> 00:42:34.360 align:middle line:84% Just not separate flow, but separate pathway for the flow. 00:42:34.360 --> 00:42:38.110 align:middle line:84% Now, then there is the question of how to aggregate these two. 00:42:38.110 --> 00:42:40.130 align:middle line:84% There is the natural training coming 00:42:40.130 --> 00:42:43.130 align:middle line:90% from my experience of working. 00:42:43.130 --> 00:42:45.450 align:middle line:84% Then there is this expert-driven training 00:42:45.450 --> 00:42:49.130 align:middle line:84% that is the extra boost that I get in my training. 00:42:49.130 --> 00:42:51.410 align:middle line:84% One way is to look at the minimum and say, 00:42:51.410 --> 00:42:53.030 align:middle line:90% these two are overlapping. 00:42:53.030 --> 00:42:55.650 align:middle line:84% So if the experts are spending time, 00:42:55.650 --> 00:42:57.850 align:middle line:84% they don't boost it unless they spend 00:42:57.850 --> 00:43:00.530 align:middle line:84% enough that goes beyond what I'm getting 00:43:00.530 --> 00:43:05.050 align:middle line:84% from my-- getting my hands dirty on a day-to-day basis. 00:43:05.050 --> 00:43:07.610 align:middle line:84% The other formulation says, no, actually they 00:43:07.610 --> 00:43:11.070 align:middle line:84% are completely independent and, therefore, they are additive. 00:43:11.070 --> 00:43:12.970 align:middle line:90% You can add them together. 00:43:12.970 --> 00:43:15.110 align:middle line:84% But if you're adding them to the other, 00:43:15.110 --> 00:43:19.250 align:middle line:84% then you need to be careful not to drain the novices too fast 00:43:19.250 --> 00:43:21.010 align:middle line:90% if you don't have any novices. 00:43:21.010 --> 00:43:23.850 align:middle line:84% So then the part that comes from the experts 00:43:23.850 --> 00:43:26.930 align:middle line:84% should get some upper bound that is bounded by minimum time 00:43:26.930 --> 00:43:30.570 align:middle line:84% to training, which is different from the times to training. 00:43:30.570 --> 00:43:38.740 align:middle line:84% So that would be perhaps a more complex, but also more realistic 00:43:38.740 --> 00:43:43.540 align:middle line:84% formulation that allows you to separate these two explicitly. 00:43:43.540 --> 00:43:46.760 align:middle line:84% Right now, you're kind of putting all of them together. 00:43:46.760 --> 00:43:50.660 align:middle line:84% This will work, but a little bit harder to explain. 00:43:50.660 --> 00:43:54.140 align:middle line:84% The other problem-- not problem, but what 00:43:54.140 --> 00:43:56.160 align:middle line:84% I, kind of looking at this, would say, 00:43:56.160 --> 00:44:01.820 align:middle line:84% OK, this is not dependent on how much work these people have. 00:44:01.820 --> 00:44:06.500 align:middle line:84% So if you have very busy schedule or no busyness 00:44:06.500 --> 00:44:08.900 align:middle line:84% of a schedule, this is not capturing it 00:44:08.900 --> 00:44:12.180 align:middle line:84% in terms of impact on training or impact 00:44:12.180 --> 00:44:15.700 align:middle line:84% on the expert time spent on training. 00:44:15.700 --> 00:44:18.440 align:middle line:84% That's kind of-- both of them are kind of missing. 00:44:18.440 --> 00:44:20.523 align:middle line:84% AUDIENCE: I feel like I had just entered the math. 00:44:20.523 --> 00:44:21.898 align:middle line:84% HAZHIR RAHMANDAD: So that's good. 00:44:21.898 --> 00:44:22.920 align:middle line:90% So we can look at that. 00:44:22.920 --> 00:44:25.220 align:middle line:84% So the expert fraction dedicated to training 00:44:25.220 --> 00:44:29.460 align:middle line:84% is impacting the maximum project completion rate. 00:44:29.460 --> 00:44:32.790 align:middle line:90% So let's look into it. 00:44:32.790 --> 00:44:37.510 align:middle line:90% It's a little bit intimidating. 00:44:37.510 --> 00:44:39.810 align:middle line:84% For now, let's put aside a certificate. 00:44:39.810 --> 00:44:42.710 align:middle line:84% I assume it is not a big deal yet. 00:44:42.710 --> 00:44:46.910 align:middle line:84% So you have the maximum novel product-- novice productivity 00:44:46.910 --> 00:44:50.110 align:middle line:84% by novices, maximum expert productivity, 1 minus fraction 00:44:50.110 --> 00:44:52.270 align:middle line:90% dedicated to training maximal. 00:44:52.270 --> 00:44:54.130 align:middle line:90% So this is the total capacity. 00:44:54.130 --> 00:44:56.430 align:middle line:84% AUDIENCE: But I guess it doesn't affect them 00:44:56.430 --> 00:44:59.350 align:middle line:90% until you hit the maximum. 00:44:59.350 --> 00:45:01.590 align:middle line:84% HAZHIR RAHMANDAD: It doesn't have an effect on what? 00:45:01.590 --> 00:45:03.590 align:middle line:84% AUDIENCE: The fraction that I take out 00:45:03.590 --> 00:45:06.550 align:middle line:84% does not have an effect on the project completion, 00:45:06.550 --> 00:45:11.048 align:middle line:84% unless the project completion hits the next one. 00:45:11.048 --> 00:45:12.590 align:middle line:84% HAZHIR RAHMANDAD: So that certificate 00:45:12.590 --> 00:45:14.890 align:middle line:90% is about project completion? 00:45:14.890 --> 00:45:16.670 align:middle line:90% AUDIENCE: That's just a policy. 00:45:16.670 --> 00:45:24.610 align:middle line:84% HAZHIR RAHMANDAD: Oh, yeah, so the part that's kind of not, 00:45:24.610 --> 00:45:27.870 align:middle line:84% in my view, fully yet consistent with what you are trying to do 00:45:27.870 --> 00:45:35.120 align:middle line:84% is that, if you have few novices, then 00:45:35.120 --> 00:45:38.720 align:middle line:84% experts are not spending any time on training. 00:45:38.720 --> 00:45:41.920 align:middle line:84% So the maximum x-- so 1 minus fraction 00:45:41.920 --> 00:45:44.240 align:middle line:84% of expert dedicated to training should not 00:45:44.240 --> 00:45:45.700 align:middle line:90% be 1 minus that fraction. 00:45:45.700 --> 00:45:49.040 align:middle line:84% It should be 1 minus a very small thing, or 0. 00:45:49.040 --> 00:45:51.640 align:middle line:84% That is not endogenously captured here. 00:45:51.640 --> 00:45:52.820 align:middle line:90% Does that make sense? 00:45:52.820 --> 00:45:53.480 align:middle line:90% AUDIENCE: Yeah. 00:45:53.480 --> 00:45:56.360 align:middle line:84% HAZHIR RAHMANDAD: So it does need 00:45:56.360 --> 00:45:59.360 align:middle line:84% to be tracking, what fraction of experts 00:45:59.360 --> 00:46:04.040 align:middle line:84% are actually being dedicated to training because 00:46:04.040 --> 00:46:05.440 align:middle line:84% of the need for training, because 00:46:05.440 --> 00:46:07.360 align:middle line:90% of existence of these novices? 00:46:07.360 --> 00:46:10.267 align:middle line:84% So there is a little bit of a demand and supply. 00:46:10.267 --> 00:46:12.100 align:middle line:84% So you are just focusing on the supply side, 00:46:12.100 --> 00:46:16.800 align:middle line:84% but what was the demand for training that got material-- 00:46:16.800 --> 00:46:18.780 align:middle line:90% that got served by experts? 00:46:18.780 --> 00:46:22.500 align:middle line:84% And then based on that, what is the realized fraction, dedicated 00:46:22.500 --> 00:46:23.000 align:middle line:90% training? 00:46:23.000 --> 00:46:27.080 align:middle line:84% That realized fraction is now what is reducing the expert 00:46:27.080 --> 00:46:28.900 align:middle line:90% availability for training. 00:46:28.900 --> 00:46:32.160 align:middle line:90% 00:46:32.160 --> 00:46:37.480 align:middle line:84% I don't remember how exactly Sterman's text handles this 00:46:37.480 --> 00:46:39.855 align:middle line:84% because he has an example that goes through this, 00:46:39.855 --> 00:46:40.980 align:middle line:90% but I don't know right now. 00:46:40.980 --> 00:46:44.827 align:middle line:84% AUDIENCE: It's not endogenously handled, maybe in a challenge 00:46:44.827 --> 00:46:45.660 align:middle line:90% that I haven't done. 00:46:45.660 --> 00:46:48.200 align:middle line:84% HAZHIR RAHMANDAD: Yeah, I think there might be a challenge that 00:46:48.200 --> 00:46:48.950 align:middle line:90% goes through that. 00:46:48.950 --> 00:46:51.880 align:middle line:84% But this is a non-trivial problem 00:46:51.880 --> 00:46:55.040 align:middle line:84% because you need to create explicit demand and supply 00:46:55.040 --> 00:46:57.980 align:middle line:84% and then calculate how much they are being served. 00:46:57.980 --> 00:47:02.160 align:middle line:84% Now, this maximum project completion rate 00:47:02.160 --> 00:47:05.160 align:middle line:84% is then compared to this desired project completion rate 00:47:05.160 --> 00:47:08.880 align:middle line:84% that we had here, that kind of drains the projects. 00:47:08.880 --> 00:47:11.680 align:middle line:84% And then you have the quality of fulfilled demand 00:47:11.680 --> 00:47:19.280 align:middle line:84% coming from novice fraction, effect on quality. 00:47:19.280 --> 00:47:25.410 align:middle line:84% So if you have novices, a bigger fraction of workers 00:47:25.410 --> 00:47:32.610 align:middle line:84% to be novices, then your quality goes down. 00:47:32.610 --> 00:47:38.170 align:middle line:84% And this is important, to have the same quality memory loss 00:47:38.170 --> 00:47:40.210 align:middle line:90% as this project depreciation. 00:47:40.210 --> 00:47:42.770 align:middle line:84% And the way you are enforcing it is 00:47:42.770 --> 00:47:45.450 align:middle line:84% using the same time constants, time to depreciate, 00:47:45.450 --> 00:47:47.270 align:middle line:90% and the same delay order. 00:47:47.270 --> 00:47:48.230 align:middle line:90% So that is fine. 00:47:48.230 --> 00:47:51.210 align:middle line:84% Mathematically, it gives you the same result, 00:47:51.210 --> 00:47:53.090 align:middle line:84% but the better way to do it is actually 00:47:53.090 --> 00:47:58.850 align:middle line:84% to use the project depreciation to drain this directly 00:47:58.850 --> 00:48:01.550 align:middle line:84% because then, if you change the time to depreciate, 00:48:01.550 --> 00:48:03.310 align:middle line:90% that's already in both places. 00:48:03.310 --> 00:48:06.250 align:middle line:84% But if you change the order of delay, for example, 00:48:06.250 --> 00:48:09.610 align:middle line:84% here, you don't need to come and change it here as well. 00:48:09.610 --> 00:48:13.330 align:middle line:84% AUDIENCE: All right, I thought the project depreciation's 00:48:13.330 --> 00:48:14.063 align:middle line:90% a project. 00:48:14.063 --> 00:48:15.230 align:middle line:90% HAZHIR RAHMANDAD: That's OK. 00:48:15.230 --> 00:48:17.830 align:middle line:84% You have a transformation of the units. 00:48:17.830 --> 00:48:21.250 align:middle line:90% So you need a unit change. 00:48:21.250 --> 00:48:27.340 align:middle line:84% AUDIENCE: I guess the novice fraction changes? 00:48:27.340 --> 00:48:31.100 align:middle line:84% So a project coming in, let's say, 00:48:31.100 --> 00:48:33.460 align:middle line:84% does not have the same [INAUDIBLE]. 00:48:33.460 --> 00:48:35.288 align:middle line:90% So I can't-- 00:48:35.288 --> 00:48:36.580 align:middle line:90% HAZHIR RAHMANDAD: Let me think. 00:48:36.580 --> 00:48:42.298 align:middle line:84% So you're tracking the quality as a-- 00:48:42.298 --> 00:48:43.340 align:middle line:90% AUDIENCE: As [INAUDIBLE]. 00:48:43.340 --> 00:48:45.680 align:middle line:84% HAZHIR RAHMANDAD: As another co-flow, basically. 00:48:45.680 --> 00:48:50.500 align:middle line:90% 00:48:50.500 --> 00:48:51.340 align:middle line:90% It works still. 00:48:51.340 --> 00:48:55.980 align:middle line:84% The way you have to do it is to calculate the average quality. 00:48:55.980 --> 00:48:58.220 align:middle line:84% So the average quality would be the total quality 00:48:58.220 --> 00:49:05.020 align:middle line:84% divided by the total number of projects that are fulfilled. 00:49:05.020 --> 00:49:08.000 align:middle line:84% And then you multiply that back into the outflow. 00:49:08.000 --> 00:49:11.380 align:middle line:84% So this is really just a co-flow. 00:49:11.380 --> 00:49:13.940 align:middle line:84% AUDIENCE: Why is it important in this closed setup 00:49:13.940 --> 00:49:21.510 align:middle line:84% that Russ has that quality, the memory of quality, 00:49:21.510 --> 00:49:24.752 align:middle line:90% co-flow with demand? 00:49:24.752 --> 00:49:27.210 align:middle line:84% HAZHIR RAHMANDAD: So this is actually a definitional issue. 00:49:27.210 --> 00:49:29.970 align:middle line:84% So, actually, here, fulfilled demand, 00:49:29.970 --> 00:49:33.790 align:middle line:84% the way Ross is defining it is basically the period over which 00:49:33.790 --> 00:49:36.030 align:middle line:84% these projects that were completed 00:49:36.030 --> 00:49:40.690 align:middle line:84% have an impact in the memory of the people in the field. 00:49:40.690 --> 00:49:43.030 align:middle line:84% So this is exactly the purpose of having 00:49:43.030 --> 00:49:46.130 align:middle line:84% this stock because you have-- as we discussed at the beginning, 00:49:46.130 --> 00:49:48.050 align:middle line:84% there is another stock, potential demand. 00:49:48.050 --> 00:49:49.430 align:middle line:84% These are also people who are not 00:49:49.430 --> 00:49:53.070 align:middle line:84% being served and are waiting to get their project into-- 00:49:53.070 --> 00:49:55.158 align:middle line:90% to get a new project going. 00:49:55.158 --> 00:49:56.950 align:middle line:84% But the difference between these two stocks 00:49:56.950 --> 00:49:59.470 align:middle line:84% is that the first one, fulfill demand, the memory 00:49:59.470 --> 00:50:02.130 align:middle line:84% matters in terms of the reputation of the field. 00:50:02.130 --> 00:50:05.010 align:middle line:84% The second one, the memory is gone. 00:50:05.010 --> 00:50:10.430 align:middle line:84% So I think, in that sense, it is doing a meaningful thing here. 00:50:10.430 --> 00:50:12.172 align:middle line:90% AUDIENCE: That makes sense. 00:50:12.172 --> 00:50:13.130 align:middle line:90% HAZHIR RAHMANDAD: Good. 00:50:13.130 --> 00:50:16.450 align:middle line:84% I think we touched on key issues here. 00:50:16.450 --> 00:50:17.870 align:middle line:90% Anything that you wanted? 00:50:17.870 --> 00:50:18.810 align:middle line:90% Oh, reputation. 00:50:18.810 --> 00:50:21.320 align:middle line:90% 00:50:21.320 --> 00:50:24.160 align:middle line:84% So reputation and recognition, we 00:50:24.160 --> 00:50:31.540 align:middle line:84% have fulfilled demand with quality of fulfilled demand, 00:50:31.540 --> 00:50:33.740 align:middle line:84% giving the average quality of fulfilled demand. 00:50:33.740 --> 00:50:36.920 align:middle line:84% So this is kind of that average quality we talked about. 00:50:36.920 --> 00:50:43.320 align:middle line:84% And there is a quality gap that-- 00:50:43.320 --> 00:50:46.592 align:middle line:90% 00:50:46.592 --> 00:50:51.780 align:middle line:84% so, basically, this is kind of a smooth function, 00:50:51.780 --> 00:50:54.080 align:middle line:84% except that it is asymmetric for the reasons 00:50:54.080 --> 00:50:55.060 align:middle line:90% that we have discussed. 00:50:55.060 --> 00:50:58.540 align:middle line:84% Makes sense to have asymmetric view. 00:50:58.540 --> 00:51:00.680 align:middle line:84% And it is kind of adjusting quality 00:51:00.680 --> 00:51:04.300 align:middle line:84% to-- reputation towards quality, asymmetric. 00:51:04.300 --> 00:51:05.740 align:middle line:90% That's largely fine. 00:51:05.740 --> 00:51:09.440 align:middle line:90% I think it should do its-- 00:51:09.440 --> 00:51:15.920 align:middle line:84% and recognition is saying something about this demand 00:51:15.920 --> 00:51:20.010 align:middle line:84% fulfillment ratio, which we talked about 00:51:20.010 --> 00:51:23.470 align:middle line:84% as being a little bit hard, complex concept, 00:51:23.470 --> 00:51:26.430 align:middle line:84% not fully capturing what you necessarily want. 00:51:26.430 --> 00:51:29.170 align:middle line:90% 00:51:29.170 --> 00:51:31.370 align:middle line:84% But, ultimately, this is anchored 00:51:31.370 --> 00:51:33.930 align:middle line:90% towards a number around-- 00:51:33.930 --> 00:51:36.010 align:middle line:90% in between 0 and 1. 00:51:36.010 --> 00:51:38.030 align:middle line:84% And this sensitivity, what does it do? 00:51:38.030 --> 00:51:40.010 align:middle line:90% It kind of is-- 00:51:40.010 --> 00:51:42.510 align:middle line:90% yeah, it's a power exponent. 00:51:42.510 --> 00:51:46.730 align:middle line:90% 00:51:46.730 --> 00:51:51.570 align:middle line:84% So a higher fulfillment ratio means 00:51:51.570 --> 00:51:54.650 align:middle line:84% there is more projects around that 00:51:54.650 --> 00:51:57.970 align:middle line:90% are in the memory of people. 00:51:57.970 --> 00:52:01.330 align:middle line:84% Now, those more projects around-- 00:52:01.330 --> 00:52:03.850 align:middle line:84% and then there is a memory loss that 00:52:03.850 --> 00:52:06.850 align:middle line:84% goes with time to depreciate that you're using. 00:52:06.850 --> 00:52:08.630 align:middle line:90% But you need to be careful here. 00:52:08.630 --> 00:52:11.490 align:middle line:84% Here, you're using a first order delay. 00:52:11.490 --> 00:52:14.390 align:middle line:84% And down here, you're using a higher order delay. 00:52:14.390 --> 00:52:16.900 align:middle line:84% So that's why you need to pick one of them 00:52:16.900 --> 00:52:19.480 align:middle line:84% and then use that to drive all of the outflows. 00:52:19.480 --> 00:52:22.700 align:middle line:84% Otherwise, the parallelism breaks down. 00:52:22.700 --> 00:52:25.420 align:middle line:84% So here, the recognition memory loss 00:52:25.420 --> 00:52:29.100 align:middle line:84% should come from project depreciation with whatever unit 00:52:29.100 --> 00:52:31.360 align:middle line:90% adjustment you may need. 00:52:31.360 --> 00:52:34.100 align:middle line:90% 00:52:34.100 --> 00:52:38.520 align:middle line:84% Although, it's also somewhat asymmetric here as well. 00:52:38.520 --> 00:52:44.800 align:middle line:84% So I'm not fully following what exactly is happening here. 00:52:44.800 --> 00:52:47.900 align:middle line:84% So you are adjusting the recognition 00:52:47.900 --> 00:52:51.320 align:middle line:84% towards indicated recognition with this time constant. 00:52:51.320 --> 00:52:56.320 align:middle line:84% But also, you're eroding the recognition. 00:52:56.320 --> 00:52:58.700 align:middle line:84% So in a steady state, this stock will always 00:52:58.700 --> 00:53:02.100 align:middle line:84% be below the indicated recognition. 00:53:02.100 --> 00:53:04.420 align:middle line:90% Is that what you wanted to do? 00:53:04.420 --> 00:53:06.860 align:middle line:84% AUDIENCE: Let's think of a more extreme case, I guess. 00:53:06.860 --> 00:53:09.060 align:middle line:84% If there's no any more project, people 00:53:09.060 --> 00:53:11.660 align:middle line:84% are going to start forgetting the recognition is 00:53:11.660 --> 00:53:12.380 align:middle line:90% going to go down. 00:53:12.380 --> 00:53:13.838 align:middle line:84% HAZHIR RAHMANDAD: Can you speak up? 00:53:13.838 --> 00:53:16.820 align:middle line:84% AUDIENCE: Oh, if there's no project anymore, 00:53:16.820 --> 00:53:18.800 align:middle line:84% then people are going to forget it over time. 00:53:18.800 --> 00:53:21.668 align:middle line:84% The recognition is going to go down? 00:53:21.668 --> 00:53:23.960 align:middle line:84% HAZHIR RAHMANDAD: Yes, but then if there is no project, 00:53:23.960 --> 00:53:26.520 align:middle line:84% your demand fulfillment ratio becomes close to zero. 00:53:26.520 --> 00:53:28.942 align:middle line:84% So your indicated recognition goes very low. 00:53:28.942 --> 00:53:30.900 align:middle line:84% AUDIENCE: I guess that's the confusion because, 00:53:30.900 --> 00:53:33.800 align:middle line:84% fulfilled demand, all sorts of depreciation. 00:53:33.800 --> 00:53:38.220 align:middle line:84% So I guess I didn't need that, recognition memory loss. 00:53:38.220 --> 00:53:40.700 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so this is creating-- 00:53:40.700 --> 00:53:45.180 align:middle line:84% this is messy because you had two different types 00:53:45.180 --> 00:53:46.320 align:middle line:90% of formulations here. 00:53:46.320 --> 00:53:47.260 align:middle line:90% One is a goal-- 00:53:47.260 --> 00:53:50.520 align:middle line:84% kind of going towards an indicated value. 00:53:50.520 --> 00:53:53.620 align:middle line:84% This is just a smooth function, pretty standard. 00:53:53.620 --> 00:53:56.380 align:middle line:90% That creates an anchor there. 00:53:56.380 --> 00:54:01.100 align:middle line:84% Then you have this recognition lasting, 00:54:01.100 --> 00:54:03.400 align:middle line:90% which is just an outflow. 00:54:03.400 --> 00:54:06.820 align:middle line:90% It's just getting things-- 00:54:06.820 --> 00:54:09.800 align:middle line:84% basically, there was nothing else. 00:54:09.800 --> 00:54:11.100 align:middle line:90% This will go to 0. 00:54:11.100 --> 00:54:13.790 align:middle line:90% So that creates another anchor. 00:54:13.790 --> 00:54:17.470 align:middle line:84% And this stock is kind of lost between these two. 00:54:17.470 --> 00:54:19.670 align:middle line:84% And the reason for why it is lost between the two 00:54:19.670 --> 00:54:20.710 align:middle line:90% is not clear. 00:54:20.710 --> 00:54:25.450 align:middle line:84% So that's why I don't necessarily like that approach. 00:54:25.450 --> 00:54:28.110 align:middle line:84% Now, let's say it was moving towards the indicated 00:54:28.110 --> 00:54:29.130 align:middle line:90% recognition. 00:54:29.130 --> 00:54:32.330 align:middle line:84% So now, what the field attractiveness does, what is it? 00:54:32.330 --> 00:54:34.790 align:middle line:90% You're multiplying the two. 00:54:34.790 --> 00:54:36.850 align:middle line:84% These are kind of apples and oranges. 00:54:36.850 --> 00:54:40.070 align:middle line:84% So one is saying, how many projects 00:54:40.070 --> 00:54:43.990 align:middle line:84% have been happening in some way, that fulfillment ratio thingy? 00:54:43.990 --> 00:54:47.310 align:middle line:84% It's basically about the scale of the impact. 00:54:47.310 --> 00:54:50.690 align:middle line:84% This one is about how good those things are. 00:54:50.690 --> 00:54:52.710 align:middle line:84% Now, putting aside all the challenges 00:54:52.710 --> 00:54:55.870 align:middle line:84% with both of the formulations, multiplying 00:54:55.870 --> 00:54:58.510 align:middle line:84% how good something is with how big something is 00:54:58.510 --> 00:55:03.430 align:middle line:84% doesn't necessarily give you a good single metric. 00:55:03.430 --> 00:55:07.990 align:middle line:84% So if the reputation is bad, let's say, 00:55:07.990 --> 00:55:11.440 align:middle line:84% knowing that-- a lot of people knowing about it 00:55:11.440 --> 00:55:17.000 align:middle line:84% should actually make it go down compared to come up. 00:55:17.000 --> 00:55:19.500 align:middle line:84% Whereas, if the reputation is good, 00:55:19.500 --> 00:55:24.480 align:middle line:84% a lot of people knowing about it should speed up the adoption. 00:55:24.480 --> 00:55:28.280 align:middle line:84% So in a way, the valence of adoption 00:55:28.280 --> 00:55:30.360 align:middle line:84% should come from the reputation, whether it 00:55:30.360 --> 00:55:32.280 align:middle line:90% is going up or coming down. 00:55:32.280 --> 00:55:35.360 align:middle line:84% The speed of adoption, the adjustment time, 00:55:35.360 --> 00:55:39.760 align:middle line:84% could come from this recognition from this size. 00:55:39.760 --> 00:55:44.040 align:middle line:84% So they should have two different types of impacts 00:55:44.040 --> 00:55:46.000 align:middle line:84% rather than lumping them together 00:55:46.000 --> 00:55:48.360 align:middle line:90% to this single attractiveness. 00:55:48.360 --> 00:55:51.600 align:middle line:90% Does that make sense? 00:55:51.600 --> 00:55:52.180 align:middle line:90% Great. 00:55:52.180 --> 00:55:54.040 align:middle line:84% So we spent quite a bit on this, but I 00:55:54.040 --> 00:55:57.000 align:middle line:84% think it was a good conversation because a lot of these equations 00:55:57.000 --> 00:55:57.620 align:middle line:90% come out. 00:55:57.620 --> 00:55:59.820 align:middle line:84% AUDIENCE: Just really curious, if there's some time, 00:55:59.820 --> 00:56:02.440 align:middle line:84% can you remind me how you did quality, 00:56:02.440 --> 00:56:03.822 align:middle line:90% how you formulated quality? 00:56:03.822 --> 00:56:05.280 align:middle line:84% HAZHIR RAHMANDAD: So the quality is 00:56:05.280 --> 00:56:12.370 align:middle line:84% coming from the ratio of novice versus experts here. 00:56:12.370 --> 00:56:14.970 align:middle line:84% There are kind of issues to tackle here 00:56:14.970 --> 00:56:19.210 align:middle line:84% because it is not just the novice versus expert fraction 00:56:19.210 --> 00:56:20.330 align:middle line:90% that matters. 