WEBVTT 00:00:00.000 --> 00:00:02.472 align:middle line:90% [SQUEAKING] 00:00:02.472 --> 00:00:04.444 align:middle line:90% [RUSTLING] 00:00:04.444 --> 00:00:06.416 align:middle line:90% [CLICKING] 00:00:06.416 --> 00:00:09.853 align:middle line:90% 00:00:09.853 --> 00:00:12.520 align:middle line:84% HAZHIR RAHMANDAD: This week, we are looking at the market growth 00:00:12.520 --> 00:00:14.520 align:middle line:90% model. 00:00:14.520 --> 00:00:19.120 align:middle line:84% So this is kind of inspired by Forrester's market growth, 00:00:19.120 --> 00:00:22.280 align:middle line:84% but has a few additional components to it. 00:00:22.280 --> 00:00:25.420 align:middle line:84% So I kind of have that there is a total market size 00:00:25.420 --> 00:00:29.520 align:middle line:84% and this is a durable product, so there is a replacement 00:00:29.520 --> 00:00:36.280 align:middle line:84% purchase and there is a first time purchase type of a dynamic. 00:00:36.280 --> 00:00:39.080 align:middle line:84% The part that is the product is sold through the direct sales 00:00:39.080 --> 00:00:44.200 align:middle line:84% force is pretty similar to Forrester. 00:00:44.200 --> 00:00:47.800 align:middle line:84% And you start with a fraction of total revenue being allocated 00:00:47.800 --> 00:00:52.040 align:middle line:84% to the sales force, and then their productivity 00:00:52.040 --> 00:00:55.560 align:middle line:84% depends on the delivery delay, but also 00:00:55.560 --> 00:00:57.260 align:middle line:90% on availability of customers. 00:00:57.260 --> 00:00:59.180 align:middle line:84% And that's kind of the additional component 00:00:59.180 --> 00:01:04.500 align:middle line:84% to the classical market growth model of Forrester. 00:01:04.500 --> 00:01:06.900 align:middle line:84% There is a manufacturing operation. 00:01:06.900 --> 00:01:10.400 align:middle line:84% You have a backlog for ordered products. 00:01:10.400 --> 00:01:11.900 align:middle line:84% You are not holding inventory, so it 00:01:11.900 --> 00:01:14.400 align:middle line:84% doesn't have the complexities of inventory management. 00:01:14.400 --> 00:01:16.060 align:middle line:84% It has a little bit of additional piece 00:01:16.060 --> 00:01:18.220 align:middle line:90% for backlog management. 00:01:18.220 --> 00:01:20.940 align:middle line:84% And then it's kind of calculation of excess delivery 00:01:20.940 --> 00:01:21.980 align:middle line:90% delay. 00:01:21.980 --> 00:01:26.260 align:middle line:84% So those are the big picture of what the scope of the problem 00:01:26.260 --> 00:01:26.900 align:middle line:90% was. 00:01:26.900 --> 00:01:32.140 align:middle line:84% Hopefully, you have worked on it and you know the challenges. 00:01:32.140 --> 00:01:35.300 align:middle line:84% It would be good to hear your thoughts on the challenges 00:01:35.300 --> 00:01:38.740 align:middle line:84% you have faced in terms of things 00:01:38.740 --> 00:01:44.380 align:middle line:84% that you noticed where-- and let me just make this screen. 00:01:44.380 --> 00:01:46.080 align:middle line:84% Oh, actually, you can just type here. 00:01:46.080 --> 00:01:49.100 align:middle line:90% 00:01:49.100 --> 00:01:50.260 align:middle line:90% Let's see. 00:01:50.260 --> 00:01:52.040 align:middle line:84% AUDIENCE: Not a specific formulation, 00:01:52.040 --> 00:01:55.010 align:middle line:84% but I just found this one very difficult to do iteratively. 00:01:55.010 --> 00:01:56.260 align:middle line:90% HAZHIR RAHMANDAD: Interesting. 00:01:56.260 --> 00:01:59.480 align:middle line:84% AUDIENCE: Like it felt like everything tied very closely 00:01:59.480 --> 00:02:00.183 align:middle line:90% into each other. 00:02:00.183 --> 00:02:02.600 align:middle line:84% So as I was trying to come up with the formulation for one 00:02:02.600 --> 00:02:05.517 align:middle line:84% thing, I found that I needed to do a lot of structures 00:02:05.517 --> 00:02:06.100 align:middle line:90% for the other. 00:02:06.100 --> 00:02:08.440 align:middle line:84% And then when I had problems, I found 00:02:08.440 --> 00:02:10.590 align:middle line:84% it more difficult to diagnose as a result. 