00:56:20.330 --> 00:56:24.810 align:middle line:84% It is also about how much of the time of experts 00:56:24.810 --> 00:56:27.810 align:middle line:84% is spent on training novices versus how much of the time 00:56:27.810 --> 00:56:29.530 align:middle line:90% is on doing the projects. 00:56:29.530 --> 00:56:32.930 align:middle line:84% So the time on doing the project by novice versus the time 00:56:32.930 --> 00:56:34.890 align:middle line:84% on doing the project by experts is 00:56:34.890 --> 00:56:39.150 align:middle line:84% what is kind of really driving it, not just the two numbers. 00:56:39.150 --> 00:56:41.670 align:middle line:84% But as a first order approximation, this is fine. 00:56:41.670 --> 00:56:43.690 align:middle line:90% AUDIENCE: I just wanted to see. 00:56:43.690 --> 00:56:45.730 align:middle line:90% HAZHIR RAHMANDAD: Yeah, OK. 00:56:45.730 --> 00:56:49.830 align:middle line:90% Let's switch to-- let's see. 00:56:49.830 --> 00:56:50.823 align:middle line:90% What's this one? 00:56:50.823 --> 00:56:51.490 align:middle line:90% This is Jason's. 00:56:51.490 --> 00:56:57.330 align:middle line:90% 00:56:57.330 --> 00:57:01.950 align:middle line:84% Let's get about-- get started with the projects. 00:57:01.950 --> 00:57:04.010 align:middle line:90% So you have? 00:57:04.010 --> 00:57:05.802 align:middle line:84% AUDIENCE: So when you're regeneration? 00:57:05.802 --> 00:57:07.260 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so you have 00:57:07.260 --> 00:57:10.680 align:middle line:84% people coming into the system with a fixed rate. 00:57:10.680 --> 00:57:14.660 align:middle line:84% This is kind of people who have new projects that 00:57:14.660 --> 00:57:16.980 align:middle line:90% might be relevant. 00:57:16.980 --> 00:57:19.160 align:middle line:84% And you're thinking of these as people. 00:57:19.160 --> 00:57:21.340 align:middle line:84% So these are as clients, basically. 00:57:21.340 --> 00:57:25.620 align:middle line:84% So we have new clients coming in from new use cases in the world. 00:57:25.620 --> 00:57:29.540 align:middle line:90% Some of them learn about it. 00:57:29.540 --> 00:57:36.260 align:middle line:84% Then with a delay, they decide to kind of go and get 00:57:36.260 --> 00:57:40.220 align:middle line:84% their project tackled by this field. 00:57:40.220 --> 00:57:41.620 align:middle line:90% AUDIENCE: Or someone else. 00:57:41.620 --> 00:57:42.520 align:middle line:84% HAZHIR RAHMANDAD: Or someone else. 00:57:42.520 --> 00:57:44.200 align:middle line:84% So there is kind of a finding solution. 00:57:44.200 --> 00:57:47.440 align:middle line:84% There is a fraction, that they become clients. 00:57:47.440 --> 00:57:49.820 align:middle line:90% That makes sense. 00:57:49.820 --> 00:57:53.320 align:middle line:84% Then these projects are started and finished. 00:57:53.320 --> 00:57:55.340 align:middle line:90% And so we can look at those. 00:57:55.340 --> 00:58:00.680 align:middle line:84% And then project delivery impacts recognition. 00:58:00.680 --> 00:58:02.100 align:middle line:90% There is a delay. 00:58:02.100 --> 00:58:04.670 align:middle line:84% There is a kind of a fading of recognition. 00:58:04.670 --> 00:58:07.950 align:middle line:84% So this is similar idea, though this one 00:58:07.950 --> 00:58:13.050 align:middle line:84% is about the scale of how people know about it. 00:58:13.050 --> 00:58:17.930 align:middle line:84% Then you have, also, the recognition of the alternatives. 00:58:17.930 --> 00:58:19.010 align:middle line:90% So that's fine. 00:58:19.010 --> 00:58:22.230 align:middle line:90% 00:58:22.230 --> 00:58:30.190 align:middle line:84% And then the backbone makes sense to me. 00:58:30.190 --> 00:58:31.735 align:middle line:84% Once projects are done, they're done. 00:58:31.735 --> 00:58:33.110 align:middle line:84% So in a way, this is a little bit 00:58:33.110 --> 00:58:34.730 align:middle line:84% of a different conceptualization, 00:58:34.730 --> 00:58:37.290 align:middle line:84% even though it's not that far from what Ross had. 00:58:37.290 --> 00:58:38.870 align:middle line:84% But this one is saying, there are 00:58:38.870 --> 00:58:41.990 align:middle line:84% kind of flow of projects coming into the system. 00:58:41.990 --> 00:58:45.170 align:middle line:84% They get allocated between these two types. 00:58:45.170 --> 00:58:49.710 align:middle line:84% And then we don't care about them later, once they're out. 00:58:49.710 --> 00:58:53.050 align:middle line:84% They just have a recognition reputation for a while. 00:58:53.050 --> 00:58:55.950 align:middle line:90% 00:58:55.950 --> 00:59:00.790 align:middle line:84% Here, another addressable market. 00:59:00.790 --> 00:59:09.670 align:middle line:84% Time to become aware depends on awareness penetration. 00:59:09.670 --> 00:59:14.570 align:middle line:90% 00:59:14.570 --> 00:59:17.110 align:middle line:90% So let's look at this time. 00:59:17.110 --> 00:59:21.950 align:middle line:84% So it has an effect of unaware fraction on time to become. 00:59:21.950 --> 00:59:24.750 align:middle line:84% And then what is unaware fraction? 00:59:24.750 --> 00:59:29.070 align:middle line:84% Fraction of population that is unaware, search time. 00:59:29.070 --> 00:59:32.690 align:middle line:84% So the bigger the fraction of population that is unaware, 00:59:32.690 --> 00:59:36.810 align:middle line:84% the slower is the time to become aware. 00:59:36.810 --> 00:59:38.570 align:middle line:90% Is that the idea? 00:59:38.570 --> 00:59:44.890 align:middle line:84% AUDIENCE: The bigger the fraction is, 00:59:44.890 --> 00:59:47.430 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so you're dividing-- 00:59:47.430 --> 00:59:51.590 align:middle line:84% so it's 1 over the fraction of the population that is unaware. 00:59:51.590 --> 00:59:56.243 align:middle line:84% So the larger that fraction, the smaller this. 00:59:56.243 --> 00:59:57.410 align:middle line:90% AUDIENCE: Yes, that's right. 00:59:57.410 --> 01:00:00.920 align:middle line:84% HAZHIR RAHMANDAD: And, therefore, the smaller this is, 01:00:00.920 --> 01:00:06.000 align:middle line:84% the smaller it's time to become a verb, which means the faster 01:00:06.000 --> 01:00:07.402 align:middle line:90% that people become aware. 01:00:07.402 --> 01:00:08.360 align:middle line:90% AUDIENCE: That's right. 01:00:08.360 --> 01:00:10.193 align:middle line:84% HAZHIR RAHMANDAD: So what's the logic there? 01:00:10.193 --> 01:00:16.320 align:middle line:84% AUDIENCE: Logic is, it's harder to-- 01:00:16.320 --> 01:00:20.560 align:middle line:90% 01:00:20.560 --> 01:00:29.480 align:middle line:84% if you have a large fraction of people who are unaware, 01:00:29.480 --> 01:00:32.800 align:middle line:84% in that situation, the people right at the margin 01:00:32.800 --> 01:00:34.520 align:middle line:84% are going to find it pretty quickly. 01:00:34.520 --> 01:00:39.200 align:middle line:84% But as that fraction decreases, those people 01:00:39.200 --> 01:00:43.480 align:middle line:84% were harder to find, or harder to connect or recognize 01:00:43.480 --> 01:00:45.760 align:middle line:90% the solution. 01:00:45.760 --> 01:00:46.860 align:middle line:90% So it takes them longer. 01:00:46.860 --> 01:00:50.280 align:middle line:84% It's like it's going to improve-- 01:00:50.280 --> 01:00:53.880 align:middle line:84% HAZHIR RAHMANDAD: There is a logic there 01:00:53.880 --> 01:00:57.160 align:middle line:84% about low-hanging fruits, but I don't see it 01:00:57.160 --> 01:00:58.885 align:middle line:90% as a first order need here. 01:00:58.885 --> 01:01:01.010 align:middle line:84% And the reason I don't see it as a first order need 01:01:01.010 --> 01:01:03.410 align:middle line:84% is that you already have that feedback directly 01:01:03.410 --> 01:01:07.410 align:middle line:84% in this link from another addressable to becoming aware. 01:01:07.410 --> 01:01:09.390 align:middle line:84% If you don't have another addressable, 01:01:09.390 --> 01:01:11.090 align:middle line:90% nothing will happen here. 01:01:11.090 --> 01:01:13.130 align:middle line:84% So you have a first order effect that 01:01:13.130 --> 01:01:15.630 align:middle line:84% just says, if there are not many unaware people, 01:01:15.630 --> 01:01:18.370 align:middle line:84% you wouldn't have newer projects. 01:01:18.370 --> 01:01:20.730 align:middle line:84% So changing the time constant means 01:01:20.730 --> 01:01:23.690 align:middle line:84% the fractional rate of becoming aware 01:01:23.690 --> 01:01:28.267 align:middle line:84% is higher when fewer-- when more people don't know about it. 01:01:28.267 --> 01:01:28.850 align:middle line:90% AUDIENCE: Yes. 01:01:28.850 --> 01:01:32.490 align:middle line:84% HAZHIR RAHMANDAD: That is a little bit 01:01:32.490 --> 01:01:34.750 align:middle line:90% of an iffy proposition. 01:01:34.750 --> 01:01:38.490 align:middle line:84% I'm not buying that just because you say that. 01:01:38.490 --> 01:01:44.850 align:middle line:84% I need some evidence to buy that proposition. 01:01:44.850 --> 01:01:49.930 align:middle line:84% But it's not going to do a lot here. 01:01:49.930 --> 01:01:52.770 align:middle line:84% Although, it actually can do quite a bit because, right now, 01:01:52.770 --> 01:01:54.330 align:middle line:84% with this sensitivity of 1, you're 01:01:54.330 --> 01:01:57.140 align:middle line:84% basically making the time constant to be 01:01:57.140 --> 01:02:00.440 align:middle line:84% 1 over the unaware addressable fraction. 01:02:00.440 --> 01:02:03.040 align:middle line:84% So it is unaware divided by total population. 01:02:03.040 --> 01:02:05.840 align:middle line:84% So when you then divide the time by time constant, 01:02:05.840 --> 01:02:08.340 align:middle line:84% the equation here becomes unaware to the power 01:02:08.340 --> 01:02:14.940 align:middle line:84% of 2 divided by this minimum time divided 01:02:14.940 --> 01:02:17.420 align:middle line:90% by total population. 01:02:17.420 --> 01:02:19.180 align:middle line:90% I don't know. 01:02:19.180 --> 01:02:21.380 align:middle line:84% AUDIENCE: Total population is a constant, yes? 01:02:21.380 --> 01:02:22.797 align:middle line:84% HAZHIR RAHMANDAD: Very good point. 01:02:22.797 --> 01:02:25.020 align:middle line:84% That's the next thing, that the total population 01:02:25.020 --> 01:02:27.700 align:middle line:90% is right now a constant. 01:02:27.700 --> 01:02:31.102 align:middle line:84% But total population, what is it actually? 01:02:31.102 --> 01:02:32.060 align:middle line:90% AUDIENCE: You're right. 01:02:32.060 --> 01:02:32.900 align:middle line:90% HAZHIR RAHMANDAD: It's like-- 01:02:32.900 --> 01:02:33.580 align:middle line:84% AUDIENCE: It's not doing anything. 01:02:33.580 --> 01:02:35.360 align:middle line:84% HAZHIR RAHMANDAD: I don't know what it is trying to do. 01:02:35.360 --> 01:02:37.620 align:middle line:84% No, the point is that, if you're talking about all 01:02:37.620 --> 01:02:39.960 align:middle line:84% of the people who are in one of these three states, 01:02:39.960 --> 01:02:43.820 align:middle line:84% or maybe perhaps even in the recognition state, 01:02:43.820 --> 01:02:46.600 align:middle line:84% you should track that explicitly by adding those people. 01:02:46.600 --> 01:02:48.540 align:middle line:84% If you mean something else, then you 01:02:48.540 --> 01:02:51.220 align:middle line:84% need to be explicit why that something else is important 01:02:51.220 --> 01:02:53.060 align:middle line:90% here. 01:02:53.060 --> 01:02:57.190 align:middle line:84% It's not very clean here for the way it is working. 01:02:57.190 --> 01:03:01.370 align:middle line:84% But then we have a time to find solutions. 01:03:01.370 --> 01:03:05.030 align:middle line:84% And then they choose which fraction 01:03:05.030 --> 01:03:10.210 align:middle line:84% becomes interested in your field versus the alternative field. 01:03:10.210 --> 01:03:13.030 align:middle line:90% And that comes from the quality. 01:03:13.030 --> 01:03:14.648 align:middle line:90% So there is a relative quality. 01:03:14.648 --> 01:03:16.190 align:middle line:84% You have a reference quality, and you 01:03:16.190 --> 01:03:17.450 align:middle line:90% have a perceived quality. 01:03:17.450 --> 01:03:20.550 align:middle line:84% So I assume reference quality is for the competitive field you're 01:03:20.550 --> 01:03:21.570 align:middle line:90% compared to. 01:03:21.570 --> 01:03:24.070 align:middle line:90% AUDIENCE: You [INAUDIBLE]. 01:03:24.070 --> 01:03:29.950 align:middle line:84% I thought about it as, what is the quality when experience 01:03:29.950 --> 01:03:31.490 align:middle line:90% is at its reference level? 01:03:31.490 --> 01:03:33.910 align:middle line:84% HAZHIR RAHMANDAD: So that's very important difference. 01:03:33.910 --> 01:03:37.650 align:middle line:84% So I don't want to interpret it in a different way. 01:03:37.650 --> 01:03:38.430 align:middle line:90% AUDIENCE: So I-- 01:03:38.430 --> 01:03:39.550 align:middle line:84% HAZHIR RAHMANDAD: Because right now, 01:03:39.550 --> 01:03:41.290 align:middle line:84% the decision that is happening here-- 01:03:41.290 --> 01:03:43.110 align:middle line:90% fraction who become clients-- 01:03:43.110 --> 01:03:45.530 align:middle line:84% is determining, what fraction go to the alternatives? 01:03:45.530 --> 01:03:47.630 align:middle line:84% And what fractions come to us, correct? 01:03:47.630 --> 01:03:48.230 align:middle line:90% AUDIENCE: Yes. 01:03:48.230 --> 01:03:50.210 align:middle line:84% HAZHIR RAHMANDAD: So if that is the definition, 01:03:50.210 --> 01:03:55.640 align:middle line:84% then this should be driven by the relative quality of us 01:03:55.640 --> 01:03:57.400 align:middle line:90% versus the alternative. 01:03:57.400 --> 01:04:01.800 align:middle line:84% So if your reference alternative is defined as them, 01:04:01.800 --> 01:04:02.880 align:middle line:90% then it works. 01:04:02.880 --> 01:04:05.800 align:middle line:84% If it is defined as r maximum quality or something 01:04:05.800 --> 01:04:08.600 align:middle line:84% along those lines, you are double-- 01:04:08.600 --> 01:04:11.660 align:middle line:84% using the same variable for two different reasons. 01:04:11.660 --> 01:04:15.440 align:middle line:84% And that's going to be problematic. 01:04:15.440 --> 01:04:18.040 align:middle line:84% AUDIENCE: That makes sense I didn't think that. 01:04:18.040 --> 01:04:19.560 align:middle line:90% HAZHIR RAHMANDAD: Good. 01:04:19.560 --> 01:04:22.880 align:middle line:84% Then with that in mind, we have relative quality coming 01:04:22.880 --> 01:04:24.460 align:middle line:90% from your perceived quality. 01:04:24.460 --> 01:04:29.360 align:middle line:84% Your perceived quality has a symmetry in its adjustment. 01:04:29.360 --> 01:04:31.200 align:middle line:90% It comes from indicated quality. 01:04:31.200 --> 01:04:34.160 align:middle line:84% That comes from your relative experience. 01:04:34.160 --> 01:04:35.700 align:middle line:90% And it has its own parameters. 01:04:35.700 --> 01:04:39.080 align:middle line:84% So we have to see where the relative experience is 01:04:39.080 --> 01:04:41.560 align:middle line:90% generated. 01:04:41.560 --> 01:04:46.040 align:middle line:84% And this recognition, it was driving what-- 01:04:46.040 --> 01:04:49.660 align:middle line:84% AUDIENCE: So the recognition is driving the allocation from-- 01:04:49.660 --> 01:04:54.140 align:middle line:84% I thought about it as driving the fraction you choose. 01:04:54.140 --> 01:04:57.760 align:middle line:90% 01:04:57.760 --> 01:04:59.460 align:middle line:84% Oh, that's the awareness of penetration. 01:04:59.460 --> 01:05:00.240 align:middle line:90% I'm sorry. 01:05:00.240 --> 01:05:05.538 align:middle line:84% So the recognition, if you look at becoming aware-- 01:05:05.538 --> 01:05:06.580 align:middle line:90% HAZHIR RAHMANDAD: Oh, OK. 01:05:06.580 --> 01:05:09.280 align:middle line:84% It is coming partly from your recognition 01:05:09.280 --> 01:05:10.960 align:middle line:90% versus the other's recognition. 01:05:10.960 --> 01:05:12.400 align:middle line:90% AUDIENCE: So it takes longer. 01:05:12.400 --> 01:05:14.500 align:middle line:84% HAZHIR RAHMANDAD: But let's see, what does this-- 01:05:14.500 --> 01:05:17.760 align:middle line:84% AUDIENCE: It's just a fraction, so what share of recognition 01:05:17.760 --> 01:05:20.637 align:middle line:84% belongs to our field versus the other field? 01:05:20.637 --> 01:05:21.720 align:middle line:90% HAZHIR RAHMANDAD: Oh, yes. 01:05:21.720 --> 01:05:27.400 align:middle line:84% That is now a weird formulation here. 01:05:27.400 --> 01:05:33.400 align:middle line:84% So if we have a lot of recognition for the other field, 01:05:33.400 --> 01:05:39.880 align:middle line:84% this formulation shots off becoming aware. 01:05:39.880 --> 01:05:43.480 align:middle line:84% And when it shots off becoming aware, 01:05:43.480 --> 01:05:50.370 align:middle line:84% no field, neither ours nor theirs, will get any solutions, 01:05:50.370 --> 01:05:52.330 align:middle line:90% will get any customers. 01:05:52.330 --> 01:05:53.825 align:middle line:90% So that just doesn't make sense. 01:05:53.825 --> 01:05:54.450 align:middle line:90% AUDIENCE: Yeah. 01:05:54.450 --> 01:06:01.110 align:middle line:90% 01:06:01.110 --> 01:06:03.550 align:middle line:84% HAZHIR RAHMANDAD: Then we have projects. 01:06:03.550 --> 01:06:07.450 align:middle line:90% 01:06:07.450 --> 01:06:09.590 align:middle line:90% Sorry, we were up here. 01:06:09.590 --> 01:06:12.050 align:middle line:84% So perceived quality is coming from indicated quality, 01:06:12.050 --> 01:06:13.370 align:middle line:90% from relative experience. 01:06:13.370 --> 01:06:15.650 align:middle line:84% Relative experience is coming from total experience 01:06:15.650 --> 01:06:17.250 align:middle line:90% versus reference experience. 01:06:17.250 --> 01:06:21.350 align:middle line:84% Reference experience is just a fixed number. 01:06:21.350 --> 01:06:23.950 align:middle line:84% That's kind of a little bit of a red flag for me, 01:06:23.950 --> 01:06:26.790 align:middle line:84% but we should come back to it and see if it makes sense. 01:06:26.790 --> 01:06:28.830 align:middle line:84% So let's see, what is total experience? 01:06:28.830 --> 01:06:36.090 align:middle line:84% So total experience is the stock of experience 01:06:36.090 --> 01:06:39.870 align:middle line:84% that is a core flow for the experienced employees. 01:06:39.870 --> 01:06:40.468 align:middle line:90% AUDIENCE: Yep. 01:06:40.468 --> 01:06:42.510 align:middle line:84% HAZHIR RAHMANDAD: So that is stock of experience. 01:06:42.510 --> 01:06:44.810 align:middle line:84% So here, OK, let's actually look at this formulation 01:06:44.810 --> 01:06:45.830 align:middle line:90% more closely. 01:06:45.830 --> 01:06:48.775 align:middle line:90% So we have rookies join. 01:06:48.775 --> 01:06:50.400 align:middle line:84% There is a time to join, time to leave. 01:06:50.400 --> 01:06:51.780 align:middle line:90% There is an indicated rookie. 01:06:51.780 --> 01:07:01.500 align:middle line:84% Indicated rookie comes from feasible annual rookie per-- 01:07:01.500 --> 01:07:05.300 align:middle line:90% annual projects per rookie. 01:07:05.300 --> 01:07:13.380 align:middle line:84% That comes from rookie productivity, work hours. 01:07:13.380 --> 01:07:21.260 align:middle line:84% This is kind of calculating how many rookies-- how many projects 01:07:21.260 --> 01:07:24.500 align:middle line:90% each rookie can serve per year. 01:07:24.500 --> 01:07:27.380 align:middle line:84% And you look at the gap between the service capacity 01:07:27.380 --> 01:07:34.220 align:middle line:84% from expert capacity and rookie capacity versus the demand 01:07:34.220 --> 01:07:36.620 align:middle line:90% for desired fulfillment rate. 01:07:36.620 --> 01:07:38.000 align:middle line:90% So that does make sense. 01:07:38.000 --> 01:07:39.720 align:middle line:84% So you have just current projects, 01:07:39.720 --> 01:07:41.180 align:middle line:90% desired fulfillment time. 01:07:41.180 --> 01:07:44.470 align:middle line:84% That gives you a service gap, comparing these. 01:07:44.470 --> 01:07:46.830 align:middle line:84% AUDIENCE: And you can only bring in rookies. 01:07:46.830 --> 01:07:49.990 align:middle line:84% HAZHIR RAHMANDAD: So when that gap is positive, 01:07:49.990 --> 01:07:56.290 align:middle line:84% you get an indicated rookie that is positive. 01:07:56.290 --> 01:07:59.390 align:middle line:84% If it is negative, you get a negative rookie. 01:07:59.390 --> 01:08:04.510 align:middle line:84% Although, you have blocked it from being negative. 01:08:04.510 --> 01:08:09.650 align:middle line:84% So, therefore, indicated rookie cannot become negative. 01:08:09.650 --> 01:08:10.630 align:middle line:90% AUDIENCE: That's true. 01:08:10.630 --> 01:08:11.850 align:middle line:90% HAZHIR RAHMANDAD: But it could. 01:08:11.850 --> 01:08:12.510 align:middle line:90% AUDIENCE: Yeah, no. 01:08:12.510 --> 01:08:13.830 align:middle line:84% HAZHIR RAHMANDAD: Maybe even it should. 01:08:13.830 --> 01:08:14.510 align:middle line:90% AUDIENCE: Yeah. 01:08:14.510 --> 01:08:16.552 align:middle line:84% HAZHIR RAHMANDAD: Because you are using this flow 01:08:16.552 --> 01:08:19.470 align:middle line:84% as a two-way channel, you don't want 01:08:19.470 --> 01:08:23.189 align:middle line:84% to remove the signal that, ah, there is not really much work 01:08:23.189 --> 01:08:24.770 align:middle line:90% to be done in this field. 01:08:24.770 --> 01:08:29.630 align:middle line:84% And based on that, there is not much of a service gap 01:08:29.630 --> 01:08:30.609 align:middle line:90% to think about. 01:08:30.609 --> 01:08:36.270 align:middle line:84% Now, service gap here is accounting 01:08:36.270 --> 01:08:42.439 align:middle line:84% for how much total capacity for service 01:08:42.439 --> 01:08:46.880 align:middle line:84% we have versus how much total demand we have. 01:08:46.880 --> 01:08:48.500 align:middle line:90% And that creates-- 01:08:48.500 --> 01:08:53.600 align:middle line:90% 01:08:53.600 --> 01:08:57.479 align:middle line:84% So that shouldn't be creating your indicated rookie. 01:08:57.479 --> 01:09:05.000 align:middle line:84% That what it should be creating is the indicated extra rookies 01:09:05.000 --> 01:09:07.545 align:middle line:84% you want to have or, yeah, extra rookies. 01:09:07.545 --> 01:09:08.920 align:middle line:84% AUDIENCE: Very incremental, yeah. 01:09:08.920 --> 01:09:10.760 align:middle line:84% HAZHIR RAHMANDAD: That is, this should not 01:09:10.760 --> 01:09:14.140 align:middle line:84% be compared with rookie employees down here. 01:09:14.140 --> 01:09:15.473 align:middle line:90% Ooh, my God. 01:09:15.473 --> 01:09:16.640 align:middle line:90% AUDIENCE: It's not that bad. 01:09:16.640 --> 01:09:18.640 align:middle line:84% HAZHIR RAHMANDAD: Indicated rookies 01:09:18.640 --> 01:09:21.040 align:middle line:84% are bigger than rookie employees. 01:09:21.040 --> 01:09:23.740 align:middle line:84% Then it is indicated rookies minus rookie employees. 01:09:23.740 --> 01:09:25.537 align:middle line:84% So this is just the time constant. 01:09:25.537 --> 01:09:28.120 align:middle line:84% You could use the if-then-else in the denominator for the time 01:09:28.120 --> 01:09:31.520 align:middle line:90% constant you use. 01:09:31.520 --> 01:09:35.319 align:middle line:84% But you have assimilation time, so I'm not sure. 01:09:35.319 --> 01:09:39.200 align:middle line:84% But for now, either way, you are comparing indicated rookies 01:09:39.200 --> 01:09:40.439 align:middle line:90% with rookie employees. 01:09:40.439 --> 01:09:41.370 align:middle line:90% AUDIENCE: Yeah. 01:09:41.370 --> 01:09:44.550 align:middle line:84% HAZHIR RAHMANDAD: So if my indicated rookie was a hundred, 01:09:44.550 --> 01:09:45.770 align:middle line:90% what does it mean? 01:09:45.770 --> 01:09:52.229 align:middle line:84% It means I'm a hundred rookies short of the existing demand. 01:09:52.229 --> 01:09:54.890 align:middle line:84% So I needed to have a hundred rookie more compared 01:09:54.890 --> 01:09:57.865 align:middle line:84% to what I currently have to be able to serve the demand. 01:09:57.865 --> 01:09:58.490 align:middle line:90% AUDIENCE: Yeah. 01:09:58.490 --> 01:10:00.170 align:middle line:84% HAZHIR RAHMANDAD: The way you are doing it here, 01:10:00.170 --> 01:10:02.070 align:middle line:84% you're taking all of the existing rookies-- 01:10:02.070 --> 01:10:04.050 align:middle line:84% you may have a thousand rookies-- 01:10:04.050 --> 01:10:06.030 align:middle line:84% and you're indicated as a hundred. 01:10:06.030 --> 01:10:07.830 align:middle line:90% So this gives you minus 900. 01:10:07.830 --> 01:10:10.610 align:middle line:84% So it says, let me fire these rookies 01:10:10.610 --> 01:10:13.182 align:middle line:84% that I have because I just need a hundred. 01:10:13.182 --> 01:10:14.390 align:middle line:90% No, you don't need a hundred. 01:10:14.390 --> 01:10:16.267 align:middle line:90% You need a thousand and 100. 01:10:16.267 --> 01:10:17.350 align:middle line:90% So that's what I'm saying. 01:10:17.350 --> 01:10:22.850 align:middle line:84% It is kind of a first order problem that is missing here. 01:10:22.850 --> 01:10:28.710 align:middle line:90% So rookies are now adopt, join. 01:10:28.710 --> 01:10:30.630 align:middle line:84% They assimilate with an assimilation time. 01:10:30.630 --> 01:10:32.790 align:middle line:84% So this is the simplest formulation. 01:10:32.790 --> 01:10:35.070 align:middle line:84% It doesn't care whether they are working or not, 01:10:35.070 --> 01:10:39.420 align:middle line:84% doesn't care whether there are experts to help them or not. 01:10:39.420 --> 01:10:41.300 align:middle line:84% And then they become experienced. 01:10:41.300 --> 01:10:45.620 align:middle line:84% The experienced folks would leave with a time constant 01:10:45.620 --> 01:10:47.620 align:middle line:90% as well, just as-- 01:10:47.620 --> 01:10:51.300 align:middle line:84% well, rookies would not leave unless they are 01:10:51.300 --> 01:10:53.720 align:middle line:90% removed because of this gap. 01:10:53.720 --> 01:10:56.060 align:middle line:84% And if the gap was correctly formulated 01:10:56.060 --> 01:10:59.160 align:middle line:84% and could go negative, then it could work fine. 01:10:59.160 --> 01:11:02.880 align:middle line:84% Right now, it is not really doing what you want it to do. 01:11:02.880 --> 01:11:05.860 align:middle line:90% 01:11:05.860 --> 01:11:08.380 align:middle line:84% It doesn't get feedback right now from availability 01:11:08.380 --> 01:11:09.800 align:middle line:90% of projects and so on. 01:11:09.800 --> 01:11:11.860 align:middle line:84% So that's something to keep in mind. 01:11:11.860 --> 01:11:14.380 align:middle line:84% Then we have this co-flow of experience. 01:11:14.380 --> 01:11:20.500 align:middle line:84% So as you assimilate, you add more experience with-- 01:11:20.500 --> 01:11:23.560 align:middle line:84% AUDIENCE: That this was kind of ad hoc, 01:11:23.560 --> 01:11:27.460 align:middle line:84% but the thought was, if there isn't a project or work 01:11:27.460 --> 01:11:32.300 align:middle line:84% availability, the quality-- the amount of experience accrued 01:11:32.300 --> 01:11:33.760 align:middle line:90% goes down. 01:11:33.760 --> 01:11:37.020 align:middle line:84% HAZHIR RAHMANDAD: So this is actually important. 01:11:37.020 --> 01:11:38.820 align:middle line:84% You are doing two different things 01:11:38.820 --> 01:11:41.877 align:middle line:84% inside this flow, experience accrual from assimilation. 01:11:41.877 --> 01:11:42.460 align:middle line:90% AUDIENCE: Yes. 01:11:42.460 --> 01:11:45.960 align:middle line:84% HAZHIR RAHMANDAD: One is you just get new people assimilated. 01:11:45.960 --> 01:11:46.860 align:middle line:90% They come in. 01:11:46.860 --> 01:11:49.700 align:middle line:84% Each person who comes has some level of experience. 01:11:49.700 --> 01:11:51.780 align:middle line:84% As a rookie that has just become-- 01:11:51.780 --> 01:11:54.700 align:middle line:84% gets the level of experience, that is usually 01:11:54.700 --> 01:11:57.300 align:middle line:90% kind of a fixed number. 01:11:57.300 --> 01:11:59.900 align:middle line:84% AUDIENCE: I think it is working out to be a fixed number here. 01:11:59.900 --> 01:12:01.900 align:middle line:84% HAZHIR RAHMANDAD: Then I don't need four or five 01:12:01.900 --> 01:12:04.060 align:middle line:90% different things going into it. 01:12:04.060 --> 01:12:06.260 align:middle line:84% And it's probably not a fixed number 01:12:06.260 --> 01:12:08.382 align:middle line:90% because you have utilization. 01:12:08.382 --> 01:12:09.840 align:middle line:84% AUDIENCE: Without that utilization, 01:12:09.840 --> 01:12:12.340 align:middle line:90% it's because I'm multiplying-- 01:12:12.340 --> 01:12:14.800 align:middle line:84% HAZHIR RAHMANDAD: But you also have assimilation time. 01:12:14.800 --> 01:12:16.020 align:middle line:84% AUDIENCE: Minus the rate, so it could-- 01:12:16.020 --> 01:12:17.603 align:middle line:84% HAZHIR RAHMANDAD: So, conceptually, it 01:12:17.603 --> 01:12:20.340 align:middle line:84% is assimilation rate multiplied by experience 01:12:20.340 --> 01:12:22.200 align:middle line:90% per assimilating rookie. 01:12:22.200 --> 01:12:22.820 align:middle line:90% AUDIENCE: Yes. 01:12:22.820 --> 01:12:24.403 align:middle line:84% HAZHIR RAHMANDAD: Something like that. 01:12:24.403 --> 01:12:25.300 align:middle line:90% That's fine. 01:12:25.300 --> 01:12:27.020 align:middle line:90% You get that. 01:12:27.020 --> 01:12:30.720 align:middle line:84% Now, maybe you're calculating it this way. 01:12:30.720 --> 01:12:33.420 align:middle line:84% So you're saying, rookie effective work hours 01:12:33.420 --> 01:12:37.550 align:middle line:84% per unit of time multiplied by assimilation time 01:12:37.550 --> 01:12:40.807 align:middle line:84% is the total hours of experience that rookies come with it. 01:12:40.807 --> 01:12:41.890 align:middle line:90% So, as such, that is fine. 01:12:41.890 --> 01:12:44.710 align:middle line:84% Only, you don't need the utilization. 01:12:44.710 --> 01:12:47.790 align:middle line:84% Then once they go in, you are saying, 01:12:47.790 --> 01:12:52.590 align:middle line:84% if they are being utilized, they gain more experience. 01:12:52.590 --> 01:12:54.290 align:middle line:84% And if they are not being utilized, 01:12:54.290 --> 01:12:55.650 align:middle line:90% they don't gain more experience. 01:12:55.650 --> 01:12:58.310 align:middle line:84% Maybe that is what you had in mind. 01:12:58.310 --> 01:13:02.010 align:middle line:90% And that is a separate inflow. 01:13:02.010 --> 01:13:04.530 align:middle line:84% It is about inflow of people who are experienced, 01:13:04.530 --> 01:13:10.350 align:middle line:84% inflow of experience from just working, which you are capturing 01:13:10.350 --> 01:13:12.910 align:middle line:84% here from experience of working because experienced employees 01:13:12.910 --> 01:13:17.310 align:middle line:84% multiplied by effective work hours multiplied by utilization. 01:13:17.310 --> 01:13:18.430 align:middle line:90% That's good. 01:13:18.430 --> 01:13:19.490 align:middle line:90% You already have that. 01:13:19.490 --> 01:13:23.830 align:middle line:84% You don't need the utilization again in this beginning flow. 01:13:23.830 --> 01:13:26.170 align:middle line:84% And then you have experience loss, which is a standard. 01:13:26.170 --> 01:13:28.430 align:middle line:90% So that's good. 01:13:28.430 --> 01:13:32.160 align:middle line:84% So this total experience is what we had up here. 01:13:32.160 --> 01:13:33.660 align:middle line:84% AUDIENCE: Yeah, that drives quality. 01:13:33.660 --> 01:13:34.577 align:middle line:90% HAZHIR RAHMANDAD: Yes. 01:13:34.577 --> 01:13:37.880 align:middle line:84% Now, that total experience is being compared 01:13:37.880 --> 01:13:41.880 align:middle line:90% to this ad hoc reference thing. 01:13:41.880 --> 01:13:45.535 align:middle line:84% You just got from there some number and put it here. 01:13:45.535 --> 01:13:46.160 align:middle line:90% AUDIENCE: Yeah. 01:13:46.160 --> 01:13:47.493 align:middle line:90% HAZHIR RAHMANDAD: Don't do that. 01:13:47.493 --> 01:13:48.580 align:middle line:90% It's not going to work. 01:13:48.580 --> 01:13:51.020 align:middle line:90% It will create headache for you. 01:13:51.020 --> 01:13:53.600 align:middle line:84% It might be orders of magnitude off. 01:13:53.600 --> 01:13:58.040 align:middle line:84% It needs to be anchored to something about, what 01:13:58.040 --> 01:14:04.940 align:middle line:84% is the experience if everybody, for example, 01:14:04.940 --> 01:14:08.080 align:middle line:84% was a rookie or something, that-- 01:14:08.080 --> 01:14:12.120 align:middle line:84% AUDIENCE: I think a lot of what I'm realizing is because I 01:14:12.120 --> 01:14:12.760 align:middle line:90% hadn't-- 01:14:12.760 --> 01:14:19.000 align:middle line:84% I didn't have a clear story on the demand side about separating 01:14:19.000 --> 01:14:23.400 align:middle line:84% the demand and what the downstream implications of that 01:14:23.400 --> 01:14:25.620 align:middle line:90% were, it all got very muddled. 01:14:25.620 --> 01:14:27.920 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so I think 01:14:27.920 --> 01:14:30.100 align:middle line:90% we covered key issues here. 01:14:30.100 --> 01:14:32.790 align:middle line:84% We could go into a couple more formulations, 01:14:32.790 --> 01:14:34.630 align:middle line:84% but I'm interested to see one more model. 01:14:34.630 --> 01:14:38.230 align:middle line:90% So let's look at another model. 01:14:38.230 --> 01:14:43.850 align:middle line:84% Anybody wants to volunteer their model for next round? 01:14:43.850 --> 01:14:44.770 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:14:44.770 --> 01:14:46.670 align:middle line:84% HAZHIR RAHMANDAD: Jake is shaking his head. 01:14:46.670 --> 01:14:48.038 align:middle line:90% [LAUGHS] 01:14:48.038 --> 01:14:49.340 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:14:49.340 --> 01:14:50.930 align:middle line:90% HAZHIR RAHMANDAD: Alban. 01:14:50.930 --> 01:14:54.570 align:middle line:90% Let's go to Alban's model. 01:14:54.570 --> 01:15:01.890 align:middle line:84% So let's look at the not awarded potential clients. 01:15:01.890 --> 01:15:08.570 align:middle line:84% They become awarded potential-- aware, sorry, not award. 01:15:08.570 --> 01:15:09.690 align:middle line:90% Aware. 01:15:09.690 --> 01:15:12.150 align:middle line:84% Awared, awared awared is the word. 01:15:12.150 --> 01:15:14.530 align:middle line:90% It's OK. 01:15:14.530 --> 01:15:20.730 align:middle line:84% And then they can become active clients or not. 01:15:20.730 --> 01:15:24.170 align:middle line:84% And you have a total aware population versus indicated 01:15:24.170 --> 01:15:25.550 align:middle line:90% total aware population. 01:15:25.550 --> 01:15:30.540 align:middle line:84% So this is potentially a two-way flow. 01:15:30.540 --> 01:15:33.140 align:middle line:84% If they indicated it's below, an indicated fraction 01:15:33.140 --> 01:15:36.460 align:middle line:84% of population that is aware comes from the completed 01:15:36.460 --> 01:15:39.960 align:middle line:90% projects and their effects. 01:15:39.960 --> 01:15:41.900 align:middle line:84% So if you have a lot of completed projects, 01:15:41.900 --> 01:15:44.480 align:middle line:90% that increases this effect. 01:15:44.480 --> 01:15:45.920 align:middle line:90% So that increases the fraction. 01:15:45.920 --> 01:15:50.360 align:middle line:84% So big picture, you make more of artifacts projects, 01:15:50.360 --> 01:15:52.540 align:middle line:84% you increase the fraction of population 01:15:52.540 --> 01:15:56.360 align:middle line:84% that may be aware of this methodology, this tool. 01:15:56.360 --> 01:15:58.780 align:middle line:84% And, therefore, that increases the population 01:15:58.780 --> 01:15:59.900 align:middle line:90% that becomes aware. 01:15:59.900 --> 01:16:02.460 align:middle line:84% Now, that's all good with the caveat 01:16:02.460 --> 01:16:03.880 align:middle line:90% that people forget about things. 01:16:03.880 --> 01:16:04.960 align:middle line:90% New generations come. 01:16:04.960 --> 01:16:07.960 align:middle line:84% So I would have an outflow from these completed projects. 01:16:07.960 --> 01:16:10.740 align:middle line:84% This is remembered completed projects and that. 01:16:10.740 --> 01:16:12.660 align:middle line:90% Hassan? 01:16:12.660 --> 01:16:15.940 align:middle line:84% AUDIENCE: In that chain, I had something similar to that chain 01:16:15.940 --> 01:16:16.820 align:middle line:90% up there. 01:16:16.820 --> 01:16:20.300 align:middle line:84% If the indicated total aware population goes down, 01:16:20.300 --> 01:16:22.860 align:middle line:84% and we have a lot of active products, 01:16:22.860 --> 01:16:24.820 align:middle line:84% can't this stock in the middle get negative? 01:16:24.820 --> 01:16:25.800 align:middle line:90% And I have trouble-- 01:16:25.800 --> 01:16:27.217 align:middle line:84% HAZHIR RAHMANDAD: Very good point. 01:16:27.217 --> 01:16:28.410 align:middle line:90% So that's absolutely right. 01:16:28.410 --> 01:16:30.190 align:middle line:84% And that is why you actually need 01:16:30.190 --> 01:16:32.910 align:middle line:84% another layer of protection, which 01:16:32.910 --> 01:16:39.150 align:middle line:84% is a flow from aware population, potential clients, 01:16:39.150 --> 01:16:41.550 align:middle line:90% back to becoming aware. 01:16:41.550 --> 01:16:45.270 align:middle line:84% And this part is fine because this is already 01:16:45.270 --> 01:16:46.910 align:middle line:90% kind of the maximum. 01:16:46.910 --> 01:16:52.110 align:middle line:84% So that the outflow moving backwards cannot exceed aware 01:16:52.110 --> 01:16:55.950 align:middle line:84% potential population divided by this awareness time constant. 01:16:55.950 --> 01:17:02.370 align:middle line:84% So that should set a maximum, minimum rate for backflow. 01:17:02.370 --> 01:17:05.110 align:middle line:84% AUDIENCE: Does that make sense, that it's like a sharp corner? 01:17:05.110 --> 01:17:09.030 align:middle line:84% HAZHIR RAHMANDAD: It is OK as a first order approximation. 01:17:09.030 --> 01:17:09.930 align:middle line:90% You're right. 01:17:09.930 --> 01:17:13.250 align:middle line:84% It would create a sharp corner, but it is not too bad. 01:17:13.250 --> 01:17:16.550 align:middle line:84% So this is actually-- in the assignment 4 of the basic SD 01:17:16.550 --> 01:17:18.050 align:middle line:90% class, we have this same issue. 01:17:18.050 --> 01:17:22.350 align:middle line:90% And we add that formulation. 01:17:22.350 --> 01:17:23.480 align:middle line:90% But that was a good catch. 01:17:23.480 --> 01:17:24.230 align:middle line:90% That is important. 01:17:24.230 --> 01:17:25.790 align:middle line:90% AUDIENCE: Oh, I had the same. 01:17:25.790 --> 01:17:28.670 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so now, total aware 01:17:28.670 --> 01:17:30.090 align:middle line:90% is pretty clear what it is. 01:17:30.090 --> 01:17:33.490 align:middle line:84% Now, people become aware after some time. 01:17:33.490 --> 01:17:36.670 align:middle line:90% 01:17:36.670 --> 01:17:41.150 align:middle line:84% So now we have another question, whether these people 01:17:41.150 --> 01:17:43.190 align:middle line:90% become active clients. 01:17:43.190 --> 01:17:45.450 align:middle line:84% They are using the-- and these are in clients. 01:17:45.450 --> 01:17:47.170 align:middle line:90% So this is in clients. 01:17:47.170 --> 01:17:49.050 align:middle line:84% And then clients can have projects. 01:17:49.050 --> 01:17:50.570 align:middle line:90% So this is demand per client. 01:17:50.570 --> 01:17:53.290 align:middle line:84% So conceptualization is a little bit different here. 01:17:53.290 --> 01:17:55.730 align:middle line:90% So these are longer term people. 01:17:55.730 --> 01:17:58.290 align:middle line:84% These are people who are proponents of the field. 01:17:58.290 --> 01:18:00.510 align:middle line:84% They are using it with some rate, 01:18:00.510 --> 01:18:04.350 align:middle line:84% that service demand per client is the rate by which they 01:18:04.350 --> 01:18:06.170 align:middle line:90% are generating service. 01:18:06.170 --> 01:18:08.290 align:middle line:90% And I kind of like this. 01:18:08.290 --> 01:18:12.630 align:middle line:84% It's kind of, in some ways, more useful way 01:18:12.630 --> 01:18:17.050 align:middle line:84% of breaking down people because it is not the same as projects. 01:18:17.050 --> 01:18:20.310 align:middle line:84% There is more inertia in people and their mentality 01:18:20.310 --> 01:18:21.170 align:middle line:90% than in projects. 01:18:21.170 --> 01:18:26.400 align:middle line:84% Projects are activating and going away. 01:18:26.400 --> 01:18:28.600 align:middle line:84% Then we have the quality reputation 01:18:28.600 --> 01:18:33.400 align:middle line:84% that impacts whether these people who are aware 01:18:33.400 --> 01:18:36.440 align:middle line:90% are willing to use it or not. 01:18:36.440 --> 01:18:40.960 align:middle line:84% So that indication comes from the total aware population 01:18:40.960 --> 01:18:42.440 align:middle line:90% and quality reputation. 01:18:42.440 --> 01:18:47.240 align:middle line:84% And quality reputation, based on this formulation, 01:18:47.240 --> 01:18:51.360 align:middle line:84% should be designed so that a quality reputation of 1 01:18:51.360 --> 01:18:55.640 align:middle line:84% basically means every aware population, every aware person, 01:18:55.640 --> 01:18:59.080 align:middle line:90% wants to be an active client. 01:18:59.080 --> 01:19:01.440 align:middle line:90% And a quality reputation of 1-- 01:19:01.440 --> 01:19:03.860 align:middle line:90% 0 means none of them want to be. 01:19:03.860 --> 01:19:06.640 align:middle line:84% So that's kind of just to have the meaning of quality 01:19:06.640 --> 01:19:09.000 align:middle line:90% to be matched here. 01:19:09.000 --> 01:19:12.840 align:middle line:84% The formulation here is just similar to the other one, 01:19:12.840 --> 01:19:17.120 align:middle line:84% indicated minus active divided by time constant. 01:19:17.120 --> 01:19:25.230 align:middle line:84% If indicated is zero, active is a number that is larger. 01:19:25.230 --> 01:19:27.970 align:middle line:84% It becomes a one-way flow the other way. 01:19:27.970 --> 01:19:29.070 align:middle line:90% This is robust. 01:19:29.070 --> 01:19:31.750 align:middle line:84% This wouldn't have a problem that we talked. 01:19:31.750 --> 01:19:32.590 align:middle line:90% So that's good. 01:19:32.590 --> 01:19:35.290 align:middle line:90% 01:19:35.290 --> 01:19:38.330 align:middle line:84% Yeah, now, we have the demand coming 01:19:38.330 --> 01:19:40.110 align:middle line:90% from these active clients. 01:19:40.110 --> 01:19:42.690 align:middle line:90% They initiate projects. 01:19:42.690 --> 01:19:45.450 align:middle line:90% These projects take some time. 01:19:45.450 --> 01:19:46.950 align:middle line:84% So there is a potential completion. 01:19:46.950 --> 01:19:50.130 align:middle line:84% There is a max completion, very standard formulation. 01:19:50.130 --> 01:19:53.890 align:middle line:84% That potential completion comes from expert 01:19:53.890 --> 01:19:57.770 align:middle line:84% versus novel completion rate that come together to create 01:19:57.770 --> 01:19:59.470 align:middle line:90% a total completion rate. 01:19:59.470 --> 01:20:03.490 align:middle line:84% The expert completion rate is coming from the productivity. 01:20:03.490 --> 01:20:08.150 align:middle line:84% Then you have feedback, effect of schedule pressure, 01:20:08.150 --> 01:20:11.490 align:middle line:84% et cetera, on the right-hand side. 01:20:11.490 --> 01:20:12.070 align:middle line:90% It's OK. 01:20:12.070 --> 01:20:14.910 align:middle line:84% It might be an overkill to have that. 01:20:14.910 --> 01:20:18.290 align:middle line:84% I'm OK with not even having that, so to have, basically, 01:20:18.290 --> 01:20:19.480 align:middle line:90% a fixed productivity. 01:20:19.480 --> 01:20:21.640 align:middle line:84% But there's nothing wrong with it. 01:20:21.640 --> 01:20:24.180 align:middle line:84% It's nice to have something extra here. 01:20:24.180 --> 01:20:26.520 align:middle line:84% Similarly, you have it for novices. 01:20:26.520 --> 01:20:29.180 align:middle line:84% They have a fractional productivity that 01:20:29.180 --> 01:20:32.540 align:middle line:90% gives their total productivity. 01:20:32.540 --> 01:20:42.380 align:middle line:84% And then, again, some feedback from schedule pressure. 01:20:42.380 --> 01:20:49.620 align:middle line:84% Now, you have two desired completion rates 01:20:49.620 --> 01:20:52.000 align:middle line:90% for novice versus experts. 01:20:52.000 --> 01:20:57.100 align:middle line:84% So you are creating these more complex schedule pressures. 01:20:57.100 --> 01:21:01.420 align:middle line:84% AUDIENCE: So the assumption is that we first 01:21:01.420 --> 01:21:06.300 align:middle line:90% allocate tasks to the expert? 01:21:06.300 --> 01:21:08.620 align:middle line:84% And then they can adjust their work hours, 01:21:08.620 --> 01:21:10.320 align:middle line:90% but to the max extent. 01:21:10.320 --> 01:21:14.438 align:middle line:84% And then whatever is that, you have to push novice to do more. 01:21:14.438 --> 01:21:15.480 align:middle line:90% HAZHIR RAHMANDAD: Got it. 01:21:15.480 --> 01:21:17.510 align:middle line:90% Got it. 01:21:17.510 --> 01:21:18.290 align:middle line:90% That's good. 01:21:18.290 --> 01:21:18.990 align:middle line:90% That's fine. 01:21:18.990 --> 01:21:22.590 align:middle line:84% I think that's actually a cleaner 01:21:22.590 --> 01:21:30.190 align:middle line:84% formulation than assuming kind of equal allocation. 01:21:30.190 --> 01:21:32.910 align:middle line:84% I might have opted for something that 01:21:32.910 --> 01:21:35.850 align:middle line:84% allows us to move between the two extremes, 01:21:35.850 --> 01:21:38.510 align:middle line:84% so that a better matching algorithm allows 01:21:38.510 --> 01:21:43.470 align:middle line:90% you to do one versus the other. 01:21:43.470 --> 01:21:49.030 align:middle line:84% But this is a good starting point and the formulation 01:21:49.030 --> 01:21:50.690 align:middle line:90% we haven't captured before. 01:21:50.690 --> 01:21:54.630 align:middle line:84% So, basically, active projects create a desired completion rate 01:21:54.630 --> 01:21:57.730 align:middle line:84% that, minus what the experts are completing, 01:21:57.730 --> 01:21:59.550 align:middle line:84% gives the desired completion rate for N. 01:21:59.550 --> 01:22:04.270 align:middle line:84% And E completion rate is coming from here. 01:22:04.270 --> 01:22:06.950 align:middle line:90% E completion rate, yes. 01:22:06.950 --> 01:22:10.230 align:middle line:84% All of that makes a lot of sense. 01:22:10.230 --> 01:22:12.590 align:middle line:84% Let's look at the learning rates now. 01:22:12.590 --> 01:22:19.720 align:middle line:84% So there is joining of novices, average monthly new joiners 01:22:19.720 --> 01:22:22.940 align:middle line:84% of the field, and effective demand on joining. 01:22:22.940 --> 01:22:28.400 align:middle line:90% 01:22:28.400 --> 01:22:31.620 align:middle line:84% Yeah, so this is basically setting a maximum. 01:22:31.620 --> 01:22:33.880 align:middle line:90% Is it average or maximum? 01:22:33.880 --> 01:22:34.680 align:middle line:90% AUDIENCE: Average. 01:22:34.680 --> 01:22:36.055 align:middle line:84% HAZHIR RAHMANDAD: So this effect, 01:22:36.055 --> 01:22:38.340 align:middle line:90% can you go above 1 and below 1? 01:22:38.340 --> 01:22:42.280 align:middle line:90% 01:22:42.280 --> 01:22:42.840 align:middle line:90% We'll see. 01:22:42.840 --> 01:22:47.240 align:middle line:84% Relative demand is coming from perceived demand. 01:22:47.240 --> 01:22:50.040 align:middle line:90% That's fine. 01:22:50.040 --> 01:22:56.480 align:middle line:84% And then average demand of alternative fields, so 01:22:56.480 --> 01:22:57.220 align:middle line:90% what's that? 01:22:57.220 --> 01:23:02.520 align:middle line:90% 01:23:02.520 --> 01:23:07.200 align:middle line:84% It's kind of comparing this field to some other alternative 01:23:07.200 --> 01:23:10.180 align:middle line:90% field to give a relative demand. 01:23:10.180 --> 01:23:13.520 align:middle line:90% 01:23:13.520 --> 01:23:16.440 align:middle line:84% So this part, I'm not quite following 01:23:16.440 --> 01:23:18.180 align:middle line:84% why you are formulating this way. 01:23:18.180 --> 01:23:20.120 align:middle line:84% Can you tell a little bit more, Alban? 01:23:20.120 --> 01:23:23.880 align:middle line:84% AUDIENCE: Yeah, so I'm thinking people deciding to join a few 01:23:23.880 --> 01:23:27.360 align:middle line:90% are being fulfilled. 01:23:27.360 --> 01:23:30.600 align:middle line:84% And assuming that their signal is only demand, 01:23:30.600 --> 01:23:34.360 align:middle line:84% so they compared this field versus another field 01:23:34.360 --> 01:23:38.160 align:middle line:90% in terms of the attraction. 01:23:38.160 --> 01:23:41.320 align:middle line:84% HAZHIR RAHMANDAD: Yeah, but the problem with that 01:23:41.320 --> 01:23:44.660 align:middle line:84% is that people don't look at the demand alone. 01:23:44.660 --> 01:23:50.040 align:middle line:84% They should be looking at demand versus supply to see, 01:23:50.040 --> 01:23:54.600 align:middle line:84% what is the room for joining, or not joining, 01:23:54.600 --> 01:23:56.360 align:middle line:90% or kind of expansion? 01:23:56.360 --> 01:24:01.080 align:middle line:84% And that usually comes from a signal of delivery delay, 01:24:01.080 --> 01:24:06.200 align:middle line:84% or reputation, or something else. 01:24:06.200 --> 01:24:10.880 align:middle line:84% So this one is kind of driving the whole joining process 01:24:10.880 --> 01:24:13.270 align:middle line:90% with two external parameters. 01:24:13.270 --> 01:24:15.690 align:middle line:84% One is this average monthly joining, 01:24:15.690 --> 01:24:19.030 align:middle line:84% and the other is this average demand of alternative field. 01:24:19.030 --> 01:24:21.270 align:middle line:84% Both of them are this external anchor, 01:24:21.270 --> 01:24:25.390 align:middle line:84% as I can't fully see how it impacts all of the dynamics. 01:24:25.390 --> 01:24:30.330 align:middle line:84% But they are disjointed from the rest of the story in some sense. 01:24:30.330 --> 01:24:36.150 align:middle line:84% So I like this to be kind of focused on serving the demand. 01:24:36.150 --> 01:24:39.470 align:middle line:84% Kind of, Jason's formulation started to get to that, 01:24:39.470 --> 01:24:46.010 align:middle line:84% but then it made some wrong adjustments. 01:24:46.010 --> 01:24:49.350 align:middle line:84% But the idea that, yes, there is this much demand. 01:24:49.350 --> 01:24:52.010 align:middle line:90% There is this much supply. 01:24:52.010 --> 01:24:53.310 align:middle line:90% There is a mismatch. 01:24:53.310 --> 01:24:56.650 align:middle line:84% That mismatch motivates people to join because they see 01:24:56.650 --> 01:24:58.110 align:middle line:90% there is room for growth here. 01:24:58.110 --> 01:25:02.290 align:middle line:84% There is kind of business coming in, something along those lines. 01:25:02.290 --> 01:25:05.430 align:middle line:84% And then you don't need to have this external parameter. 01:25:05.430 --> 01:25:10.020 align:middle line:84% It can be all fairly endogenized into the adjustment 01:25:10.020 --> 01:25:11.100 align:middle line:90% of the field. 01:25:11.100 --> 01:25:13.500 align:middle line:84% Then you have these people quitting. 01:25:13.500 --> 01:25:14.940 align:middle line:84% Some of it is coming from the fact 01:25:14.940 --> 01:25:16.400 align:middle line:84% that they don't have enough demand. 01:25:16.400 --> 01:25:17.080 align:middle line:90% That's all good. 01:25:17.080 --> 01:25:20.780 align:middle line:84% I haven't looked into it, but the idea is correct. 01:25:20.780 --> 01:25:22.500 align:middle line:84% Then you have the learning time that 01:25:22.500 --> 01:25:26.480 align:middle line:84% is coming from just some maximum and minimum. 01:25:26.480 --> 01:25:28.780 align:middle line:84% And then there is this adjustment 01:25:28.780 --> 01:25:34.180 align:middle line:90% based on practical experience. 01:25:34.180 --> 01:25:38.100 align:middle line:84% So you are calculating how much the novices are actually 01:25:38.100 --> 01:25:39.040 align:middle line:90% actively working. 01:25:39.040 --> 01:25:40.780 align:middle line:84% So if they are not actively working, 01:25:40.780 --> 01:25:43.080 align:middle line:84% it becomes the maximum learning time. 01:25:43.080 --> 01:25:44.820 align:middle line:84% And if they are actively working there, 01:25:44.820 --> 01:25:46.860 align:middle line:84% it becomes the minimum learning time. 01:25:46.860 --> 01:25:47.600 align:middle line:90% That's nice. 01:25:47.600 --> 01:25:48.140 align:middle line:90% I like that. 01:25:48.140 --> 01:25:51.900 align:middle line:84% It's kind of a clean way of capturing the fact 01:25:51.900 --> 01:25:54.560 align:middle line:84% that, if novices are not given projects, 01:25:54.560 --> 01:25:57.580 align:middle line:84% it takes more time for them to grow. 01:25:57.580 --> 01:26:01.980 align:middle line:84% Now, you don't have the impact of experience, folks, 01:26:01.980 --> 01:26:04.100 align:middle line:90% training novices, perhaps. 01:26:04.100 --> 01:26:10.730 align:middle line:84% So in a well-coordinated field, when there is not much demand, 01:26:10.730 --> 01:26:14.610 align:middle line:84% experts can actually help more with training of the novices. 01:26:14.610 --> 01:26:18.390 align:middle line:84% So the training time may actually shrink, perhaps. 01:26:18.390 --> 01:26:20.650 align:middle line:84% So we could come up with alternatives there. 01:26:20.650 --> 01:26:22.847 align:middle line:84% But this is a good starting point. 01:26:22.847 --> 01:26:24.430 align:middle line:84% And then you have the effective demand 01:26:24.430 --> 01:26:27.630 align:middle line:84% for quitting of the experts as well. 01:26:27.630 --> 01:26:30.710 align:middle line:90% That's all fine. 01:26:30.710 --> 01:26:34.870 align:middle line:84% I think the rest comes together cleanly. 01:26:34.870 --> 01:26:37.770 align:middle line:84% Then let's look at the reputation. 01:26:37.770 --> 01:26:40.190 align:middle line:84% So quality reputation comes from the expert 01:26:40.190 --> 01:26:49.390 align:middle line:84% versus novel completion rates, so E completion plus N 01:26:49.390 --> 01:26:54.190 align:middle line:84% completion multiplied by this quality fraction divided by E 01:26:54.190 --> 01:26:56.030 align:middle line:90% completion plus N completion. 01:26:56.030 --> 01:27:02.170 align:middle line:84% So, yeah, that that's fine for indicated quality. 01:27:02.170 --> 01:27:03.990 align:middle line:84% Now, what I want to highlight here 01:27:03.990 --> 01:27:08.860 align:middle line:84% is that, right now, this is a little bit more subtle. 01:27:08.860 --> 01:27:13.080 align:middle line:84% So the way indicated quality is formulated right now 01:27:13.080 --> 01:27:16.600 align:middle line:84% is very tightly connected to your stock and flow of novice 01:27:16.600 --> 01:27:19.200 align:middle line:90% versus experienced. 01:27:19.200 --> 01:27:20.380 align:middle line:90% Does that make sense? 01:27:20.380 --> 01:27:23.180 align:middle line:84% So it's really about this relative quality 01:27:23.180 --> 01:27:26.520 align:middle line:84% and what fraction of people are in these two stocks. 01:27:26.520 --> 01:27:29.340 align:middle line:84% What fraction of people are in these two stocks really, 01:27:29.340 --> 01:27:31.880 align:middle line:84% at the end of it, comes back to these time constants 01:27:31.880 --> 01:27:36.460 align:middle line:84% for quitting and for assimilation of the two, guys. 01:27:36.460 --> 01:27:40.600 align:middle line:84% That's kind of what determines the core under it. 01:27:40.600 --> 01:27:43.920 align:middle line:84% So when these two come together, they 01:27:43.920 --> 01:27:48.300 align:middle line:84% indicated quality has an anchor based on those time constants 01:27:48.300 --> 01:27:53.640 align:middle line:84% and based on this quality fraction. 01:27:53.640 --> 01:27:57.720 align:middle line:84% That anchor may not map exactly into the anchor 01:27:57.720 --> 01:28:04.770 align:middle line:84% you want for quality reputation impacting the indicated A. 01:28:04.770 --> 01:28:07.490 align:middle line:84% That is, the built-in implicit anchor 01:28:07.490 --> 01:28:12.890 align:middle line:84% that you have from the workforce sector 01:28:12.890 --> 01:28:16.690 align:middle line:84% doesn't necessarily give you the minimum and maximum 01:28:16.690 --> 01:28:21.250 align:middle line:84% of this quality reputation that you need for indicated A. 01:28:21.250 --> 01:28:23.050 align:middle line:90% Remember what indicated A was. 01:28:23.050 --> 01:28:25.890 align:middle line:84% Indicated A is about people deciding, 01:28:25.890 --> 01:28:27.970 align:middle line:90% this is an attractive service. 01:28:27.970 --> 01:28:32.410 align:middle line:84% This is a good quality service that they want to use. 01:28:32.410 --> 01:28:35.170 align:middle line:84% So to give you an extreme example, if the indicated 01:28:35.170 --> 01:28:39.090 align:middle line:84% quality only changes between 0.5 and 0.6, 01:28:39.090 --> 01:28:41.110 align:middle line:84% that's the range of feasible things 01:28:41.110 --> 01:28:43.650 align:middle line:84% that may happen with these parameters. 01:28:43.650 --> 01:28:48.270 align:middle line:84% Then your indicated A will only change between 0.5 and 0.6. 01:28:48.270 --> 01:28:51.390 align:middle line:84% It doesn't have much flexibility beyond that. 01:28:51.390 --> 01:28:54.850 align:middle line:90% 01:28:54.850 --> 01:28:59.490 align:middle line:84% AUDIENCE: But it is not only the stock of the expert novice 01:28:59.490 --> 01:29:01.500 align:middle line:84% because this is a completion rate. 01:29:01.500 --> 01:29:05.100 align:middle line:84% And then if you don't have demand, none of us 01:29:05.100 --> 01:29:07.080 align:middle line:90% will push to do no work. 01:29:07.080 --> 01:29:09.460 align:middle line:84% So the completion rate by novice is zero. 01:29:09.460 --> 01:29:10.660 align:middle line:90% HAZHIR RAHMANDAD: Yes. 01:29:10.660 --> 01:29:13.670 align:middle line:84% AUDIENCE: So they can go a wide range, not just 01:29:13.670 --> 01:29:14.420 align:middle line:90% incremental stuff. 01:29:14.420 --> 01:29:15.670 align:middle line:90% HAZHIR RAHMANDAD: That's fair. 01:29:15.670 --> 01:29:17.980 align:middle line:84% So I think it can go to 1 when you don't 01:29:17.980 --> 01:29:21.380 align:middle line:90% have projects for novices. 01:29:21.380 --> 01:29:26.540 align:middle line:90% It cannot go to 0. 01:29:26.540 --> 01:29:29.860 align:middle line:84% The lower bound might not be very high. 01:29:29.860 --> 01:29:34.380 align:middle line:84% So if you have a lot of demand, it is still-- 01:29:34.380 --> 01:29:36.400 align:middle line:84% then it is bounded by that parameter. 01:29:36.400 --> 01:29:38.120 align:middle line:84% So the range is probably on the top. 01:29:38.120 --> 01:29:39.793 align:middle line:90% It goes to 1, of course. 01:29:39.793 --> 01:29:40.960 align:middle line:90% On the bottom, I don't know. 01:29:40.960 --> 01:29:44.000 align:middle line:84% But there is this inherent anchor question there. 01:29:44.000 --> 01:29:46.580 align:middle line:84% Now, you do have a little bit of control over it 01:29:46.580 --> 01:29:49.220 align:middle line:84% with this quality fraction of N. If you think of it 01:29:49.220 --> 01:29:52.340 align:middle line:84% as a free parameter to play with, 01:29:52.340 --> 01:29:57.100 align:middle line:84% that could shift the lower bound a little bit. 01:29:57.100 --> 01:30:00.020 align:middle line:84% But you don't want that because now it 01:30:00.020 --> 01:30:02.040 align:middle line:90% is playing two different roles. 01:30:02.040 --> 01:30:03.580 align:middle line:90% It's a little bit tricky. 01:30:03.580 --> 01:30:08.260 align:middle line:84% So, anyway, it's an invitation to have just one more free 01:30:08.260 --> 01:30:12.100 align:middle line:84% parameter to make this adjustment of quality reputation 01:30:12.100 --> 01:30:14.740 align:middle line:84% to indicated A. Something there might benefit 01:30:14.740 --> 01:30:18.315 align:middle line:90% from another degree of freedom. 01:30:18.315 --> 01:30:19.940 align:middle line:84% But, otherwise, this looks pretty good. 01:30:19.940 --> 01:30:23.340 align:middle line:84% I think kind of-- so that factor and this anchoring 01:30:23.340 --> 01:30:27.240 align:middle line:84% here are the two that I would highlight. 01:30:27.240 --> 01:30:30.740 align:middle line:90% And if we run this before we-- 01:30:30.740 --> 01:30:35.620 align:middle line:84% oh, now, you have a field that is growing. 01:30:35.620 --> 01:30:42.940 align:middle line:84% Now, if these folks are pretty lousy, 01:30:42.940 --> 01:30:45.600 align:middle line:84% you still get them to be growing. 01:30:45.600 --> 01:30:51.900 align:middle line:90% 01:30:51.900 --> 01:30:54.540 align:middle line:84% And part of it is that the completed projects continues 01:30:54.540 --> 01:30:59.510 align:middle line:90% to push the not awarded-- 01:30:59.510 --> 01:31:01.790 align:middle line:90% not aware potential clients. 01:31:01.790 --> 01:31:09.310 align:middle line:84% So this will always be this total desire-indicated fraction 01:31:09.310 --> 01:31:10.290 align:middle line:90% aware will go up. 01:31:10.290 --> 01:31:12.450 align:middle line:84% Although, it is already going to 1 very quickly. 01:31:12.450 --> 01:31:16.530 align:middle line:84% So then there are these parametric assumptions. 01:31:16.530 --> 01:31:18.310 align:middle line:84% Is it reasonable to assume, in five years, 01:31:18.310 --> 01:31:22.490 align:middle line:84% everybody would want to know about it and so on? 01:31:22.490 --> 01:31:26.070 align:middle line:90% 01:31:26.070 --> 01:31:27.350 align:middle line:90% But it's interesting. 01:31:27.350 --> 01:31:30.630 align:middle line:84% If this indicated total aware population 01:31:30.630 --> 01:31:37.150 align:middle line:84% is 1, that, very quickly, why don't we see more movement? 01:31:37.150 --> 01:31:43.310 align:middle line:84% Because not aware is actually pretty high. 01:31:43.310 --> 01:31:48.030 align:middle line:84% And this time constant is only six months. 01:31:48.030 --> 01:32:00.380 align:middle line:84% So something-- there is a big gap here. 01:32:00.380 --> 01:32:03.760 align:middle line:90% 01:32:03.760 --> 01:32:07.900 align:middle line:84% That's not clear to me, what is happening here. 01:32:07.900 --> 01:32:11.880 align:middle line:90% 01:32:11.880 --> 01:32:13.640 align:middle line:84% So I think it's a good time to take 01:32:13.640 --> 01:32:17.760 align:middle line:84% a break, but lots of good formulations 01:32:17.760 --> 01:32:19.060 align:middle line:90% and interesting challenges. 01:32:19.060 --> 01:32:22.140 align:middle line:84% I think this was a more challenging project. 01:32:22.140 --> 01:32:24.480 align:middle line:84% And I think you guys, each of you 01:32:24.480 --> 01:32:28.060 align:middle line:84% have some good formulations that capture one part of it. 01:32:28.060 --> 01:32:31.040 align:middle line:84% We haven't yet had one that brings all of them 01:32:31.040 --> 01:32:33.120 align:middle line:84% fully together, but this is good. 01:32:33.120 --> 01:32:35.720 align:middle line:84% So let's take 10 minutes of a break. 01:32:35.720 --> 01:32:41.880 align:middle line:84% When we come back, we can focus more on your project proposals, 01:32:41.880 --> 01:32:43.940 align:middle line:84% get any questions there answered, 01:32:43.940 --> 01:32:46.103 align:middle line:84% get some of your feedback for your friends. 01:32:46.103 --> 01:32:48.520 align:middle line:84% And if you had more time, we can go through another couple 01:32:48.520 --> 01:32:50.860 align:middle line:90% more formulations as well. 01:32:50.860 --> 01:32:54.270 align:middle line:90% 01:32:54.270 --> 01:32:57.890 align:middle line:84% Let's come back together, get us started 01:32:57.890 --> 01:33:07.890 align:middle line:84% with your discussion of problems, problem 01:33:07.890 --> 01:33:10.430 align:middle line:90% definitions of your friends. 01:33:10.430 --> 01:33:11.950 align:middle line:90% So I'm going to those. 01:33:11.950 --> 01:33:18.530 align:middle line:90% 01:33:18.530 --> 01:33:19.490 align:middle line:90% So we had it. 01:33:19.490 --> 01:33:26.450 align:middle line:84% Was it with comments from classmates? 01:33:26.450 --> 01:33:26.950 align:middle line:90% Good. 01:33:26.950 --> 01:33:29.830 align:middle line:90% So we have a few here already. 01:33:29.830 --> 01:33:30.690 align:middle line:90% That's great. 01:33:30.690 --> 01:33:35.050 align:middle line:84% So I think several of you have provided 01:33:35.050 --> 01:33:37.930 align:middle line:90% feedback for different people. 01:33:37.930 --> 01:33:41.810 align:middle line:84% Sercin and Jason have been lucky to have lots of feedback, 01:33:41.810 --> 01:33:43.570 align:middle line:90% somehow by chance. 01:33:43.570 --> 01:33:47.170 align:middle line:84% AUDIENCE: Wait, oh, that is weird. 01:33:47.170 --> 01:33:49.470 align:middle line:90% And I haven't done one yet. 01:33:49.470 --> 01:33:52.000 align:middle line:84% HAZHIR RAHMANDAD: So you're the free rider today. 01:33:52.000 --> 01:33:53.174 align:middle line:90% [LAUGHTER] 01:33:53.174 --> 01:33:57.540 align:middle line:90% 01:33:57.540 --> 01:34:01.860 align:middle line:84% Why don't we just go in this order and get reactions? 01:34:01.860 --> 01:34:05.260 align:middle line:90% So I think this is Hannah. 01:34:05.260 --> 01:34:07.600 align:middle line:84% She had a feedback on Cathy's project. 01:34:07.600 --> 01:34:10.340 align:middle line:84% So, first of all, because nobody knows 01:34:10.340 --> 01:34:13.920 align:middle line:84% what each person's project was about, 01:34:13.920 --> 01:34:18.060 align:middle line:90% I am happy to open these, but-- 01:34:18.060 --> 01:34:20.440 align:middle line:84% and these are on the first versions of the project. 01:34:20.440 --> 01:34:28.960 align:middle line:84% So Cathy's project was about literacy and book sales. 01:34:28.960 --> 01:34:31.700 align:middle line:84% So her starting point was something 01:34:31.700 --> 01:34:38.320 align:middle line:84% that looks like a puzzle, that we have had declining literacy, 01:34:38.320 --> 01:34:40.680 align:middle line:84% but book sales seem to be increasing. 01:34:40.680 --> 01:34:43.660 align:middle line:84% And then she's coming up with hypothesis for what 01:34:43.660 --> 01:34:47.180 align:middle line:90% might be explaining that. 01:34:47.180 --> 01:34:49.808 align:middle line:84% So, Hannah, kind of what was your reaction 01:34:49.808 --> 01:34:50.850 align:middle line:90% as you went through this? 01:34:50.850 --> 01:34:54.610 align:middle line:84% And did you comment on the first version or on the later version? 01:34:54.610 --> 01:34:55.150 align:middle line:90% I haven't-- 01:34:55.150 --> 01:34:56.650 align:middle line:84% AUDIENCE: This is the first version. 01:34:56.650 --> 01:34:56.850 align:middle line:90% HAZHIR RAHMANDAD: Good. 01:34:56.850 --> 01:34:59.270 align:middle line:84% So that was the plan, just because I wanted everybody 01:34:59.270 --> 01:35:04.510 align:middle line:84% to go through a more raw problem definition that 01:35:04.510 --> 01:35:07.350 align:middle line:84% helps see the challenges in defining 01:35:07.350 --> 01:35:08.690 align:middle line:90% the problem more clearly. 01:35:08.690 --> 01:35:11.390 align:middle line:90% 01:35:11.390 --> 01:35:13.650 align:middle line:84% So rather than reading through this, 01:35:13.650 --> 01:35:16.070 align:middle line:84% maybe you share, Hannah, what was your reactions? 01:35:16.070 --> 01:35:18.590 align:middle line:84% And what did you find potentially interesting, 01:35:18.590 --> 01:35:19.630 align:middle line:90% challenging? 01:35:19.630 --> 01:35:23.710 align:middle line:84% How would you think about making the problem more clear? 01:35:23.710 --> 01:35:25.310 align:middle line:84% AUDIENCE: Mm, my first impression 01:35:25.310 --> 01:35:28.350 align:middle line:84% was that the problem was very interesting. 01:35:28.350 --> 01:35:30.150 align:middle line:84% It was something I never heard of before, 01:35:30.150 --> 01:35:31.690 align:middle line:84% and the graphs were very helpful. 01:35:31.690 --> 01:35:35.230 align:middle line:84% There was graphs towards the end. 01:35:35.230 --> 01:35:40.230 align:middle line:84% And I think, for me, the only thing that I was a little bit-- 01:35:40.230 --> 01:35:42.870 align:middle line:84% I thought could use a little bit more improvement 01:35:42.870 --> 01:35:49.110 align:middle line:84% was just making the problem itself a bit more specific. 01:35:49.110 --> 01:35:51.270 align:middle line:90% It seemed very broad. 01:35:51.270 --> 01:35:54.150 align:middle line:84% I think because there was a lot of really great background 01:35:54.150 --> 01:35:58.240 align:middle line:84% information, but the problem itself towards the end 01:35:58.240 --> 01:35:59.490 align:middle line:90% seemed like it was very broad. 01:35:59.490 --> 01:36:01.952 align:middle line:84% It was just like on goods, generally. 01:36:01.952 --> 01:36:04.410 align:middle line:84% And maybe it could have been just focused more on the books 01:36:04.410 --> 01:36:06.130 align:middle line:84% that you mentioned in the beginning. 01:36:06.130 --> 01:36:08.790 align:middle line:90% 01:36:08.790 --> 01:36:11.630 align:middle line:84% Yeah, but I think, overall, it was very interesting. 01:36:11.630 --> 01:36:14.670 align:middle line:84% I think Cathy explained the dynamics pretty well 01:36:14.670 --> 01:36:17.232 align:middle line:90% in the background information. 01:36:17.232 --> 01:36:18.190 align:middle line:90% HAZHIR RAHMANDAD: Good. 01:36:18.190 --> 01:36:19.023 align:middle line:90% AUDIENCE: Thank you. 01:36:19.023 --> 01:36:21.418 align:middle line:84% Yeah, I think in the revised version-- 01:36:21.418 --> 01:36:23.210 align:middle line:84% HAZHIR RAHMANDAD: Yeah, could you speak up? 01:36:23.210 --> 01:36:24.010 align:middle line:90% AUDIENCE: Oh, yeah, sorry. 01:36:24.010 --> 01:36:25.750 align:middle line:84% In the revised version, I think you 01:36:25.750 --> 01:36:30.590 align:middle line:84% gave a similar comment of, don't generalize to everything, 01:36:30.590 --> 01:36:32.170 align:middle line:90% but just focus on books. 01:36:32.170 --> 01:36:33.038 align:middle line:90% And so, yeah-- 01:36:33.038 --> 01:36:34.830 align:middle line:84% HAZHIR RAHMANDAD: So can you now summarize, 01:36:34.830 --> 01:36:38.025 align:middle line:84% what is your updated kind of project definition based 01:36:38.025 --> 01:36:39.650 align:middle line:84% on the feedback that you have received? 01:36:39.650 --> 01:36:41.358 align:middle line:84% AUDIENCE: Yeah, so based on the feedback, 01:36:41.358 --> 01:36:45.280 align:middle line:84% I think, so the puzzle is kind of increasing book 01:36:45.280 --> 01:36:48.940 align:middle line:84% sales and decreasing reading rates among adults. 01:36:48.940 --> 01:36:55.180 align:middle line:84% And I was looking around for book sales data, 01:36:55.180 --> 01:36:58.440 align:middle line:84% and I actually think there's better data for library 01:36:58.440 --> 01:36:59.260 align:middle line:90% circulation. 01:36:59.260 --> 01:37:02.580 align:middle line:84% Although, I haven't found that anybody's compiled it. 01:37:02.580 --> 01:37:04.240 align:middle line:84% So I might have to do some little-- 01:37:04.240 --> 01:37:07.440 align:middle line:84% take a couple of states or something 01:37:07.440 --> 01:37:12.440 align:middle line:84% and do some quick analysis for that. 01:37:12.440 --> 01:37:15.920 align:middle line:84% And my hypothesis for what's driving 01:37:15.920 --> 01:37:19.680 align:middle line:84% these kind of divergent trends in sales versus reading, 01:37:19.680 --> 01:37:26.000 align:middle line:84% as surveyed, is that people have decoupled the value that you 01:37:26.000 --> 01:37:28.680 align:middle line:84% get out of reading a book from the value 01:37:28.680 --> 01:37:35.000 align:middle line:84% that you get, kind of signaling that you are among pop culture 01:37:35.000 --> 01:37:38.440 align:middle line:84% circles who have the book and who are talking about the book, 01:37:38.440 --> 01:37:41.940 align:middle line:84% and how these-- mostly on social media, 01:37:41.940 --> 01:37:46.770 align:middle line:84% like the conversation about the books on social media 01:37:46.770 --> 01:37:49.352 align:middle line:84% includes kind of full summaries and all the plot points. 01:37:49.352 --> 01:37:51.810 align:middle line:84% So you don't even really need to sit down and read the book 01:37:51.810 --> 01:37:56.010 align:middle line:84% to have something to add to the conversation any longer. 01:37:56.010 --> 01:37:58.230 align:middle line:84% And then I think last time we spoke about it, 01:37:58.230 --> 01:37:59.770 align:middle line:84% it was pointed out that that's kind 01:37:59.770 --> 01:38:03.030 align:middle line:84% of one hypothesis for how these trends could be generated. 01:38:03.030 --> 01:38:04.730 align:middle line:84% And then another hypothesis for which 01:38:04.730 --> 01:38:10.890 align:middle line:84% I found support online and looking into it 01:38:10.890 --> 01:38:13.870 align:middle line:84% is that there are heterogeneous groups of readers, 01:38:13.870 --> 01:38:20.290 align:middle line:84% and that the differences between people who are reading 01:38:20.290 --> 01:38:22.290 align:middle line:84% a lot more and people who are reading a lot less 01:38:22.290 --> 01:38:23.470 align:middle line:90% are actually growing. 01:38:23.470 --> 01:38:26.890 align:middle line:84% So, whereas, maybe the average of group A and group B 01:38:26.890 --> 01:38:29.910 align:middle line:90% would be eight and 12. 01:38:29.910 --> 01:38:34.430 align:middle line:84% Now it's like three and 20 or something like that. 01:38:34.430 --> 01:38:38.370 align:middle line:84% And so these divergent trends could also 01:38:38.370 --> 01:38:42.300 align:middle line:84% potentially explain the difference 01:38:42.300 --> 01:38:44.140 align:middle line:90% between consumption and sales. 01:38:44.140 --> 01:38:46.413 align:middle line:90% And so I'm going to model-- 01:38:46.413 --> 01:38:48.580 align:middle line:84% I'm going to try to build a model with both of those 01:38:48.580 --> 01:38:53.940 align:middle line:84% and see whether both of those mechanisms 01:38:53.940 --> 01:38:55.720 align:middle line:90% can generate the pattern. 01:38:55.720 --> 01:38:59.793 align:middle line:84% And I think maybe it's some combination of both as well. 01:38:59.793 --> 01:39:00.960 align:middle line:90% HAZHIR RAHMANDAD: Very good. 01:39:00.960 --> 01:39:02.085 align:middle line:90% So that was a nice summary. 01:39:02.085 --> 01:39:05.380 align:middle line:90% So, basically, this is kind of-- 01:39:05.380 --> 01:39:10.220 align:middle line:84% so it starts with a nice clean puzzle, kind of surprising. 01:39:10.220 --> 01:39:12.260 align:middle line:84% Assuming that the data is OK, now we 01:39:12.260 --> 01:39:14.140 align:middle line:84% don't really have or haven't seen 01:39:14.140 --> 01:39:17.920 align:middle line:84% the data on sales of the books, so to be 100% sure on that. 01:39:17.920 --> 01:39:21.940 align:middle line:84% But literacy, at least, it is a couple of graphs 01:39:21.940 --> 01:39:25.040 align:middle line:84% in terms of average number of books read in the last year, 01:39:25.040 --> 01:39:27.500 align:middle line:84% for example, has been going down. 01:39:27.500 --> 01:39:31.380 align:middle line:84% Percentage of people who read any books in the last year, 01:39:31.380 --> 01:39:34.140 align:middle line:84% it's kind of stable, a little bit going down. 01:39:34.140 --> 01:39:38.500 align:middle line:84% So there is evidence on the reading side. 01:39:38.500 --> 01:39:40.590 align:middle line:84% Now, assuming that the book sales hasn't 01:39:40.590 --> 01:39:44.590 align:middle line:84% gone down or perhaps even increased, then 01:39:44.590 --> 01:39:45.970 align:middle line:90% we have an interesting puzzle. 01:39:45.970 --> 01:39:48.650 align:middle line:84% AUDIENCE: Which I think is a relatively safe assumption. 01:39:48.650 --> 01:39:50.150 align:middle line:84% One thing you might have heard about 01:39:50.150 --> 01:39:54.150 align:middle line:84% is that Barnes and Nobles were slated to close, 01:39:54.150 --> 01:39:57.050 align:middle line:84% and now they're opening stores instead. 01:39:57.050 --> 01:40:02.510 align:middle line:84% And Barnes and Noble, which is a huge book retailer here, 01:40:02.510 --> 01:40:04.635 align:middle line:84% has said that all of their trends have reversed 01:40:04.635 --> 01:40:06.510 align:middle line:84% and they have really positive outlooks thanks 01:40:06.510 --> 01:40:08.130 align:middle line:90% to the social media trends. 01:40:08.130 --> 01:40:10.210 align:middle line:84% HAZHIR RAHMANDAD: Yeah, that's very interesting. 01:40:10.210 --> 01:40:13.630 align:middle line:84% So, anyway, good starting problem. 01:40:13.630 --> 01:40:15.430 align:middle line:84% The first version of the proposal 01:40:15.430 --> 01:40:17.870 align:middle line:84% was a little bit kind of unclear, 01:40:17.870 --> 01:40:20.970 align:middle line:90% what was the exact hypothesis? 01:40:20.970 --> 01:40:23.910 align:middle line:84% You had this idea of social media 01:40:23.910 --> 01:40:27.710 align:middle line:84% creating some kind of a demand for books 01:40:27.710 --> 01:40:28.980 align:middle line:90% that is not real demand. 01:40:28.980 --> 01:40:30.730 align:middle line:84% That is, people are not reading the books. 01:40:30.730 --> 01:40:31.813 align:middle line:90% They're just buying books. 01:40:31.813 --> 01:40:35.430 align:middle line:84% So that's, ultimately, one hypothesis that is now clear. 01:40:35.430 --> 01:40:38.640 align:middle line:84% We added in the conversation the kind of possibility 01:40:38.640 --> 01:40:42.400 align:middle line:84% of bifurcation, of having subpopulations that are reading 01:40:42.400 --> 01:40:44.320 align:middle line:90% a lot more, and then there's-- 01:40:44.320 --> 01:40:48.400 align:middle line:84% and I think making those hypotheses, turning them 01:40:48.400 --> 01:40:52.700 align:middle line:84% into quantitative model alternatives, is a nice one. 01:40:52.700 --> 01:40:54.460 align:middle line:90% It's kind of in the right scope. 01:40:54.460 --> 01:40:56.480 align:middle line:90% So great. 01:40:56.480 --> 01:41:01.680 align:middle line:84% Go with it, and we'll see how things come together. 01:41:01.680 --> 01:41:16.100 align:middle line:84% Next one, let's-- so you should share your overview, 01:41:16.100 --> 01:41:19.300 align:middle line:84% and then Gerardo can share his feedback. 01:41:19.300 --> 01:41:26.400 align:middle line:84% AUDIENCE: So the overview of my problem is on diabetes care, 01:41:26.400 --> 01:41:31.240 align:middle line:84% that the prevalence of diabetes has been increasing 01:41:31.240 --> 01:41:32.740 align:middle line:90% in the past few decades. 01:41:32.740 --> 01:41:36.600 align:middle line:84% But the problem is that the control of this condition 01:41:36.600 --> 01:41:39.400 align:middle line:90% is very low in many countries. 01:41:39.400 --> 01:41:46.120 align:middle line:84% So treatment adherence for people that have diabetes 01:41:46.120 --> 01:41:49.240 align:middle line:84% is very important, but this treatment adherence 01:41:49.240 --> 01:41:54.200 align:middle line:84% is also unsatisfactory in many countries. 01:41:54.200 --> 01:41:56.120 align:middle line:84% And its treatment adherence depends 01:41:56.120 --> 01:41:59.580 align:middle line:84% on two factors, patients' adherence, 01:41:59.580 --> 01:42:04.360 align:middle line:84% like changing their lifestyle or being aware of their condition 01:42:04.360 --> 01:42:05.900 align:middle line:90% to taking the medication. 01:42:05.900 --> 01:42:08.200 align:middle line:84% And the other factor is the health system 01:42:08.200 --> 01:42:11.320 align:middle line:84% that has an important role to improve 01:42:11.320 --> 01:42:14.040 align:middle line:90% the accessibility of patients. 01:42:14.040 --> 01:42:21.400 align:middle line:84% So in this model, I try to have some reinforcing loop 01:42:21.400 --> 01:42:23.600 align:middle line:84% that treatment adherence influence 01:42:23.600 --> 01:42:25.300 align:middle line:90% on the control of patients. 01:42:25.300 --> 01:42:29.720 align:middle line:84% If this control increases, it reduces 01:42:29.720 --> 01:42:31.740 align:middle line:90% the incidence of complications. 01:42:31.740 --> 01:42:37.630 align:middle line:84% So diabetes is a disease that, if it is not controlled, 01:42:37.630 --> 01:42:40.790 align:middle line:84% it leads to some complicated conditions, 01:42:40.790 --> 01:42:47.530 align:middle line:84% such as cardiovascular diseases or other conditions that 01:42:47.530 --> 01:42:52.210 align:middle line:84% pose more financial difficulties to patients. 01:42:52.210 --> 01:42:55.810 align:middle line:90% So this is the overall problem. 01:42:55.810 --> 01:43:00.050 align:middle line:84% So I have received some feedback from Hashir and another feedback 01:43:00.050 --> 01:43:04.970 align:middle line:84% from Gerardo that is really interesting. 01:43:04.970 --> 01:43:11.650 align:middle line:84% For this model, I try to conceptualize or structure 01:43:11.650 --> 01:43:15.330 align:middle line:84% the progress of patients from the healthy population, 01:43:15.330 --> 01:43:16.710 align:middle line:90% that they are at risk. 01:43:16.710 --> 01:43:20.510 align:middle line:84% So patients, like people over age of 30, 01:43:20.510 --> 01:43:22.730 align:middle line:90% they are at risk of diabetes. 01:43:22.730 --> 01:43:29.790 align:middle line:84% And they might go to the next level of prediabetes. 01:43:29.790 --> 01:43:32.480 align:middle line:84% So prediabetes can be controlled, 01:43:32.480 --> 01:43:36.400 align:middle line:84% and then people can go back to their healthy population. 01:43:36.400 --> 01:43:39.900 align:middle line:84% And then we have the diabetic patients and control 01:43:39.900 --> 01:43:44.160 align:middle line:84% population, or uncontrolled, or population with complications. 01:43:44.160 --> 01:43:49.300 align:middle line:84% And then I'm going to have maybe the other-- 01:43:49.300 --> 01:43:53.860 align:middle line:84% next step of this model is to subscribe the population 01:43:53.860 --> 01:43:56.220 align:middle line:84% with different ages because the age is 01:43:56.220 --> 01:44:02.260 align:middle line:84% important for people that are at different urban areas 01:44:02.260 --> 01:44:03.800 align:middle line:90% or rural areas. 01:44:03.800 --> 01:44:09.340 align:middle line:84% This kind of stuff also affect the patients' situation 01:44:09.340 --> 01:44:12.460 align:middle line:84% to this progress of their disease. 01:44:12.460 --> 01:44:16.520 align:middle line:84% And then I'm going to have the test, 01:44:16.520 --> 01:44:19.300 align:middle line:84% the different interventions, like 01:44:19.300 --> 01:44:24.380 align:middle line:84% if we have the increased awareness, how it influence 01:44:24.380 --> 01:44:26.980 align:middle line:84% on this progress of this disease, 01:44:26.980 --> 01:44:30.390 align:middle line:84% or some other possible interventions that 01:44:30.390 --> 01:44:32.215 align:middle line:90% are more effective. 01:44:32.215 --> 01:44:33.590 align:middle line:84% HAZHIR RAHMANDAD: So, Gerardo, do 01:44:33.590 --> 01:44:36.350 align:middle line:84% you want to share some of your feedback, what 01:44:36.350 --> 01:44:39.510 align:middle line:84% kind of-- at least on the first version of the draft 01:44:39.510 --> 01:44:40.750 align:middle line:90% that you read through? 01:44:40.750 --> 01:44:44.990 align:middle line:84% And I think Sepida highlighted a few more kind of solutions 01:44:44.990 --> 01:44:47.610 align:middle line:84% as well or updates as well, but, yep. 01:44:47.610 --> 01:44:58.650 align:middle line:84% AUDIENCE: [INAUDIBLE] but [INAUDIBLE] really being said. 01:44:58.650 --> 01:45:03.190 align:middle line:84% So, basically, I just say, my feedback 01:45:03.190 --> 01:45:10.990 align:middle line:84% is to improve how the model will be built. 01:45:10.990 --> 01:45:16.190 align:middle line:84% And so apart from the description in the [INAUDIBLE], 01:45:16.190 --> 01:45:19.850 align:middle line:90% step-by-step workflow. 01:45:19.850 --> 01:45:26.130 align:middle line:84% And so with this, just a feedback, one, for example, 01:45:26.130 --> 01:45:27.800 align:middle line:90% is to try to-- 01:45:27.800 --> 01:45:33.700 align:middle line:84% or my mind came up on these perception of importance, 01:45:33.700 --> 01:45:36.160 align:middle line:84% for example, of screening in the population. 01:45:36.160 --> 01:45:43.440 align:middle line:84% So it's sort of a feedback loop related to awareness of disease 01:45:43.440 --> 01:45:48.880 align:middle line:84% and awareness of how the private screening could 01:45:48.880 --> 01:45:50.660 align:middle line:90% be important to this problem. 01:45:50.660 --> 01:45:54.400 align:middle line:84% And so it could be really important feedback loop. 01:45:54.400 --> 01:45:58.440 align:middle line:84% And another one is related to the-- 01:45:58.440 --> 01:46:06.500 align:middle line:84% and e-health device that kind of improve the other-- 01:46:06.500 --> 01:46:07.700 align:middle line:90% the problem. 01:46:07.700 --> 01:46:11.133 align:middle line:84% So I just-- inside a feedback loop to-- 01:46:11.133 --> 01:46:12.300 align:middle line:90% HAZHIR RAHMANDAD: Thank you. 01:46:12.300 --> 01:46:15.880 align:middle line:84% AUDIENCE: [INAUDIBLE] that, to me, 01:46:15.880 --> 01:46:18.720 align:middle line:84% would be-- that's an interesting problem for her. 01:46:18.720 --> 01:46:20.400 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so this is 01:46:20.400 --> 01:46:26.340 align:middle line:84% kind of a pretty clean problem in terms of well-defined space. 01:46:26.340 --> 01:46:28.630 align:middle line:84% It's kind of, diabetes is well known. 01:46:28.630 --> 01:46:33.530 align:middle line:84% And for tackling-- but it is also a complex problem. 01:46:33.530 --> 01:46:35.310 align:middle line:84% There's a lot of research in this space, 01:46:35.310 --> 01:46:37.390 align:middle line:84% so you're not the first one tackling it. 01:46:37.390 --> 01:46:39.830 align:middle line:84% That is good in terms of learning from it. 01:46:39.830 --> 01:46:42.250 align:middle line:84% But also, that is complex for adding, 01:46:42.250 --> 01:46:44.950 align:middle line:84% what is now new that you would be adding? 01:46:44.950 --> 01:46:48.130 align:middle line:84% I think, as we discussed in my feedback 01:46:48.130 --> 01:46:51.930 align:middle line:84% as well, starting with a solid backbone of stock and flows, 01:46:51.930 --> 01:46:55.270 align:middle line:84% for a lot of these health scare problems with patient flows, 01:46:55.270 --> 01:46:56.750 align:middle line:84% this is a very good starting point. 01:46:56.750 --> 01:46:59.650 align:middle line:84% So it's one of those situations that I'm not 01:46:59.650 --> 01:47:02.610 align:middle line:84% too worried about seeing feedback loops explicitly 01:47:02.610 --> 01:47:04.490 align:middle line:90% at the beginning of the project. 01:47:04.490 --> 01:47:07.130 align:middle line:84% I want a solid stock and flow that kind of cuts 01:47:07.130 --> 01:47:10.650 align:middle line:84% the patients in the right groups, 01:47:10.650 --> 01:47:13.755 align:middle line:84% establishes the baseline without any longer term feedback. 01:47:13.755 --> 01:47:15.130 align:middle line:84% There are kind of small feedbacks 01:47:15.130 --> 01:47:16.530 align:middle line:84% in every stock, how it is managed 01:47:16.530 --> 01:47:19.190 align:middle line:84% and how people flow across different states. 01:47:19.190 --> 01:47:22.070 align:middle line:84% But if you get that together nicely, 01:47:22.070 --> 01:47:23.810 align:middle line:84% you have a very solid starting point 01:47:23.810 --> 01:47:28.780 align:middle line:84% for any additional new mechanism that you want to capture. 01:47:28.780 --> 01:47:31.095 align:middle line:84% So that was kind of my main feedback, 01:47:31.095 --> 01:47:32.720 align:middle line:84% and I think there were a couple others. 01:47:32.720 --> 01:47:35.560 align:middle line:84% I think Maria's project has a similar flavor to it. 01:47:35.560 --> 01:47:41.380 align:middle line:84% That solid stock and flow, make that your primary goal. 01:47:41.380 --> 01:47:43.460 align:middle line:84% If you got a chance to add a feedback, 01:47:43.460 --> 01:47:45.880 align:middle line:90% great, but don't force yourself. 01:47:45.880 --> 01:47:47.380 align:middle line:84% Given the limited time you have, you 01:47:47.380 --> 01:47:50.660 align:middle line:84% might just be able to get a good stock and flow 01:47:50.660 --> 01:47:54.940 align:middle line:90% structure working for that. 01:47:54.940 --> 01:48:00.740 align:middle line:84% But I think nice background, but also, good cutting of the space 01:48:00.740 --> 01:48:04.360 align:middle line:84% that you just described, people who are fine, 01:48:04.360 --> 01:48:06.900 align:middle line:84% people who are at risk, prediabetic, then 01:48:06.900 --> 01:48:09.420 align:middle line:84% diabetic with different levels of complication 01:48:09.420 --> 01:48:11.160 align:middle line:90% or not and diagnosed or not. 01:48:11.160 --> 01:48:18.220 align:middle line:84% And that kind of sets the larger state for the later work. 01:48:18.220 --> 01:48:22.360 align:middle line:90% Let's look at Jason's project. 01:48:22.360 --> 01:48:25.120 align:middle line:84% So, Jason, kind of big picture, quick overview. 01:48:25.120 --> 01:48:28.780 align:middle line:84% Then we get a little bit of feedback from Jake and others. 01:48:28.780 --> 01:48:31.060 align:middle line:84% AUDIENCE: Well, this really has changed quite a bit. 01:48:31.060 --> 01:48:32.018 align:middle line:90% HAZHIR RAHMANDAD: Good. 01:48:32.018 --> 01:48:34.900 align:middle line:84% AUDIENCE: So this first iteration 01:48:34.900 --> 01:48:43.260 align:middle line:84% is more of a perspective agenda-type question, 01:48:43.260 --> 01:48:49.660 align:middle line:84% but it's not well-bounded or concrete. 01:48:49.660 --> 01:48:53.140 align:middle line:90% And so what is it about? 01:48:53.140 --> 01:48:58.020 align:middle line:84% Basically, it's about capacity planning 01:48:58.020 --> 01:49:05.060 align:middle line:84% and how airlines are able to evaluate 01:49:05.060 --> 01:49:08.460 align:middle line:84% the quality of long-term decisions 01:49:08.460 --> 01:49:14.540 align:middle line:84% when they make shorter term downstream decisions that 01:49:14.540 --> 01:49:22.150 align:middle line:84% are changing the performance, and also altering 01:49:22.150 --> 01:49:25.190 align:middle line:90% the demand landscape itself. 01:49:25.190 --> 01:49:29.930 align:middle line:84% So this statement-- I don't remember what I said here. 01:49:29.930 --> 01:49:31.775 align:middle line:90% 01:49:31.775 --> 01:49:33.150 align:middle line:84% HAZHIR RAHMANDAD: So I would like 01:49:33.150 --> 01:49:35.630 align:middle line:84% to investigate the impact of planning process 01:49:35.630 --> 01:49:37.290 align:middle line:90% on organizational learning. 01:49:37.290 --> 01:49:39.810 align:middle line:84% Each stage of the process has its own planning horizon 01:49:39.810 --> 01:49:42.870 align:middle line:84% and informational cues, downstream decisions. 01:49:42.870 --> 01:49:45.490 align:middle line:84% So that was kind of the starting point. 01:49:45.490 --> 01:49:47.210 align:middle line:84% Of course, now, this has evolved. 01:49:47.210 --> 01:49:49.490 align:middle line:84% But let's get a feedback on that, kind of. 01:49:49.490 --> 01:49:53.450 align:middle line:84% So Jake, you have shared some nice notes here. 01:49:53.450 --> 01:49:57.270 align:middle line:84% So can you just summarize what you saw here 01:49:57.270 --> 01:49:59.590 align:middle line:90% that could be improved? 01:49:59.590 --> 01:50:03.030 align:middle line:84% AUDIENCE: Yeah, so it sounds like maybe it's 01:50:03.030 --> 01:50:05.830 align:middle line:84% deviated a little bit from sort of looking 01:50:05.830 --> 01:50:10.730 align:middle line:84% at organizational learning as the exact end goal. 01:50:10.730 --> 01:50:18.350 align:middle line:84% But I guess I was curious what the actual planning levers 01:50:18.350 --> 01:50:20.280 align:middle line:84% that you were going to be looking at 01:50:20.280 --> 01:50:23.420 align:middle line:84% to see the impact on overall learning, 01:50:23.420 --> 01:50:27.960 align:middle line:84% like how you know if your planning process is actually 01:50:27.960 --> 01:50:30.920 align:middle line:84% helping, or the adjustments to your planning process 01:50:30.920 --> 01:50:35.640 align:middle line:84% is actually helping or hurting your overall deviations 01:50:35.640 --> 01:50:39.440 align:middle line:90% from matching demand. 01:50:39.440 --> 01:50:44.760 align:middle line:84% And then I think that spelling out some of the types of shocks 01:50:44.760 --> 01:50:47.400 align:middle line:84% that you need to be able to respond-- 01:50:47.400 --> 01:50:50.880 align:middle line:84% to have the model respond to, like COVID, 01:50:50.880 --> 01:50:53.840 align:middle line:84% and like 9/11 sort of specifically, 01:50:53.840 --> 01:51:01.740 align:middle line:84% why planning in those areas is a modelable problem, 01:51:01.740 --> 01:51:06.040 align:middle line:84% even though it seems like it's just like, everything goes to 0, 01:51:06.040 --> 01:51:10.480 align:middle line:84% but spelling out some of those levers and what the expected 01:51:10.480 --> 01:51:12.920 align:middle line:90% behaviors to shocks would be. 01:51:12.920 --> 01:51:14.920 align:middle line:84% HAZHIR RAHMANDAD: Yeah, thank you. 01:51:14.920 --> 01:51:20.370 align:middle line:84% So let's hear from others who had, also, feedback for Jason. 01:51:20.370 --> 01:51:23.410 align:middle line:90% So MA. 01:51:23.410 --> 01:51:24.583 align:middle line:90% AUDIENCE: That's me. 01:51:24.583 --> 01:51:25.750 align:middle line:90% HAZHIR RAHMANDAD: OK, Maria. 01:51:25.750 --> 01:51:28.130 align:middle line:90% Go ahead. 01:51:28.130 --> 01:51:29.610 align:middle line:84% AUDIENCE: I was just trying to be 01:51:29.610 --> 01:51:32.730 align:middle line:84% helpful in terms of thinking about how he could 01:51:32.730 --> 01:51:37.430 align:middle line:84% operationalize the learning and what would be the signal for-- 01:51:37.430 --> 01:51:41.130 align:middle line:84% I wasn't sure if you would want to measure performance 01:51:41.130 --> 01:51:44.890 align:middle line:84% or what would be then the signal for learning. 01:51:44.890 --> 01:51:50.370 align:middle line:84% And yeah, I'm just trying to [INAUDIBLE] start 01:51:50.370 --> 01:51:51.578 align:middle line:90% thinking about that. 01:51:51.578 --> 01:51:52.870 align:middle line:90% HAZHIR RAHMANDAD: That's great. 01:51:52.870 --> 01:51:54.890 align:middle line:84% And I think kind of a specificity and clarity 01:51:54.890 --> 01:51:58.050 align:middle line:84% here is kind of the key feedback that I also 01:51:58.050 --> 01:52:04.490 align:middle line:84% had earlier for Jason, kind of how you get to a more 01:52:04.490 --> 01:52:05.850 align:middle line:90% concrete problem efficient. 01:52:05.850 --> 01:52:07.145 align:middle line:90% And, Ross, I think? 01:52:07.145 --> 01:52:07.770 align:middle line:90% AUDIENCE: Yeah. 01:52:07.770 --> 01:52:09.930 align:middle line:90% HAZHIR RAHMANDAD: You had-- yep. 01:52:09.930 --> 01:52:12.330 align:middle line:84% AUDIENCE: I guess it's pretty similar feedback 01:52:12.330 --> 01:52:15.833 align:middle line:84% about being more concrete about what you're actually modeling. 01:52:15.833 --> 01:52:16.500 align:middle line:90% AUDIENCE: Right. 01:52:16.500 --> 01:52:18.980 align:middle line:84% AUDIENCE: I guess the purpose of the model 01:52:18.980 --> 01:52:23.540 align:middle line:84% is to demonstrate where the learning opportunity is 01:52:23.540 --> 01:52:27.300 align:middle line:84% missed because you're making sort of corrections. 01:52:27.300 --> 01:52:29.020 align:middle line:90% AUDIENCE: My kind of-- 01:52:29.020 --> 01:52:31.540 align:middle line:84% so it's evolved since now, but one 01:52:31.540 --> 01:52:35.740 align:middle line:84% of the ways its evolved is specifying learning about what-- 01:52:35.740 --> 01:52:38.700 align:middle line:84% and for the new iteration, it's learning 01:52:38.700 --> 01:52:42.740 align:middle line:90% about demand environment. 01:52:42.740 --> 01:52:47.880 align:middle line:84% There's other things you could learn, like for example, how-- 01:52:47.880 --> 01:52:49.380 align:middle line:84% there's learning parameters that you 01:52:49.380 --> 01:52:51.980 align:middle line:84% could try to learn about that govern 01:52:51.980 --> 01:52:58.180 align:middle line:84% how quickly each airline learns about the demand environment. 01:52:58.180 --> 01:53:00.420 align:middle line:90% That could be another question. 01:53:00.420 --> 01:53:04.000 align:middle line:84% I think-- what was your question? 01:53:04.000 --> 01:53:04.980 align:middle line:90% Sorry 01:53:04.980 --> 01:53:06.660 align:middle line:84% AUDIENCE: Oh, so is it to demonstrate-- 01:53:06.660 --> 01:53:08.515 align:middle line:90% AUDIENCE: What's the goal? 01:53:08.515 --> 01:53:09.140 align:middle line:90% AUDIENCE: Yeah. 01:53:09.140 --> 01:53:11.280 align:middle line:84% AUDIENCE: Yeah, in the new version, 01:53:11.280 --> 01:53:18.030 align:middle line:84% the goal is to kind of investigate 01:53:18.030 --> 01:53:20.470 align:middle line:90% the consequences of-- 01:53:20.470 --> 01:53:22.590 align:middle line:84% basically, show that there's path dependence 01:53:22.590 --> 01:53:29.710 align:middle line:84% and you can have a good market environment, a good demand 01:53:29.710 --> 01:53:30.750 align:middle line:90% environment. 01:53:30.750 --> 01:53:35.710 align:middle line:84% And when you couple forecasting, which both-- 01:53:35.710 --> 01:53:38.070 align:middle line:84% which two competitors are engaged 01:53:38.070 --> 01:53:39.810 align:middle line:84% in forecasting in that environment, 01:53:39.810 --> 01:53:43.030 align:middle line:84% it can really shift kind of the realized demand 01:53:43.030 --> 01:53:45.750 align:middle line:90% and the performance outcomes. 01:53:45.750 --> 01:53:50.770 align:middle line:84% Just based on the pairing of those forecast sensitivities, 01:53:50.770 --> 01:53:55.642 align:middle line:84% it can change in the same type of demand environment. 01:53:55.642 --> 01:53:57.850 align:middle line:84% AUDIENCE: So you still have the different timescales, 01:53:57.850 --> 01:54:00.230 align:middle line:90% different time frames. 01:54:00.230 --> 01:54:02.230 align:middle line:84% AUDIENCE: The time scale issue of planning 01:54:02.230 --> 01:54:05.390 align:middle line:90% went away in this version, yeah. 01:54:05.390 --> 01:54:08.030 align:middle line:84% There is temporal complexity that I 01:54:08.030 --> 01:54:11.270 align:middle line:84% think will be important because you're 01:54:11.270 --> 01:54:16.910 align:middle line:84% kind of making a forecast, comparing forecasts, and then 01:54:16.910 --> 01:54:17.730 align:middle line:90% doing two things. 01:54:17.730 --> 01:54:19.590 align:middle line:84% You're adjusting your prices, and you're 01:54:19.590 --> 01:54:23.990 align:middle line:84% adjusting your forecast of remaining demand. 01:54:23.990 --> 01:54:27.650 align:middle line:84% Yeah, there's two people doing pricing. 01:54:27.650 --> 01:54:29.650 align:middle line:84% HAZHIR RAHMANDAD: So I take kind of big picture, 01:54:29.650 --> 01:54:32.110 align:middle line:84% everybody had that feedback of, OK, 01:54:32.110 --> 01:54:34.650 align:middle line:84% this is not very clear what the exact problem is. 01:54:34.650 --> 01:54:36.950 align:middle line:84% And, sometimes, these theoretical words, 01:54:36.950 --> 01:54:40.090 align:middle line:84% like organizational learning, temporal complexity, 01:54:40.090 --> 01:54:42.870 align:middle line:84% they're nice for selling the paper, ultimately, 01:54:42.870 --> 01:54:44.650 align:middle line:90% to the academic audiences. 01:54:44.650 --> 01:54:47.150 align:middle line:84% But it's not helpful as a starting point 01:54:47.150 --> 01:54:50.550 align:middle line:84% because it doesn't tie you to a very concrete problem to model 01:54:50.550 --> 01:54:51.690 align:middle line:90% and question to model. 01:54:51.690 --> 01:54:54.270 align:middle line:84% We have talked about these different ways of defining 01:54:54.270 --> 01:54:57.230 align:middle line:84% problems, whether you're trying to explain a reference mode, 01:54:57.230 --> 01:55:01.190 align:middle line:84% you're trying to look at a policy or a few policy levers, 01:55:01.190 --> 01:55:04.030 align:middle line:84% or you are picking a mechanism that you have observed 01:55:04.030 --> 01:55:05.750 align:middle line:84% as kind of a structure of the system 01:55:05.750 --> 01:55:09.480 align:middle line:84% and trying to explore the implications. 01:55:09.480 --> 01:55:11.860 align:middle line:84% Do you now have a better feel which of these three 01:55:11.860 --> 01:55:13.660 align:middle line:90% may best fit what you're doing? 01:55:13.660 --> 01:55:15.258 align:middle line:90% AUDIENCE: Third category. 01:55:15.258 --> 01:55:16.800 align:middle line:84% HAZHIR RAHMANDAD: So, basically, what 01:55:16.800 --> 01:55:19.480 align:middle line:84% are the mechanisms that you are focusing on 01:55:19.480 --> 01:55:21.183 align:middle line:90% to explore their implications? 01:55:21.183 --> 01:55:22.600 align:middle line:84% AUDIENCE: Yeah, the main mechanism 01:55:22.600 --> 01:55:30.080 align:middle line:84% is this process of revenue management, where 01:55:30.080 --> 01:55:34.680 align:middle line:84% you have a flight taking off in a year from now. 01:55:34.680 --> 01:55:37.120 align:middle line:84% And every day you take bookings, you 01:55:37.120 --> 01:55:39.360 align:middle line:90% have to adjust kind of your-- 01:55:39.360 --> 01:55:48.240 align:middle line:84% so it's the mechanism of dynamic forecasting and revenue 01:55:48.240 --> 01:55:50.440 align:middle line:84% management coupled with competition. 01:55:50.440 --> 01:55:54.880 align:middle line:84% And I want to see, when you put those two things together, 01:55:54.880 --> 01:55:55.900 align:middle line:90% how do they interact? 01:55:55.900 --> 01:55:57.300 align:middle line:90% What's the kind of-- 01:55:57.300 --> 01:56:01.040 align:middle line:90% 01:56:01.040 --> 01:56:03.600 align:middle line:84% yeah, I'm interested in that parameter space 01:56:03.600 --> 01:56:07.210 align:middle line:84% when you have different kind of adjustment or nerve sensitivity. 01:56:07.210 --> 01:56:09.910 align:middle line:84% HAZHIR RAHMANDAD: And you have, basically, some hypothesis. 01:56:09.910 --> 01:56:11.330 align:middle line:84% And the hypothesis is, basically, 01:56:11.330 --> 01:56:14.970 align:middle line:84% a connection between a structure and a behavior that you think 01:56:14.970 --> 01:56:16.130 align:middle line:90% may be happening. 01:56:16.130 --> 01:56:18.010 align:middle line:84% And your hypothesis is that, when 01:56:18.010 --> 01:56:23.050 align:middle line:84% you have competition and these kind of continuous adjustment of 01:56:23.050 --> 01:56:25.480 align:middle line:90% pricing and-- 01:56:25.480 --> 01:56:26.443 align:middle line:90% AUDIENCE: It's pricing. 01:56:26.443 --> 01:56:28.610 align:middle line:84% HAZHIR RAHMANDAD: It's largely pricing in this case. 01:56:28.610 --> 01:56:31.985 align:middle line:84% Then you may get some inefficiencies in the system. 01:56:31.985 --> 01:56:32.610 align:middle line:90% AUDIENCE: Yeah. 01:56:32.610 --> 01:56:34.818 align:middle line:84% HAZHIR RAHMANDAD: And that's your hypothesis as well. 01:56:34.818 --> 01:56:37.070 align:middle line:84% So it's a little bit of the first approach as well. 01:56:37.070 --> 01:56:40.490 align:middle line:84% That is, you actually came into this by the feeling 01:56:40.490 --> 01:56:44.410 align:middle line:84% internally that there are some inefficiencies in the planning 01:56:44.410 --> 01:56:47.710 align:middle line:84% among the airlines that, in the bigger picture, 01:56:47.710 --> 01:56:51.450 align:middle line:84% combines pricing and longer term capacity planning. 01:56:51.450 --> 01:56:53.310 align:middle line:84% Then you narrowed it down into, OK, 01:56:53.310 --> 01:56:57.210 align:middle line:84% let me understand the pricing side under competition 01:56:57.210 --> 01:56:59.970 align:middle line:84% to see whether I can see inefficiencies, problems, 01:56:59.970 --> 01:57:03.470 align:middle line:84% that may just come from the standard pricing process, 01:57:03.470 --> 01:57:06.660 align:middle line:84% and then perhaps connect it to the capacity planning. 01:57:06.660 --> 01:57:15.260 align:middle line:84% AUDIENCE: Because there's two kind of time horizons, 01:57:15.260 --> 01:57:19.980 align:middle line:84% and it's too hard to think about them interacting with each other 01:57:19.980 --> 01:57:24.860 align:middle line:84% unless you've figured out how one works by itself. 01:57:24.860 --> 01:57:27.780 align:middle line:84% And then you can, perhaps, have a clear story 01:57:27.780 --> 01:57:32.780 align:middle line:84% about that second kind of longer term capacity planning 01:57:32.780 --> 01:57:33.400 align:middle line:90% discussion. 01:57:33.400 --> 01:57:37.460 align:middle line:84% And then you could couple them together and see what happens. 01:57:37.460 --> 01:57:38.940 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so I think 01:57:38.940 --> 01:57:41.940 align:middle line:84% this was a good example of going through a more 01:57:41.940 --> 01:57:44.780 align:middle line:84% challenging problem definition that started with a feeling 01:57:44.780 --> 01:57:48.300 align:middle line:84% that things are not really doing that well in the airline 01:57:48.300 --> 01:57:50.780 align:middle line:90% industry in some planning space. 01:57:50.780 --> 01:57:54.540 align:middle line:84% So Jason has experience in that space and brought that concern, 01:57:54.540 --> 01:57:58.000 align:middle line:84% but it wasn't a very well-defined problem. 01:57:58.000 --> 01:58:00.040 align:middle line:84% Then he brought it into the question of, OK, 01:58:00.040 --> 01:58:02.430 align:middle line:84% there is this shorter term and longer term time 01:58:02.430 --> 01:58:03.850 align:middle line:90% horizons of planning. 01:58:03.850 --> 01:58:08.150 align:middle line:84% And I think something is not working perfectly there. 01:58:08.150 --> 01:58:09.990 align:middle line:84% And then from there into, OK, let 01:58:09.990 --> 01:58:14.470 align:middle line:84% me understand this shorter term horizon with competition 01:58:14.470 --> 01:58:17.190 align:middle line:84% because I think competition matters here. 01:58:17.190 --> 01:58:23.130 align:middle line:84% So now we shifted from a reference mode type E, 01:58:23.130 --> 01:58:26.390 align:middle line:84% but not a clean reference mode into more of a mechanism 01:58:26.390 --> 01:58:28.550 align:middle line:84% that you need to model, with the idea 01:58:28.550 --> 01:58:32.130 align:middle line:84% that maybe it contributes to a problematic reference mode. 01:58:32.130 --> 01:58:34.250 align:middle line:84% And maybe the result would be that it doesn't. 01:58:34.250 --> 01:58:36.430 align:middle line:90% It's perfectly fine. 01:58:36.430 --> 01:58:39.430 align:middle line:84% AUDIENCE: That part of it is I'm used to-- 01:58:39.430 --> 01:58:43.190 align:middle line:84% part of my challenge with this exercise is like, 01:58:43.190 --> 01:58:45.430 align:middle line:84% do I need to know in advance that it's 01:58:45.430 --> 01:58:51.030 align:middle line:84% going to yield a problem or give a specific behavior? 01:58:51.030 --> 01:58:53.030 align:middle line:84% If you did know that, then why model it? 01:58:53.030 --> 01:58:54.113 align:middle line:90% HAZHIR RAHMANDAD: Exactly. 01:58:54.113 --> 01:58:55.850 align:middle line:84% So the answer to that is always no. 01:58:55.850 --> 01:58:57.470 align:middle line:84% You don't need to know that it will 01:58:57.470 --> 01:58:59.050 align:middle line:90% create a problematic behavior. 01:58:59.050 --> 01:59:02.063 align:middle line:84% Otherwise, why would I build the model in the first place? 01:59:02.063 --> 01:59:03.980 align:middle line:84% AUDIENCE: So then what is a problem statement? 01:59:03.980 --> 01:59:10.600 align:middle line:84% The problem statement is one of these three categories, where 01:59:10.600 --> 01:59:16.240 align:middle line:84% you're observing that there's some conflict or tension, 01:59:16.240 --> 01:59:19.175 align:middle line:90% potential conflict or tension. 01:59:19.175 --> 01:59:20.800 align:middle line:84% HAZHIR RAHMANDAD: Yeah, and in a sense, 01:59:20.800 --> 01:59:23.740 align:middle line:84% it doesn't need to be any of them exclusively. 01:59:23.740 --> 01:59:26.560 align:middle line:84% It is kind of a mental conversation 01:59:26.560 --> 01:59:29.320 align:middle line:84% between these different ways of anchoring what 01:59:29.320 --> 01:59:30.660 align:middle line:90% you're doing in the modeling. 01:59:30.660 --> 01:59:32.640 align:middle line:84% So one way is to think of it, I'm 01:59:32.640 --> 01:59:34.420 align:middle line:84% trying to explain this mode of behavior. 01:59:34.420 --> 01:59:37.240 align:middle line:84% But in this case, you don't have a very clean mode of behavior 01:59:37.240 --> 01:59:40.560 align:middle line:84% that you can explain with certainty because you don't have 01:59:40.560 --> 01:59:44.800 align:middle line:84% actual evidence of significant inefficiencies, 01:59:44.800 --> 01:59:47.600 align:middle line:84% that you can come and show me and convince me 01:59:47.600 --> 01:59:48.960 align:middle line:90% that there is some positive-- 01:59:48.960 --> 01:59:54.327 align:middle line:84% Cathy's problem does have a more clear kind of behavior over-- 01:59:54.327 --> 01:59:56.160 align:middle line:84% AUDIENCE: Just don't make sense [INAUDIBLE]. 01:59:56.160 --> 01:59:57.220 align:middle line:84% HAZHIR RAHMANDAD: Just, in the first order, 01:59:57.220 --> 01:59:58.137 align:middle line:90% they don't make sense. 01:59:58.137 --> 02:00:01.730 align:middle line:90% So now everybody is hooked. 02:00:01.730 --> 02:00:04.770 align:middle line:84% But in your case, you see, I have this nagging feeling 02:00:04.770 --> 02:00:06.710 align:middle line:90% that there is something. 02:00:06.710 --> 02:00:10.390 align:middle line:84% And then the mechanism behind it, let me talk about mechanism. 02:00:10.390 --> 02:00:13.530 align:middle line:84% So now it is bringing you to the space of mechanisms. 02:00:13.530 --> 02:00:16.610 align:middle line:84% And you can anchor it now to the space of mechanisms 02:00:16.610 --> 02:00:18.270 align:middle line:84% and defining the problem around that. 02:00:18.270 --> 02:00:20.450 align:middle line:90% And that's perfectly fine. 02:00:20.450 --> 02:00:23.290 align:middle line:84% But you can see these different problem 02:00:23.290 --> 02:00:25.370 align:middle line:84% definition approaches are emerging out 02:00:25.370 --> 02:00:27.403 align:middle line:84% of these problem definitions that you guys have 02:00:27.403 --> 02:00:28.070 align:middle line:90% been working on. 02:00:28.070 --> 02:00:30.170 align:middle line:84% And that's exactly what we want to have. 02:00:30.170 --> 02:00:35.170 align:middle line:84% So Sercin's problem definition is-- 02:00:35.170 --> 02:00:38.110 align:middle line:84% actually, he's not here today, but-- 02:00:38.110 --> 02:00:40.990 align:middle line:90% 02:00:40.990 --> 02:00:44.558 align:middle line:90% AUDIENCE: [INAUDIBLE] 02:00:44.558 --> 02:00:45.850 align:middle line:90% HAZHIR RAHMANDAD: Oh, yeah, OK. 02:00:45.850 --> 02:00:48.550 align:middle line:84% So aggregate business behavior under uncertainty, 02:00:48.550 --> 02:00:52.812 align:middle line:84% endogenous sentiments, macroeconomic-- 02:00:52.812 --> 02:00:53.770 align:middle line:90% AUDIENCE: Fluctuations. 02:00:53.770 --> 02:00:56.210 align:middle line:90% HAZHIR RAHMANDAD: Fluctuations. 02:00:56.210 --> 02:01:01.590 align:middle line:84% This is actually a pretty broad problem. 02:01:01.590 --> 02:01:04.730 align:middle line:90% 02:01:04.730 --> 02:01:09.650 align:middle line:84% So this is kind of focused on an academic audience. 02:01:09.650 --> 02:01:17.990 align:middle line:84% And I think we have a few feedback items. 02:01:17.990 --> 02:01:30.250 align:middle line:84% So we have-- who wants to go first? 02:01:30.250 --> 02:01:33.110 align:middle line:84% Jake, do you want to share your feedback? 02:01:33.110 --> 02:01:35.745 align:middle line:90% And then Maria, and then Ross? 02:01:35.745 --> 02:01:36.370 align:middle line:90% AUDIENCE: Sure. 02:01:36.370 --> 02:01:39.730 align:middle line:90% 02:01:39.730 --> 02:01:41.870 align:middle line:84% So I think I found it interesting. 02:01:41.870 --> 02:01:46.450 align:middle line:84% I got a little bit confused kind of in the distinction 02:01:46.450 --> 02:01:48.130 align:middle line:84% he's making between risk and uncertainty 02:01:48.130 --> 02:01:51.010 align:middle line:84% versus this idea of uncertainty and sentiment 02:01:51.010 --> 02:01:53.170 align:middle line:90% that's being modeled. 02:01:53.170 --> 02:01:56.300 align:middle line:84% And, particularly, I guess it's being applied 02:01:56.300 --> 02:02:02.000 align:middle line:84% in the context of operations and sort of buffer planning, 02:02:02.000 --> 02:02:05.860 align:middle line:84% I guess, is the idea or what I took away from it. 02:02:05.860 --> 02:02:08.740 align:middle line:84% And it mentions as well that you're 02:02:08.740 --> 02:02:12.740 align:middle line:84% going to be working in both banks 02:02:12.740 --> 02:02:17.100 align:middle line:84% and household-like reactions to everything, as well 02:02:17.100 --> 02:02:19.140 align:middle line:90% as the firms. 02:02:19.140 --> 02:02:24.080 align:middle line:84% And it just feels generally very broad to me. 02:02:24.080 --> 02:02:25.300 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 02:02:25.300 --> 02:02:30.220 align:middle line:84% AUDIENCE: But so I guess I'm unsure where the crux is. 02:02:30.220 --> 02:02:33.300 align:middle line:84% Is it monitoring operations within-- 02:02:33.300 --> 02:02:36.320 align:middle line:84% operations buffer planning with sentiment? 02:02:36.320 --> 02:02:38.372 align:middle line:90% Is that the idea? 02:02:38.372 --> 02:02:40.080 align:middle line:84% HAZHIR RAHMANDAD: That's a good question. 02:02:40.080 --> 02:02:41.622 align:middle line:84% And I think, kind of, it is something 02:02:41.622 --> 02:02:46.620 align:middle line:84% that Sercin needs to answer because, basically, he 02:02:46.620 --> 02:02:49.700 align:middle line:84% starts with a very fundamental question of, 02:02:49.700 --> 02:02:52.100 align:middle line:84% there is uncertainty, there is risk, 02:02:52.100 --> 02:02:54.030 align:middle line:84% and there is buffering against those 02:02:54.030 --> 02:02:56.590 align:middle line:84% that households and businesses do. 02:02:56.590 --> 02:03:02.190 align:middle line:84% And how that buffering against risk may happen dynamically 02:03:02.190 --> 02:03:03.930 align:middle line:90% is kind of the problem space. 02:03:03.930 --> 02:03:08.930 align:middle line:84% But it is less clear exactly how you narrow that down 02:03:08.930 --> 02:03:10.710 align:middle line:84% because that becomes, potentially, 02:03:10.710 --> 02:03:12.530 align:middle line:90% modeling the whole economy. 02:03:12.530 --> 02:03:17.430 align:middle line:84% And that's way too complex a problem to tackle, 02:03:17.430 --> 02:03:21.210 align:middle line:84% even too ill-defined a problem to tackle in the first place. 02:03:21.210 --> 02:03:22.970 align:middle line:84% Modeling the economy is not enough. 02:03:22.970 --> 02:03:25.170 align:middle line:90% It's just too general. 02:03:25.170 --> 02:03:27.910 align:middle line:90% You need a problem within that. 02:03:27.910 --> 02:03:28.830 align:middle line:90% Yep. 02:03:28.830 --> 02:03:32.250 align:middle line:84% Maria, do you want to share any thoughts you had? 02:03:32.250 --> 02:03:37.910 align:middle line:90% 02:03:37.910 --> 02:03:40.790 align:middle line:84% AUDIENCE: Sorry, I forgot to unmute myself. 02:03:40.790 --> 02:03:43.890 align:middle line:84% So kind of similar to what you said, 02:03:43.890 --> 02:03:48.550 align:middle line:84% I thought it was very broad still 02:03:48.550 --> 02:03:50.810 align:middle line:84% and written for an academic audience, 02:03:50.810 --> 02:03:54.600 align:middle line:84% so kind of very far from becoming a modeling 02:03:54.600 --> 02:03:55.900 align:middle line:90% problem at this stage. 02:03:55.900 --> 02:03:59.900 align:middle line:84% So I started from the basics, let's say. 02:03:59.900 --> 02:04:03.020 align:middle line:84% So what would be the intended use of the model? 02:04:03.020 --> 02:04:06.000 align:middle line:84% Would it be to explain the reference mode, 02:04:06.000 --> 02:04:08.840 align:middle line:84% assess policy interventions, or find 02:04:08.840 --> 02:04:12.560 align:middle line:90% the implications of mechanisms? 02:04:12.560 --> 02:04:19.000 align:middle line:84% Yeah, is it for theory building or sensitivity exploration? 02:04:19.000 --> 02:04:23.120 align:middle line:84% Yeah, something about the model boundary 02:04:23.120 --> 02:04:27.240 align:middle line:84% as well could have been useful, things like that. 02:04:27.240 --> 02:04:29.680 align:middle line:84% Yeah, kind of going to the basics, I would say. 02:04:29.680 --> 02:04:31.320 align:middle line:84% HAZHIR RAHMANDAD: Yeah, now, so I 02:04:31.320 --> 02:04:34.680 align:middle line:84% think kind of in the second part, towards the end, 02:04:34.680 --> 02:04:38.380 align:middle line:84% it tries to become more specific, which is very helpful. 02:04:38.380 --> 02:04:42.420 align:middle line:84% So Sercin is highlighting that you have agents. 02:04:42.420 --> 02:04:45.320 align:middle line:84% Each agent is represented with a balance sheet. 02:04:45.320 --> 02:04:49.490 align:middle line:84% They have fund and goods that are 02:04:49.490 --> 02:04:54.490 align:middle line:84% flowing between different agents that follow the accounting 02:04:54.490 --> 02:04:56.370 align:middle line:90% principles. 02:04:56.370 --> 02:04:59.310 align:middle line:84% So it kind of gives a big picture view. 02:04:59.310 --> 02:05:01.950 align:middle line:84% In fact, the agents are divided into households. 02:05:01.950 --> 02:05:05.770 align:middle line:84% So they may buy goods, which means you give money to somebody 02:05:05.770 --> 02:05:07.450 align:middle line:90% and get goods from them. 02:05:07.450 --> 02:05:10.310 align:middle line:84% You have firms that are selling goods, basically, 02:05:10.310 --> 02:05:11.850 align:middle line:90% and producing goods. 02:05:11.850 --> 02:05:14.810 align:middle line:84% You have banks that lend money to people. 02:05:14.810 --> 02:05:18.890 align:middle line:84% But if you try to build all of these agents, 02:05:18.890 --> 02:05:22.150 align:middle line:84% this is really a full economy model. 02:05:22.150 --> 02:05:24.690 align:middle line:84% There are models like this that people build. 02:05:24.690 --> 02:05:29.610 align:middle line:84% I'm not saying you can't do it, but it is pretty ambitious. 02:05:29.610 --> 02:05:32.530 align:middle line:84% Sources of inertia, it's capital stock 02:05:32.530 --> 02:05:34.710 align:middle line:84% for firms, inventories of the firms. 02:05:34.710 --> 02:05:37.450 align:middle line:84% So if you're accounting for inventories in these dynamics, 02:05:37.450 --> 02:05:40.710 align:middle line:84% that's yet another layer of complexity. 02:05:40.710 --> 02:05:45.490 align:middle line:84% It's kind of, you're going into more detail for businesses. 02:05:45.490 --> 02:05:49.220 align:middle line:84% Liabilities, cash is stock, and then you 02:05:49.220 --> 02:05:54.540 align:middle line:84% have sentiment, which is a whole different space. 02:05:54.540 --> 02:05:58.340 align:middle line:84% And then reference modes, income, 02:05:58.340 --> 02:06:02.900 align:middle line:84% sentiments, potential oscillations, 02:06:02.900 --> 02:06:06.140 align:middle line:84% sensitivity of parameters of the real sector with respect 02:06:06.140 --> 02:06:08.800 align:middle line:90% to sentiments and so on. 02:06:08.800 --> 02:06:11.020 align:middle line:84% Ross, do you want to add something? 02:06:11.020 --> 02:06:13.500 align:middle line:84% AUDIENCE: Yeah, I guess it's similar in the sense that 02:06:13.500 --> 02:06:17.720 align:middle line:84% simplification, maybe you can use household demand instead 02:06:17.720 --> 02:06:21.660 align:middle line:84% of household-- or interest rate instead of banks, 02:06:21.660 --> 02:06:24.740 align:middle line:84% and simplify from capital [INAUDIBLE] 02:06:24.740 --> 02:06:27.020 align:middle line:90% to assets and liabilities. 02:06:27.020 --> 02:06:29.460 align:middle line:84% Another question was like, sentiment 02:06:29.460 --> 02:06:31.980 align:middle line:84% sounds like something irrational, 02:06:31.980 --> 02:06:35.300 align:middle line:84% in that it is a function of market signals, 02:06:35.300 --> 02:06:37.660 align:middle line:90% like growth and volatility. 02:06:37.660 --> 02:06:41.020 align:middle line:84% I was curious how, in general, someone would 02:06:41.020 --> 02:06:43.580 align:middle line:90% model something irrational. 02:06:43.580 --> 02:06:45.900 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so this is-- 02:06:45.900 --> 02:06:48.740 align:middle line:84% now, Sercin is an economist, and so he has experience 02:06:48.740 --> 02:06:50.320 align:middle line:90% working on these problems. 02:06:50.320 --> 02:06:53.220 align:middle line:84% So if somebody without experience 02:06:53.220 --> 02:06:54.900 align:middle line:84% would have come and proposed this, 02:06:54.900 --> 02:06:57.920 align:middle line:84% I would say, no, don't even get us started. 02:06:57.920 --> 02:07:00.500 align:middle line:84% But this is kind of what-- he has spent enough time 02:07:00.500 --> 02:07:02.440 align:middle line:90% to know the lay of the land. 02:07:02.440 --> 02:07:05.080 align:middle line:84% And there is kind of good modeling 02:07:05.080 --> 02:07:08.260 align:middle line:84% that is a starting point he can draw upon. 02:07:08.260 --> 02:07:13.140 align:middle line:90% 02:07:13.140 --> 02:07:16.420 align:middle line:84% But still, without a more clear problem definition 02:07:16.420 --> 02:07:18.520 align:middle line:84% of what you want to do with this model, 02:07:18.520 --> 02:07:21.180 align:middle line:84% it is a little bit harder to say, what is in? 02:07:21.180 --> 02:07:22.300 align:middle line:90% And what is out? 02:07:22.300 --> 02:07:25.260 align:middle line:84% So is it a prediction tool at the end of the day? 02:07:25.260 --> 02:07:28.180 align:middle line:84% Is it a new theory development tool at the end of the day? 02:07:28.180 --> 02:07:30.660 align:middle line:84% That you would say, OK, nobody has 02:07:30.660 --> 02:07:37.020 align:middle line:84% looked at these alternative ways of projection 02:07:37.020 --> 02:07:39.700 align:middle line:84% into the future of opportunities to invest 02:07:39.700 --> 02:07:42.900 align:middle line:90% or opportunities to consume. 02:07:42.900 --> 02:07:46.910 align:middle line:84% And this alternative way of future projections 02:07:46.910 --> 02:07:50.970 align:middle line:84% makes a big difference into how we understand the economy. 02:07:50.970 --> 02:07:57.630 align:middle line:84% So that's kind of going to this expectations or sentiment 02:07:57.630 --> 02:07:59.990 align:middle line:90% question. 02:07:59.990 --> 02:08:02.630 align:middle line:84% Then you need to narrow down, what exactly 02:08:02.630 --> 02:08:07.270 align:middle line:84% is the sentiment of space, expectation space, that has not 02:08:07.270 --> 02:08:10.390 align:middle line:84% been explored, that you are bringing in a new formulation 02:08:10.390 --> 02:08:12.550 align:middle line:90% there? 02:08:12.550 --> 02:08:15.110 align:middle line:84% It might be-- so that's the type of question 02:08:15.110 --> 02:08:19.230 align:middle line:84% that I've been asking, but there is also 02:08:19.230 --> 02:08:22.710 align:middle line:84% work that does basics of this machinery. 02:08:22.710 --> 02:08:24.370 align:middle line:84% So you can actually draw upon it. 02:08:24.370 --> 02:08:27.515 align:middle line:84% There is agent-based modeling that has done this. 02:08:27.515 --> 02:08:29.890 align:middle line:84% There's kind of history in system dynamics of doing this. 02:08:29.890 --> 02:08:33.430 align:middle line:84% There are these models of macroeconomic behavior 02:08:33.430 --> 02:08:37.110 align:middle line:84% with firms and households in those models, 02:08:37.110 --> 02:08:38.950 align:middle line:90% potentially adding banks. 02:08:38.950 --> 02:08:42.460 align:middle line:84% It's can become very complex, very quickly. 02:08:42.460 --> 02:08:44.900 align:middle line:90% So I'm worried about that. 02:08:44.900 --> 02:08:48.680 align:middle line:84% But, otherwise, the definition of stocks 02:08:48.680 --> 02:08:54.440 align:middle line:84% flows ideas are a solid space to explore. 02:08:54.440 --> 02:08:58.340 align:middle line:84% So let's see what else we have here. 02:08:58.340 --> 02:09:02.820 align:middle line:84% So we have feedback from Hannah for Oriana. 02:09:02.820 --> 02:09:08.573 align:middle line:84% So let's take a quick look at Oriana's problem. 02:09:08.573 --> 02:09:10.240 align:middle line:84% So this is the first version, and Oriana 02:09:10.240 --> 02:09:11.620 align:middle line:90% has updated hers as well. 02:09:11.620 --> 02:09:13.400 align:middle line:84% So do you want to give a quick overview 02:09:13.400 --> 02:09:15.360 align:middle line:90% of what you were interested in? 02:09:15.360 --> 02:09:17.160 align:middle line:84% AUDIENCE: So I'll only go over proposal one 02:09:17.160 --> 02:09:20.640 align:middle line:84% because that was already pretty convoluted. 02:09:20.640 --> 02:09:23.980 align:middle line:84% And then if I pare it down even further to clothing brands, 02:09:23.980 --> 02:09:28.080 align:middle line:84% I was thinking about the ways that the secondhand market 02:09:28.080 --> 02:09:29.720 align:middle line:84% interacts with the primary market 02:09:29.720 --> 02:09:33.420 align:middle line:84% when it comes to that because, depending on what the brand is, 02:09:33.420 --> 02:09:36.400 align:middle line:84% let's say, take, for example, Hermes versus Supreme. 02:09:36.400 --> 02:09:38.400 align:middle line:84% For Supreme, the secondhand market, 02:09:38.400 --> 02:09:40.570 align:middle line:84% it being sold out actually increases 02:09:40.570 --> 02:09:43.690 align:middle line:84% the prestige and hype around the brand versus if you, say, look 02:09:43.690 --> 02:09:46.530 align:middle line:84% at something like Hermes, being on the secondhand market 02:09:46.530 --> 02:09:48.690 align:middle line:84% and falling into the wrong kind of consumer 02:09:48.690 --> 02:09:50.710 align:middle line:84% actually destroys brand prestige. 02:09:50.710 --> 02:09:52.210 align:middle line:90% And that's its own problem. 02:09:52.210 --> 02:09:55.690 align:middle line:84% But the feedback from both Kajir and Hannah 02:09:55.690 --> 02:09:58.090 align:middle line:84% was that these are way too broad and I need 02:09:58.090 --> 02:10:01.030 align:middle line:90% to focus on just one thing. 02:10:01.030 --> 02:10:03.610 align:middle line:84% So in the updated proposal, I'm focusing specifically 02:10:03.610 --> 02:10:05.210 align:middle line:90% on Supreme. 02:10:05.210 --> 02:10:07.890 align:middle line:84% And so it's like, Supreme doesn't actually 02:10:07.890 --> 02:10:11.570 align:middle line:84% offer any of physical, better product. 02:10:11.570 --> 02:10:13.530 align:middle line:84% The t-shirt you buy at Supreme is just 02:10:13.530 --> 02:10:16.190 align:middle line:84% like a Hanes t-shirt with the Supreme logo on it. 02:10:16.190 --> 02:10:18.270 align:middle line:84% So they're not actually competing on quality. 02:10:18.270 --> 02:10:21.050 align:middle line:84% So what they are competing on is like prestige, 02:10:21.050 --> 02:10:23.250 align:middle line:84% like clout and cultural relevance. 02:10:23.250 --> 02:10:25.530 align:middle line:84% But an interesting issue that comes 02:10:25.530 --> 02:10:28.050 align:middle line:84% from that, which is the people that made Supreme culturally 02:10:28.050 --> 02:10:29.670 align:middle line:90% relevant, are skaters. 02:10:29.670 --> 02:10:33.970 align:middle line:84% Skaters aren't trying to spend $700 on a t-shirt. 02:10:33.970 --> 02:10:37.110 align:middle line:84% But the reason Supreme has the prestige it does 02:10:37.110 --> 02:10:38.860 align:middle line:84% and it sells out so quickly because, 02:10:38.860 --> 02:10:41.900 align:middle line:84% on the secondhand market, people will buy Supreme and resell it 02:10:41.900 --> 02:10:43.300 align:middle line:90% for $700. 02:10:43.300 --> 02:10:45.100 align:middle line:84% And so there the problem becomes, 02:10:45.100 --> 02:10:46.740 align:middle line:90% Supreme is losing out on money. 02:10:46.740 --> 02:10:49.880 align:middle line:84% If retail is $100 and it's getting resold for $700, 02:10:49.880 --> 02:10:52.200 align:middle line:84% that's $600 of profit that they've lost. 02:10:52.200 --> 02:10:54.140 align:middle line:84% But without the secondhand market, 02:10:54.140 --> 02:10:57.460 align:middle line:84% they lose a bunch of that prestige that 02:10:57.460 --> 02:11:01.340 align:middle line:84% allows them to sell out in the first place, which is also 02:11:01.340 --> 02:11:03.620 align:middle line:84% mixed with, they still maybe need 02:11:03.620 --> 02:11:05.693 align:middle line:84% to keep their cultural relevance because, 02:11:05.693 --> 02:11:07.860 align:middle line:84% without the skaters and the people who actually made 02:11:07.860 --> 02:11:09.560 align:middle line:84% the brand cool in the first place, 02:11:09.560 --> 02:11:11.700 align:middle line:84% being able to wear the brand, then 02:11:11.700 --> 02:11:13.540 align:middle line:84% you just get filled with all these, 02:11:13.540 --> 02:11:17.440 align:middle line:84% I guess, quote, unquote, posers wearing Supreme. 02:11:17.440 --> 02:11:19.160 align:middle line:90% And then it becomes very uncool. 02:11:19.160 --> 02:11:22.500 align:middle line:84% And then they, for example, no longer sell out in minutes. 02:11:22.500 --> 02:11:25.600 align:middle line:84% This was like a brief period of, let's say, four to five years. 02:11:25.600 --> 02:11:29.660 align:middle line:84% And so looking at the ways that the secondhand market 02:11:29.660 --> 02:11:33.100 align:middle line:84% and the wealth disparity interacts with both profits 02:11:33.100 --> 02:11:35.380 align:middle line:84% and cultural relevance for a brand like Supreme 02:11:35.380 --> 02:11:40.550 align:middle line:84% that really is only fighting on prestige and cultural relevance 02:11:40.550 --> 02:11:42.330 align:middle line:84% and no actual indicators of quality, 02:11:42.330 --> 02:11:46.310 align:middle line:90% unlike other luxury firms. 02:11:46.310 --> 02:11:48.230 align:middle line:84% HAZHIR RAHMANDAD: Yeah, go ahead. 02:11:48.230 --> 02:11:51.430 align:middle line:84% AUDIENCE: I was thinking about how 02:11:51.430 --> 02:11:54.870 align:middle line:84% this might be analogous to some parts of ticket scalping 02:11:54.870 --> 02:11:58.150 align:middle line:84% and how, if you're buying the resold stuff 02:11:58.150 --> 02:12:00.630 align:middle line:84% and-- you wouldn't want to brag like, oh, yeah, 02:12:00.630 --> 02:12:02.230 align:middle line:84% I bought this on the secondhand market 02:12:02.230 --> 02:12:06.550 align:middle line:84% and paid $700 for the Hanes t-shirt. 02:12:06.550 --> 02:12:09.450 align:middle line:84% And if you're buying a scalped ticket, 02:12:09.450 --> 02:12:12.923 align:middle line:84% similarly, you're not bragging about buying 02:12:12.923 --> 02:12:14.590 align:middle line:84% the more expensive version of the ticket 02:12:14.590 --> 02:12:19.470 align:middle line:84% or supporting the people who are just buying to resell it. 02:12:19.470 --> 02:12:23.270 align:middle line:84% And I wonder, yeah, if you have any thoughts on the people who 02:12:23.270 --> 02:12:28.350 align:middle line:84% are just getting in the queue to buy it up really fast, just 02:12:28.350 --> 02:12:32.850 align:middle line:84% so that they can sell it at the increased markup and how that-- 02:12:32.850 --> 02:12:35.510 align:middle line:84% AUDIENCE: How does selling out impact a-- 02:12:35.510 --> 02:12:38.390 align:middle line:84% AUDIENCE: Yeah, but then people's perception of, oh, 02:12:38.390 --> 02:12:40.510 align:middle line:90% why did they sell out? 02:12:40.510 --> 02:12:43.330 align:middle line:84% It would shift from, oh, because it's very popular 02:12:43.330 --> 02:12:47.150 align:middle line:84% and everybody who bought this wants to wear it to, oh, this 02:12:47.150 --> 02:12:50.510 align:middle line:84% is kind of like a washed out or saturated brand, 02:12:50.510 --> 02:12:56.150 align:middle line:84% and it's just all of these resellers buying it up and then 02:12:56.150 --> 02:12:59.030 align:middle line:84% selling it at a super high margin. 02:12:59.030 --> 02:13:04.360 align:middle line:84% AUDIENCE: Yeah, I feel like they're similar in the sense of, 02:13:04.360 --> 02:13:07.510 align:middle line:84% it's kind of that idea of, oh, it's 02:13:07.510 --> 02:13:09.470 align:middle line:84% all these posers or people who don't really 02:13:09.470 --> 02:13:10.810 align:middle line:90% like the music getting it. 02:13:10.810 --> 02:13:13.060 align:middle line:84% But I think maybe it's slightly different because what 02:13:13.060 --> 02:13:16.310 align:middle line:84% I'm thinking with artists, let's say a musician, 02:13:16.310 --> 02:13:19.430 align:middle line:84% is like, at the end of the day, anyone can listen to the music. 02:13:19.430 --> 02:13:21.830 align:middle line:84% So you're still going to have fans coming just 02:13:21.830 --> 02:13:25.615 align:middle line:84% from enjoying the albums that they're putting out versus, 02:13:25.615 --> 02:13:26.990 align:middle line:84% with Supreme, if no one's wearing 02:13:26.990 --> 02:13:29.350 align:middle line:84% the clothes, the skaters can't access 02:13:29.350 --> 02:13:33.200 align:middle line:84% the clothes, which is actually the draw, let's say, to Supreme. 02:13:33.200 --> 02:13:35.320 align:middle line:84% So if you completely cut off their relationship 02:13:35.320 --> 02:13:40.680 align:middle line:84% to the clothes, I think there is an actual collapse 02:13:40.680 --> 02:13:42.960 align:middle line:84% of the prestige of the brand versus, 02:13:42.960 --> 02:13:46.880 align:middle line:84% if Taylor Swift concerts were pretty much only 02:13:46.880 --> 02:13:49.840 align:middle line:84% like super wealthy millionaires and billionaires who 02:13:49.840 --> 02:13:51.480 align:middle line:84% could see the show, and then there's 02:13:51.480 --> 02:13:54.200 align:middle line:84% 10 people in the nosebleeds, I think 02:13:54.200 --> 02:13:56.520 align:middle line:84% she can still retain a lot of her audience, 02:13:56.520 --> 02:13:58.920 align:middle line:84% purely because the actual music she's selling 02:13:58.920 --> 02:14:02.560 align:middle line:84% and the reason she's popular will 02:14:02.560 --> 02:14:04.620 align:middle line:84% continue to remain available for everyone. 02:14:04.620 --> 02:14:05.995 align:middle line:84% AUDIENCE: Yeah, that makes sense, 02:14:05.995 --> 02:14:07.330 align:middle line:90% like multi-channel versus-- 02:14:07.330 --> 02:14:07.480 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 02:14:07.480 --> 02:14:08.105 align:middle line:90% AUDIENCE: Yeah. 02:14:08.105 --> 02:14:11.300 align:middle line:90% 02:14:11.300 --> 02:14:13.760 align:middle line:84% HAZHIR RAHMANDAD: Other thoughts? 02:14:13.760 --> 02:14:16.960 align:middle line:84% You wanted to say something, Jason. 02:14:16.960 --> 02:14:21.560 align:middle line:84% AUDIENCE: You reminded me in that there's 02:14:21.560 --> 02:14:23.140 align:middle line:90% lots of questions in here. 02:14:23.140 --> 02:14:30.950 align:middle line:84% And I think it could benefit from taking a slice 02:14:30.950 --> 02:14:39.715 align:middle line:84% or perhaps moving slightly adjacent to the original thing 02:14:39.715 --> 02:14:41.090 align:middle line:84% that is interesting to you, which 02:14:41.090 --> 02:14:47.410 align:middle line:84% is this question about the properties of the products 02:14:47.410 --> 02:14:49.850 align:middle line:90% and the connection. 02:14:49.850 --> 02:14:54.030 align:middle line:84% You're asking this question about prestige and kind 02:14:54.030 --> 02:14:57.970 align:middle line:84% of losing contact with the original population the thing is 02:14:57.970 --> 02:15:01.450 align:middle line:84% designed for, which is what's motivating you. 02:15:01.450 --> 02:15:06.770 align:middle line:84% But maybe there's a smaller, more concrete question 02:15:06.770 --> 02:15:07.790 align:middle line:90% you can ask. 02:15:07.790 --> 02:15:11.330 align:middle line:90% 02:15:11.330 --> 02:15:13.170 align:middle line:90% What that is-- 02:15:13.170 --> 02:15:16.930 align:middle line:84% One place to explore would be like a question 02:15:16.930 --> 02:15:18.930 align:middle line:90% about market failures. 02:15:18.930 --> 02:15:22.370 align:middle line:84% Or not market failure, but, basically, 02:15:22.370 --> 02:15:28.620 align:middle line:84% you could ask the question, it seems inefficient 02:15:28.620 --> 02:15:31.080 align:middle line:84% that there's this huge secondary market. 02:15:31.080 --> 02:15:33.900 align:middle line:84% Why isn't the primary market able 02:15:33.900 --> 02:15:36.040 align:middle line:90% to capture all of that revenue? 02:15:36.040 --> 02:15:41.260 align:middle line:84% And in what ways do they rely on the secondary market 02:15:41.260 --> 02:15:43.720 align:middle line:90% to sustain the primary? 02:15:43.720 --> 02:15:46.020 align:middle line:84% So just very focused on the relationship 02:15:46.020 --> 02:15:49.460 align:middle line:84% between primary and secondary market. 02:15:49.460 --> 02:15:51.660 align:middle line:84% AUDIENCE: As in like cutting out, 02:15:51.660 --> 02:15:53.860 align:middle line:84% let's say, for Supreme, the skater group entirely? 02:15:53.860 --> 02:15:54.800 align:middle line:90% Or what do you mean? 02:15:54.800 --> 02:15:57.067 align:middle line:90% 02:15:57.067 --> 02:15:58.400 align:middle line:90% AUDIENCE: Maybe this is just me. 02:15:58.400 --> 02:16:00.460 align:middle line:90% This is what I thought about it. 02:16:00.460 --> 02:16:03.860 align:middle line:84% In the ticket sale space, if the retail price 02:16:03.860 --> 02:16:06.780 align:middle line:84% is $100 for a ticket, but then on Ticketmaster 02:16:06.780 --> 02:16:08.420 align:middle line:84% or the secondary platform, people 02:16:08.420 --> 02:16:10.540 align:middle line:90% are willing to pay $800 or-- 02:16:10.540 --> 02:16:13.700 align:middle line:84% you can sell them for a lot more. 02:16:13.700 --> 02:16:17.737 align:middle line:84% You would think the primary vendor would just sell them 02:16:17.737 --> 02:16:20.320 align:middle line:84% for that higher price, and that would be the end of the story. 02:16:20.320 --> 02:16:23.660 align:middle line:84% Why do these secondary markets kind of exist? 02:16:23.660 --> 02:16:27.950 align:middle line:84% Why are they a necessary consequence? 02:16:27.950 --> 02:16:30.170 align:middle line:84% AUDIENCE: I guess in the fashion case, 02:16:30.170 --> 02:16:33.670 align:middle line:84% it's actually really important what specific type of brand 02:16:33.670 --> 02:16:34.709 align:middle line:90% we're talking about. 02:16:34.709 --> 02:16:36.150 align:middle line:84% So for Supreme, I think the reason 02:16:36.150 --> 02:16:39.590 align:middle line:84% why there's some Goldilocks zone between the resale 02:16:39.590 --> 02:16:42.670 align:middle line:84% and the retail is Supreme needs to figure out 02:16:42.670 --> 02:16:47.490 align:middle line:84% how to actually allow the skater group to still buy their brand. 02:16:47.490 --> 02:16:50.750 align:middle line:84% So skaters, we'll say, are a lot less-- 02:16:50.750 --> 02:16:52.290 align:middle line:90% or a lot more price elastic. 02:16:52.290 --> 02:16:56.969 align:middle line:84% And so if Supreme starts selling their stuff at $400 or $500, 02:16:56.969 --> 02:16:59.990 align:middle line:84% they've now entered just the luxury brand pricing 02:16:59.990 --> 02:17:01.770 align:middle line:90% with Fendi and whoever else. 02:17:01.770 --> 02:17:04.910 align:middle line:84% And they can't compete with those brands on price 02:17:04.910 --> 02:17:07.209 align:middle line:84% because that's just not the branding they are. 02:17:07.209 --> 02:17:09.555 align:middle line:84% So they have to keep it lower and-- 02:17:09.555 --> 02:17:11.430 align:middle line:84% AUDIENCE: Which is [INAUDIBLE] the secondary. 02:17:11.430 --> 02:17:13.830 align:middle line:84% AUDIENCE: The cultural relevance can still stay there. 02:17:13.830 --> 02:17:17.430 align:middle line:84% And the secondhand market exists to create the hype 02:17:17.430 --> 02:17:21.709 align:middle line:90% for these other people. 02:17:21.709 --> 02:17:24.980 align:middle line:84% So, yeah, I think there's an interesting tension there, 02:17:24.980 --> 02:17:27.400 align:middle line:84% but I don't think you can separate that 02:17:27.400 --> 02:17:29.040 align:middle line:84% from the cultural relevance, unless I 02:17:29.040 --> 02:17:30.478 align:middle line:90% chose a different brand. 02:17:30.478 --> 02:17:33.020 align:middle line:84% HAZHIR RAHMANDAD: It's a pretty interesting, complex problem, 02:17:33.020 --> 02:17:34.400 align:middle line:84% I think, kind of because, as soon 02:17:34.400 --> 02:17:38.520 align:middle line:84% as you go into goods that are not just utility focused, 02:17:38.520 --> 02:17:42.480 align:middle line:84% and hype and prestige play a role, 02:17:42.480 --> 02:17:47.120 align:middle line:84% you have very weird, different economic implications. 02:17:47.120 --> 02:17:50.219 align:middle line:84% And that's kind of creates interesting problems. 02:17:50.219 --> 02:17:52.320 align:middle line:84% So in that sense, this problem definition 02:17:52.320 --> 02:17:53.959 align:middle line:84% more starts with a mechanism that 02:17:53.959 --> 02:17:57.600 align:middle line:84% is, OK, the mechanism is demand is not driven only by price 02:17:57.600 --> 02:17:59.920 align:middle line:90% or some underlying utility. 02:17:59.920 --> 02:18:03.340 align:middle line:84% It is actually driven more by prestige and so on. 02:18:03.340 --> 02:18:06.260 align:middle line:84% And then within that, you can go in many different directions. 02:18:06.260 --> 02:18:08.700 align:middle line:84% So I think kind of focusing on one of those directions, 02:18:08.700 --> 02:18:10.980 align:middle line:84% which is what you are doing, makes sense. 02:18:10.980 --> 02:18:13.480 align:middle line:84% So let's think about secondhand market 02:18:13.480 --> 02:18:17.320 align:middle line:84% and its interaction with the firsthand market. 02:18:17.320 --> 02:18:23.030 align:middle line:84% That's kind of a nice focus area to just, OK, what would-- 02:18:23.030 --> 02:18:26.450 align:middle line:84% and then it has it a little bit of the policy lever question 02:18:26.450 --> 02:18:26.950 align:middle line:90% as well. 02:18:26.950 --> 02:18:29.410 align:middle line:84% What is the right pricing approach? 02:18:29.410 --> 02:18:31.230 align:middle line:84% And what is the right quantity approach? 02:18:31.230 --> 02:18:34.969 align:middle line:84% These two are kind of hand-in-hand for a brand 02:18:34.969 --> 02:18:38.309 align:middle line:84% like Supreme that could benefit from secondhand market. 02:18:38.309 --> 02:18:40.650 align:middle line:84% But how far, how long, will it benefit? 02:18:40.650 --> 02:18:42.530 align:middle line:84% At what price point would it benefit 02:18:42.530 --> 02:18:43.990 align:middle line:90% from that secondhand market? 02:18:43.990 --> 02:18:46.170 align:middle line:84% And what are the longer term implications 02:18:46.170 --> 02:18:50.129 align:middle line:84% about novelty, about being able to get it to the hand of people 02:18:50.129 --> 02:18:52.025 align:middle line:90% who create the word of mouth? 02:18:52.025 --> 02:18:54.150 align:middle line:84% So there is quite a bit that you can explore there. 02:18:54.150 --> 02:18:56.490 align:middle line:84% And I think kind of your updated problem definition 02:18:56.490 --> 02:18:57.790 align:middle line:90% is moving in that direction. 02:18:57.790 --> 02:19:00.129 align:middle line:90% So it's kind in a good space. 02:19:00.129 --> 02:19:01.010 align:middle line:90% Great. 02:19:01.010 --> 02:19:04.389 align:middle line:84% I think last one is Ross's problem. 02:19:04.389 --> 02:19:15.070 align:middle line:84% And Gerardo has some feedback on that. 02:19:15.070 --> 02:19:18.350 align:middle line:84% So, Ross, you're also looking at this space. 02:19:18.350 --> 02:19:22.370 align:middle line:84% So maybe you can give a very quick overview of what 02:19:22.370 --> 02:19:23.590 align:middle line:90% you're thinking of doing. 02:19:23.590 --> 02:19:25.190 align:middle line:90% And Gerardo can add. 02:19:25.190 --> 02:19:27.690 align:middle line:84% AUDIENCE: Yeah, I'm mostly focused on the primary market 02:19:27.690 --> 02:19:30.209 align:middle line:84% and how the company's going to price 02:19:30.209 --> 02:19:32.570 align:middle line:84% its goods because the price itself 02:19:32.570 --> 02:19:34.270 align:middle line:90% comes with certain trade offs. 02:19:34.270 --> 02:19:36.850 align:middle line:84% If you increase the price, you increase the desirability 02:19:36.850 --> 02:19:37.870 align:middle line:90% for a luxury brand. 02:19:37.870 --> 02:19:41.650 align:middle line:84% At the same time, the accessibility goes down. 02:19:41.650 --> 02:19:43.023 align:middle line:90% There's that trade-off there. 02:19:43.023 --> 02:19:44.690 align:middle line:84% And then there's also different segments 02:19:44.690 --> 02:19:49.170 align:middle line:84% of demand, people who are more price sensitive, 02:19:49.170 --> 02:19:54.890 align:middle line:84% which is a mass market, who would buy less per customer, 02:19:54.890 --> 02:20:00.810 align:middle line:84% but has a larger customer base, and then less price sensitive 02:20:00.810 --> 02:20:02.430 align:middle line:90% and more sensitive to quality. 02:20:02.430 --> 02:20:04.290 align:middle line:84% In the original version, I was also 02:20:04.290 --> 02:20:07.090 align:middle line:84% thinking about optimizing [INAUDIBLE] 02:20:07.090 --> 02:20:11.770 align:middle line:84% price to see how much profit they can get sustainably. 02:20:11.770 --> 02:20:15.250 align:middle line:84% But, yeah, in the end, I just decided 02:20:15.250 --> 02:20:19.620 align:middle line:84% to go with the fixed quality cost and competition, 02:20:19.620 --> 02:20:22.900 align:middle line:84% and just look at the price strategy. 02:20:22.900 --> 02:20:24.540 align:middle line:90% HAZHIR RAHMANDAD: Very nice. 02:20:24.540 --> 02:20:26.300 align:middle line:84% Gerardo, do you want to add anything 02:20:26.300 --> 02:20:31.220 align:middle line:90% to what your observations were? 02:20:31.220 --> 02:20:34.260 align:middle line:84% AUDIENCE: Yeah, so the model really is quite well-described. 02:20:34.260 --> 02:20:38.520 align:middle line:84% The [INAUDIBLE] models are really nice. 02:20:38.520 --> 02:20:42.140 align:middle line:84% And, yeah, I just had some clarification [INAUDIBLE] 02:20:42.140 --> 02:20:45.860 align:middle line:84% about, clarification of the clients 02:20:45.860 --> 02:20:49.000 align:middle line:90% to clients that are described. 02:20:49.000 --> 02:20:51.980 align:middle line:84% So my comment is that one of the two 02:20:51.980 --> 02:20:58.600 align:middle line:84% can be more related to activity and quality. 02:20:58.600 --> 02:21:02.980 align:middle line:84% And then the other one is more on ratio between quality 02:21:02.980 --> 02:21:05.420 align:middle line:90% and price, something like that. 02:21:05.420 --> 02:21:08.660 align:middle line:84% And another comment, yeah, the clarification 02:21:08.660 --> 02:21:11.400 align:middle line:90% also about brand prestige. 02:21:11.400 --> 02:21:15.660 align:middle line:84% And the other word, I don't remember, 02:21:15.660 --> 02:21:20.190 align:middle line:90% is just we're using [INAUDIBLE]. 02:21:20.190 --> 02:21:24.670 align:middle line:84% And the last one is on the competitive phase, 02:21:24.670 --> 02:21:28.230 align:middle line:84% where in the description, it's not clear 02:21:28.230 --> 02:21:33.050 align:middle line:84% what kind of mechanism, like more endogenous mechanisms 02:21:33.050 --> 02:21:38.230 align:middle line:84% [INAUDIBLE] that provide more detail about it [INAUDIBLE] 02:21:38.230 --> 02:21:39.390 align:middle line:90% but [INAUDIBLE]. 02:21:39.390 --> 02:21:43.743 align:middle line:90% 02:21:43.743 --> 02:21:44.910 align:middle line:90% HAZHIR RAHMANDAD: You have-- 02:21:44.910 --> 02:21:48.870 align:middle line:84% AUDIENCE: I was wondering, in your model, what else is driving 02:21:48.870 --> 02:21:51.550 align:middle line:90% desirability beyond price? 02:21:51.550 --> 02:21:57.690 align:middle line:84% 2 and quality, which is reflected in the cost structure. 02:21:57.690 --> 02:22:00.788 align:middle line:84% It's supposed to be, but I guess [INAUDIBLE]. 02:22:00.788 --> 02:22:02.830 align:middle line:84% HAZHIR RAHMANDAD: So if you fix the quality, then 02:22:02.830 --> 02:22:06.930 align:middle line:84% how does price impact, basically, demand? 02:22:06.930 --> 02:22:09.730 align:middle line:84% So how are you seeing it in the updated version? 02:22:09.730 --> 02:22:13.350 align:middle line:84% AUDIENCE: Yeah, there's an inertia in the price itself, 02:22:13.350 --> 02:22:16.700 align:middle line:84% the people's expectation of whether prices go up or down. 02:22:16.700 --> 02:22:19.040 align:middle line:84% And then there's a competition price, 02:22:19.040 --> 02:22:25.200 align:middle line:84% I guess like the relative price [INAUDIBLE] 02:22:25.200 --> 02:22:30.080 align:middle line:84% prestige, which is going to decide 02:22:30.080 --> 02:22:31.600 align:middle line:90% the utility of the demand. 02:22:31.600 --> 02:22:35.980 align:middle line:84% And then how much is consumed is actually going to affect it 02:22:35.980 --> 02:22:37.200 align:middle line:90% too, the price. 02:22:37.200 --> 02:22:40.060 align:middle line:84% It's like the more you see people having it, 02:22:40.060 --> 02:22:42.040 align:middle line:90% it's less desirable. 02:22:42.040 --> 02:22:45.517 align:middle line:84% So I think there are certain mechanisms there. 02:22:45.517 --> 02:22:47.100 align:middle line:84% AUDIENCE: What brand are you thinking? 02:22:47.100 --> 02:22:50.040 align:middle line:90% 02:22:50.040 --> 02:22:51.040 align:middle line:90% HAZHIR RAHMANDAD: Rolex. 02:22:51.040 --> 02:22:52.840 align:middle line:84% AUDIENCE: Yeah, something like that. 02:22:52.840 --> 02:22:55.800 align:middle line:84% AUDIENCE: But can you really say the more you see people wearing 02:22:55.800 --> 02:22:57.420 align:middle line:90% it, it's less desirable? 02:22:57.420 --> 02:22:59.262 align:middle line:90% AUDIENCE: I feel like-- 02:22:59.262 --> 02:23:01.220 align:middle line:84% AUDIENCE: It's definitely true for some brands. 02:23:01.220 --> 02:23:04.920 align:middle line:84% AUDIENCE: Yeah, what type of people, I think, 02:23:04.920 --> 02:23:05.900 align:middle line:90% is the more important. 02:23:05.900 --> 02:23:07.900 align:middle line:84% AUDIENCE: So you have to model who's wearing it. 02:23:07.900 --> 02:23:11.280 align:middle line:84% AUDIENCE: Yeah, like conspicuous consumption, 02:23:11.280 --> 02:23:12.945 align:middle line:84% like the thoroughbred market, that's 02:23:12.945 --> 02:23:14.070 align:middle line:90% where I got the motivation. 02:23:14.070 --> 02:23:14.690 align:middle line:90% AUDIENCE: The what? 02:23:14.690 --> 02:23:15.810 align:middle line:90% AUDIENCE: The thoroughbred. 02:23:15.810 --> 02:23:18.060 align:middle line:84% AUDIENCE: Thoroughbred, yeah, conspicuous [INAUDIBLE]. 02:23:18.060 --> 02:23:19.768 align:middle line:84% AUDIENCE: Yeah, that's right [INAUDIBLE]. 02:23:19.768 --> 02:23:21.143 align:middle line:84% HAZHIR RAHMANDAD: So I think this 02:23:21.143 --> 02:23:22.470 align:middle line:90% is pretty interesting space. 02:23:22.470 --> 02:23:25.090 align:middle line:90% I think there was a-- 02:23:25.090 --> 02:23:27.783 align:middle line:84% Ross's problem definition is very operational. 02:23:27.783 --> 02:23:29.950 align:middle line:84% And the first version, also, was pretty operational. 02:23:29.950 --> 02:23:31.110 align:middle line:90% And now he has updated it. 02:23:31.110 --> 02:23:33.670 align:middle line:90% So it's kind of a nice one. 02:23:33.670 --> 02:23:35.270 align:middle line:84% We'll see how things come together. 02:23:35.270 --> 02:23:37.210 align:middle line:84% And I think with the overlap with yours, 02:23:37.210 --> 02:23:42.090 align:middle line:84% you can compare notes as well and look at each other's models. 02:23:42.090 --> 02:23:45.090 align:middle line:84% So thanks, everybody, for those who provided 02:23:45.090 --> 02:23:46.410 align:middle line:90% feedback for their friends. 02:23:46.410 --> 02:23:49.530 align:middle line:84% And the feedbacks are there in-- and we kind of 02:23:49.530 --> 02:23:54.490 align:middle line:84% discussed their outlines as well together. 02:23:54.490 --> 02:23:56.850 align:middle line:84% Maybe we have time to go through one more model 02:23:56.850 --> 02:23:58.790 align:middle line:90% before closing the class today. 02:23:58.790 --> 02:24:04.090 align:middle line:84% So anybody wants their model to be the Guinea pig? 02:24:04.090 --> 02:24:05.350 align:middle line:90% Hassan seems to be-- 02:24:05.350 --> 02:24:06.693 align:middle line:90% AUDIENCE: Yeah, [INAUDIBLE]. 02:24:06.693 --> 02:24:07.610 align:middle line:90% Different [INAUDIBLE]. 02:24:07.610 --> 02:24:09.193 align:middle line:84% HAZHIR RAHMANDAD: Sure, let's do that. 02:24:09.193 --> 02:24:12.860 align:middle line:90% So is this Hassan's model? 02:24:12.860 --> 02:24:13.860 align:middle line:90% Yes. 02:24:13.860 --> 02:24:18.440 align:middle line:84% So as we kind of, at the beginning, talked about it, 02:24:18.440 --> 02:24:22.860 align:middle line:84% you have novices and experts working on projects 02:24:22.860 --> 02:24:24.480 align:middle line:90% moving into these stocks. 02:24:24.480 --> 02:24:26.120 align:middle line:84% So let's first go to the projects. 02:24:26.120 --> 02:24:28.040 align:middle line:90% You have potential projects. 02:24:28.040 --> 02:24:30.780 align:middle line:84% There is an exogenous number coming in. 02:24:30.780 --> 02:24:34.620 align:middle line:84% And this is in the form of projects. 02:24:34.620 --> 02:24:38.820 align:middle line:84% So new projects come for the field. 02:24:38.820 --> 02:24:40.120 align:middle line:90% It's exogenous. 02:24:40.120 --> 02:24:43.464 align:middle line:90% 02:24:43.464 --> 02:24:46.540 align:middle line:84% AUDIENCE: Yeah, and then the reputation 02:24:46.540 --> 02:24:50.480 align:middle line:84% and recognition feed into the indicated demand. 02:24:50.480 --> 02:24:55.820 align:middle line:84% HAZHIR RAHMANDAD: Yes, so then the recent indicated demand 02:24:55.820 --> 02:25:07.780 align:middle line:84% is based on what is the potential multiplied replied 02:25:07.780 --> 02:25:12.940 align:middle line:84% by the sum of potential demand and-- 02:25:12.940 --> 02:25:15.040 align:middle line:84% AUDIENCE: So sum of all the demand. 02:25:15.040 --> 02:25:18.180 align:middle line:84% So I'm looking at the percentage of all demand that 02:25:18.180 --> 02:25:21.765 align:middle line:90% is interested in the field. 02:25:21.765 --> 02:25:23.640 align:middle line:84% HAZHIR RAHMANDAD: What was the recent demand? 02:25:23.640 --> 02:25:24.100 align:middle line:90% Sorry. 02:25:24.100 --> 02:25:26.475 align:middle line:84% AUDIENCE: The recent demand is the sum of the other three 02:25:26.475 --> 02:25:27.500 align:middle line:90% stocks. 02:25:27.500 --> 02:25:30.380 align:middle line:90% HAZHIR RAHMANDAD: Oh, OK. 02:25:30.380 --> 02:25:36.060 align:middle line:84% So, basically, everything added up. 02:25:36.060 --> 02:25:38.020 align:middle line:84% AUDIENCE: But multiply it by a couple of-- 02:25:38.020 --> 02:25:40.760 align:middle line:84% reputation and recognition are both between 0 and 1, 02:25:40.760 --> 02:25:42.260 align:middle line:90% so the percentages. 02:25:42.260 --> 02:25:46.260 align:middle line:84% HAZHIR RAHMANDAD: So, basically, all the possible-- 02:25:46.260 --> 02:25:48.740 align:middle line:84% all the demand that is in any of-- all the projects 02:25:48.740 --> 02:25:52.560 align:middle line:84% that are in any of these stocks multiplied by reputation 02:25:52.560 --> 02:25:58.900 align:middle line:90% and the other thing. 02:25:58.900 --> 02:26:02.820 align:middle line:84% The recognition gives you some indicated demand, 02:26:02.820 --> 02:26:06.750 align:middle line:84% that then your demand is moving towards that 02:26:06.750 --> 02:26:10.130 align:middle line:84% with this indicated minus recent. 02:26:10.130 --> 02:26:11.870 align:middle line:84% AUDIENCE: We still have that risk of-- 02:26:11.870 --> 02:26:14.630 align:middle line:84% HAZHIR RAHMANDAD: That has the problem of reverse 02:26:14.630 --> 02:26:16.430 align:middle line:90% may get too much. 02:26:16.430 --> 02:26:17.770 align:middle line:90% So that's part of the problem. 02:26:17.770 --> 02:26:18.770 align:middle line:90% The other part is-- 02:26:18.770 --> 02:26:32.030 align:middle line:90% 02:26:32.030 --> 02:26:35.330 align:middle line:84% so, usually, recent demand might be much bigger. 02:26:35.330 --> 02:26:38.170 align:middle line:84% So this becomes negative quite often. 02:26:38.170 --> 02:26:42.190 align:middle line:90% And that's the problem, yeah. 02:26:42.190 --> 02:26:45.750 align:middle line:84% And then if this indicator demand goes up, 02:26:45.750 --> 02:26:47.430 align:middle line:84% now you have more unaddressed demand 02:26:47.430 --> 02:26:56.630 align:middle line:84% for the field, which experts and novices start projects 02:26:56.630 --> 02:27:01.550 align:middle line:84% because it becomes addressing demand. 02:27:01.550 --> 02:27:04.490 align:middle line:84% And then these are ongoing projects. 02:27:04.490 --> 02:27:08.080 align:middle line:84% And then you have novices and experts delivering projects. 02:27:08.080 --> 02:27:10.480 align:middle line:90% That becomes project delivery. 02:27:10.480 --> 02:27:14.080 align:middle line:84% And then this recent projects informs reputation. 02:27:14.080 --> 02:27:16.880 align:middle line:84% Or recognition, not reputation yet. 02:27:16.880 --> 02:27:21.260 align:middle line:84% And then this expert is starting projects and novices 02:27:21.260 --> 02:27:24.200 align:middle line:84% starting projects is this flow that 02:27:24.200 --> 02:27:28.820 align:middle line:84% looks at the unaddressed demand in the field. 02:27:28.820 --> 02:27:30.840 align:middle line:84% AUDIENCE: With the experts if we have-- 02:27:30.840 --> 02:27:31.840 align:middle line:90% HAZHIR RAHMANDAD: I see. 02:27:31.840 --> 02:27:36.600 align:middle line:84% So the unaddressed demand in the whole field 02:27:36.600 --> 02:27:41.920 align:middle line:84% is going to be unaddressed demand for experts. 02:27:41.920 --> 02:27:42.923 align:middle line:90% AUDIENCE: That's it. 02:27:42.923 --> 02:27:45.340 align:middle line:84% HAZHIR RAHMANDAD: And then the part that they can fulfill, 02:27:45.340 --> 02:27:46.710 align:middle line:90% they will fulfill. 02:27:46.710 --> 02:27:47.360 align:middle line:90% AUDIENCE: Yep. 02:27:47.360 --> 02:27:49.470 align:middle line:84% HAZHIR RAHMANDAD: And then what remains-- hmm? 02:27:49.470 --> 02:27:51.720 align:middle line:84% AUDIENCE: The function there is that the function that 02:27:51.720 --> 02:27:56.640 align:middle line:84% is tangent to y equal to x and then goes-- saturates at y 02:27:56.640 --> 02:27:57.242 align:middle line:90% equal to 1. 02:27:57.242 --> 02:27:58.200 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 02:27:58.200 --> 02:28:03.610 align:middle line:84% AUDIENCE: And then the gap goes towards small business. 02:28:03.610 --> 02:28:07.250 align:middle line:84% HAZHIR RAHMANDAD: Yeah, and then the rest goes to the novices. 02:28:07.250 --> 02:28:09.390 align:middle line:84% That allows novices to get started. 02:28:09.390 --> 02:28:12.350 align:middle line:84% So, basically, you have a bunch of projects in this stock. 02:28:12.350 --> 02:28:15.930 align:middle line:90% 02:28:15.930 --> 02:28:20.593 align:middle line:84% AUDIENCE: Yeah, whatever can be picked up by experts. 02:28:20.593 --> 02:28:22.010 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so you are 02:28:22.010 --> 02:28:24.170 align:middle line:84% kind of missing a time constant to make 02:28:24.170 --> 02:28:26.170 align:middle line:90% this transition a little-- 02:28:26.170 --> 02:28:29.010 align:middle line:84% built into the project per person 02:28:29.010 --> 02:28:31.530 align:middle line:84% because this needs to be in the units of-- 02:28:31.530 --> 02:28:33.950 align:middle line:84% AUDIENCE: You have the time to initiate projects there, 02:28:33.950 --> 02:28:34.950 align:middle line:90% but, yeah. 02:28:34.950 --> 02:28:36.617 align:middle line:84% HAZHIR RAHMANDAD: Yeah, but the concept, 02:28:36.617 --> 02:28:38.390 align:middle line:84% really, if it is demand, it should be-- 02:28:38.390 --> 02:28:42.610 align:middle line:84% so this is the projects that are in the backlog. 02:28:42.610 --> 02:28:46.530 align:middle line:84% Then you want to turn it into demand for projects or kind 02:28:46.530 --> 02:28:47.730 align:middle line:90% of allocation. 02:28:47.730 --> 02:28:51.390 align:middle line:84% And, yes, the time constant can come somewhere down here, 02:28:51.390 --> 02:28:55.810 align:middle line:84% but it is better for it to be turned into demand right 02:28:55.810 --> 02:28:58.070 align:middle line:84% when you are trying to match it with experts. 02:28:58.070 --> 02:29:02.360 align:middle line:84% That's a little bit more taste in putting the equations, 02:29:02.360 --> 02:29:05.340 align:middle line:84% but makes it easier to read because unaddressed demand 02:29:05.340 --> 02:29:08.900 align:middle line:84% for experts feels like a bunch of projects 02:29:08.900 --> 02:29:12.260 align:middle line:90% are now allocated to experts. 02:29:12.260 --> 02:29:14.300 align:middle line:90% You want to flow off. 02:29:14.300 --> 02:29:15.940 align:middle line:90% I don't know. 02:29:15.940 --> 02:29:19.460 align:middle line:84% AUDIENCE: I didn't like what I was doing there myself because I 02:29:19.460 --> 02:29:25.140 align:middle line:84% felt like it's a little too much binary choices 02:29:25.140 --> 02:29:29.360 align:middle line:84% and discrete choices [INAUDIBLE] between everything, sort of. 02:29:29.360 --> 02:29:32.180 align:middle line:84% HAZHIR RAHMANDAD: Yeah, but, anyway, you 02:29:32.180 --> 02:29:35.220 align:middle line:84% have the sufficiency availability of experts 02:29:35.220 --> 02:29:38.540 align:middle line:90% that creates a bottleneck here. 02:29:38.540 --> 02:29:41.300 align:middle line:84% And this formulation is also messy. 02:29:41.300 --> 02:29:42.420 align:middle line:90% Oh, God, yeah. 02:29:42.420 --> 02:29:43.420 align:middle line:90% AUDIENCE: Yeah, that's-- 02:29:43.420 --> 02:29:47.980 align:middle line:84% HAZHIR RAHMANDAD: Yeah, I know, but unaddressed demand 02:29:47.980 --> 02:29:51.140 align:middle line:84% for experts divided by time to initiate 02:29:51.140 --> 02:29:55.100 align:middle line:90% projects multiplied by this. 02:29:55.100 --> 02:29:55.905 align:middle line:90% So you have the-- 02:29:55.905 --> 02:29:56.780 align:middle line:90% AUDIENCE: [INAUDIBLE] 02:29:56.780 --> 02:29:58.730 align:middle line:84% HAZHIR RAHMANDAD: Yeah, utilization ratio. 02:29:58.730 --> 02:30:00.990 align:middle line:90% So take it out of the equation. 02:30:00.990 --> 02:30:03.792 align:middle line:84% You never put so many things together, 02:30:03.792 --> 02:30:05.250 align:middle line:84% especially with that complex thing. 02:30:05.250 --> 02:30:08.970 align:middle line:84% You can never track what is exactly happening. 02:30:08.970 --> 02:30:09.890 align:middle line:90% But that's the idea. 02:30:09.890 --> 02:30:12.950 align:middle line:84% And then once these guys-- it takes just some time for them 02:30:12.950 --> 02:30:14.190 align:middle line:90% to do it. 02:30:14.190 --> 02:30:16.710 align:middle line:84% The way it is working, this is basically 02:30:16.710 --> 02:30:18.410 align:middle line:90% the initiation of new projects. 02:30:18.410 --> 02:30:20.290 align:middle line:84% So it goes into the ongoing projects. 02:30:20.290 --> 02:30:23.030 align:middle line:84% And then when they finish, this goes out. 02:30:23.030 --> 02:30:27.830 align:middle line:84% So that all is sensible, at least, at the first cut 02:30:27.830 --> 02:30:28.990 align:middle line:90% is sensible. 02:30:28.990 --> 02:30:33.070 align:middle line:84% Then you have these kind of quality measures, the reputation 02:30:33.070 --> 02:30:35.630 align:middle line:84% coming from whether the experts are delivering it 02:30:35.630 --> 02:30:37.070 align:middle line:90% or the novices. 02:30:37.070 --> 02:30:42.350 align:middle line:84% That creates a reputation that is just this prevalence 02:30:42.350 --> 02:30:43.485 align:middle line:90% of expert delivery. 02:30:43.485 --> 02:30:45.110 align:middle line:84% So you could actually combine these two 02:30:45.110 --> 02:30:46.530 align:middle line:90% into a single parameter. 02:30:46.530 --> 02:30:49.070 align:middle line:90% You have asymmetric adjustment. 02:30:49.070 --> 02:30:52.670 align:middle line:84% So all of it is fine, but goes back to this field reputation 02:30:52.670 --> 02:30:55.110 align:middle line:84% now, coming back and feeding here 02:30:55.110 --> 02:30:58.070 align:middle line:84% directly into the indicated demand, exactly 02:30:58.070 --> 02:31:01.830 align:middle line:84% the same conversation we had with Alban's model, 02:31:01.830 --> 02:31:03.990 align:middle line:84% that this needs another degree of freedom 02:31:03.990 --> 02:31:07.780 align:middle line:90% to have the right effect. 02:31:07.780 --> 02:31:09.530 align:middle line:84% Otherwise, you don't have control over it. 02:31:09.530 --> 02:31:13.070 align:middle line:84% This is about expert versus novice availability. 02:31:13.070 --> 02:31:16.130 align:middle line:84% Then we have the joining of the field. 02:31:16.130 --> 02:31:19.950 align:middle line:84% You have exogenous joining from the novices that 02:31:19.950 --> 02:31:22.010 align:middle line:90% come into the picture. 02:31:22.010 --> 02:31:24.050 align:middle line:90% So that is a little bit-- 02:31:24.050 --> 02:31:26.558 align:middle line:90% 02:31:26.558 --> 02:31:28.350 align:middle line:84% AUDIENCE: I think the only thing that I did 02:31:28.350 --> 02:31:30.048 align:middle line:90% model endogenously is that-- 02:31:30.048 --> 02:31:31.590 align:middle line:84% HAZHIR RAHMANDAD: Yeah, opportunities 02:31:31.590 --> 02:31:36.158 align:middle line:84% are impacting the flow of people to stay in or leave. 02:31:36.158 --> 02:31:37.950 align:middle line:84% AUDIENCE: And the leaving happens for those 02:31:37.950 --> 02:31:39.952 align:middle line:90% who are not busy, essentially. 02:31:39.952 --> 02:31:40.910 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 02:31:40.910 --> 02:31:42.630 align:middle line:84% AUDIENCE: Which is a lot of the stuff-- 02:31:42.630 --> 02:31:45.410 align:middle line:84% HAZHIR RAHMANDAD: Yeah, but so I think the leaving part is fine. 02:31:45.410 --> 02:31:48.870 align:middle line:84% But you probably also need some feedback for the entering. 02:31:48.870 --> 02:31:51.128 align:middle line:84% Fixed entering is a little bit too extreme. 02:31:51.128 --> 02:31:51.670 align:middle line:90% This is good. 02:31:51.670 --> 02:31:52.930 align:middle line:90% I think we are out of time. 02:31:52.930 --> 02:31:56.200 align:middle line:84% So, basically, great job on these. 02:31:56.200 --> 02:31:58.760 align:middle line:84% So starting next week, we are moving 02:31:58.760 --> 02:32:01.600 align:middle line:90% into your personal projects. 02:32:01.600 --> 02:32:05.360 align:middle line:84% If you haven't already, please put your problem definition 02:32:05.360 --> 02:32:08.248 align:middle line:84% in the final version of the problem definition, 02:32:08.248 --> 02:32:09.540 align:middle line:90% whatever is your final version. 02:32:09.540 --> 02:32:12.400 align:middle line:84% If you have edited it based on the feedback that you 02:32:12.400 --> 02:32:14.100 align:middle line:84% have received, please just put it there. 02:32:14.100 --> 02:32:17.200 align:middle line:84% Nelson will be looking at those when he comes to the class. 02:32:17.200 --> 02:32:18.800 align:middle line:84% He needs a little bit of a background 02:32:18.800 --> 02:32:23.120 align:middle line:84% to be able to provide meaningful feedback in the class. 02:32:23.120 --> 02:32:24.600 align:middle line:90% That's number one. 02:32:24.600 --> 02:32:28.720 align:middle line:84% If you decided at the end that, OK, the personal project is 02:32:28.720 --> 02:32:30.860 align:middle line:84% too complex or you don't want to work on it, 02:32:30.860 --> 02:32:34.760 align:middle line:84% you're welcome to work on any of these problem assignments 02:32:34.760 --> 02:32:38.160 align:middle line:84% that we have had, and improve that, and submit that 02:32:38.160 --> 02:32:41.340 align:middle line:84% into the relevant folder and get feedback on that. 02:32:41.340 --> 02:32:44.140 align:middle line:84% So that is a legitimate alternative 02:32:44.140 --> 02:32:47.168 align:middle line:84% if you don't feel good about your problem definitions, 02:32:47.168 --> 02:32:49.460 align:middle line:84% or you don't want to work on those for whatever reason. 02:32:49.460 --> 02:32:52.530 align:middle line:90% That's also perfectly fine. 02:32:52.530 --> 02:32:55.870 align:middle line:84% Any questions for plans for the coming weeks? 02:32:55.870 --> 02:33:00.410 align:middle line:84% So, basically, the plan is, each week, submit any update 02:33:00.410 --> 02:33:01.310 align:middle line:90% that you have. 02:33:01.310 --> 02:33:02.470 align:middle line:90% We have a conversation. 02:33:02.470 --> 02:33:03.490 align:middle line:90% We try to give feedback. 02:33:03.490 --> 02:33:05.270 align:middle line:84% And, hopefully, in a couple of weeks, 02:33:05.270 --> 02:33:08.070 align:middle line:84% we will have some interesting, clean models. 02:33:08.070 --> 02:33:11.170 align:middle line:90% 02:33:11.170 --> 02:33:11.790 align:middle line:90% Great. 02:33:11.790 --> 02:33:13.130 align:middle line:90% Thanks so much. 02:33:13.130 --> 02:33:16.786 align:middle line:90% Oh, go ahead, Maria. 02:33:16.786 --> 02:33:18.310 align:middle line:90% AUDIENCE: [INAUDIBLE], sorry. 02:33:18.310 --> 02:33:19.270 align:middle line:90% Just a quick question. 02:33:19.270 --> 02:33:22.690 align:middle line:84% So for next week, we can also submit some income version 02:33:22.690 --> 02:33:27.167 align:middle line:90% of the model in Benson. 02:33:27.167 --> 02:33:29.250 align:middle line:84% AUDIENCE: Submit an initial version of the project 02:33:29.250 --> 02:33:30.790 align:middle line:90% model in Benson for next week 02:33:30.790 --> 02:33:33.910 align:middle line:84% HAZHIR RAHMANDAD: Yes so submit and-- 02:33:33.910 --> 02:33:35.010 align:middle line:90% [LAUGHTER] 02:33:35.010 --> 02:33:36.095 align:middle line:90% And she got disconnected. 02:33:36.095 --> 02:33:36.970 align:middle line:90% AUDIENCE: [INAUDIBLE] 02:33:36.970 --> 02:33:40.670 align:middle line:84% HAZHIR RAHMANDAD: Yes, submit the initial version in Benson. 02:33:40.670 --> 02:33:44.370 align:middle line:90% And the whole system is gone. 02:33:44.370 --> 02:33:45.710 align:middle line:90% So that's it. 02:33:45.710 --> 02:33:46.230 align:middle line:90% Thank you. 02:33:46.230 --> 02:33:47.130 align:middle line:90% [LAUGHS] 02:33:47.130 --> 02:33:49.580 align:middle line:90% [SIDE CONVERSATIONS] 02:33:49.580 --> 02:33:52.000 align:middle line:90%