00:02:10.590 --> 00:02:11.840 align:middle line:90% HAZHIR RAHMANDAD: Interesting. 00:02:11.840 --> 00:02:16.060 align:middle line:84% So can you tell us what was your process? 00:02:16.060 --> 00:02:18.360 align:middle line:84% Where did you start, and where did you 00:02:18.360 --> 00:02:24.320 align:middle line:84% find it more challenging to isolate? 00:02:24.320 --> 00:02:33.000 align:middle line:84% AUDIENCE: Yeah, so I started, I think, with the order backlog 00:02:33.000 --> 00:02:35.640 align:middle line:84% through to delivered goods, just trying 00:02:35.640 --> 00:02:42.220 align:middle line:84% to make the very basic framework there. 00:02:42.220 --> 00:02:45.840 align:middle line:84% But then I felt like I needed to go a step up from that 00:02:45.840 --> 00:02:48.500 align:middle line:90% of how many orders there were. 00:02:48.500 --> 00:02:52.778 align:middle line:84% I felt like I needed some level of the sales force immediately. 00:02:52.778 --> 00:02:54.820 align:middle line:84% And because I needed the sales force immediately, 00:02:54.820 --> 00:02:57.198 align:middle line:84% I felt like I needed revenue immediately, 00:02:57.198 --> 00:02:59.240 align:middle line:84% because I felt like I needed revenue immediately. 00:02:59.240 --> 00:03:01.468 align:middle line:84% It all just played in together at once. 00:03:01.468 --> 00:03:02.760 align:middle line:90% HAZHIR RAHMANDAD: I see, I see. 00:03:02.760 --> 00:03:06.280 align:middle line:84% So this is actually a very good question and approach. 00:03:06.280 --> 00:03:08.940 align:middle line:84% That is, you want as much as possible to have chunks 00:03:08.940 --> 00:03:12.398 align:middle line:84% of model that are fairly clean, independent, 00:03:12.398 --> 00:03:13.440 align:middle line:90% and then you add a chunk. 00:03:13.440 --> 00:03:16.700 align:middle line:84% And so I like Jake's approach, trying to do that. 00:03:16.700 --> 00:03:20.780 align:middle line:84% And this model, in my view, is not necessarily that much harder 00:03:20.780 --> 00:03:24.460 align:middle line:84% to do it compared to many other models. 00:03:24.460 --> 00:03:26.840 align:middle line:84% Perhaps even some of the chunks are easier. 00:03:26.840 --> 00:03:28.280 align:middle line:90% And your approach was good. 00:03:28.280 --> 00:03:29.940 align:middle line:84% So you have a kind of a production 00:03:29.940 --> 00:03:31.360 align:middle line:90% sector, the production. 00:03:31.360 --> 00:03:33.780 align:middle line:84% There is a production capacity, orders are coming in 00:03:33.780 --> 00:03:36.180 align:middle line:84% and are going into the backlog, and then 00:03:36.180 --> 00:03:37.600 align:middle line:90% delivered out of the backlog. 00:03:37.600 --> 00:03:41.240 align:middle line:84% So that part you can formulate and have exogenous orders. 00:03:41.240 --> 00:03:43.320 align:middle line:84% That's actually how you would do it. 00:03:43.320 --> 00:03:46.320 align:middle line:84% You can create exogenous orders, have a step function, 00:03:46.320 --> 00:03:49.060 align:middle line:84% go into it, see how things go, have exponential growth 00:03:49.060 --> 00:03:53.640 align:middle line:84% going into it, how it works, and be done with that part. 00:03:53.640 --> 00:03:56.320 align:middle line:84% Then there is a mechanism for ordering, 00:03:56.320 --> 00:04:02.280 align:middle line:84% and there the exogenous drivers might be available 00:04:02.280 --> 00:04:05.900 align:middle line:84% customers and delivery delay that is perceived. 00:04:05.900 --> 00:04:09.320 align:middle line:84% So if you make those two as exogenous inputs 00:04:09.320 --> 00:04:15.280 align:middle line:84% into the sales force productivity effectiveness part, 00:04:15.280 --> 00:04:20.120 align:middle line:84% that can also create its own chunk of a structure. 00:04:20.120 --> 00:04:26.720 align:middle line:90% And then what remains? 00:04:26.720 --> 00:04:30.600 align:middle line:84% So the delivery delay is coming from the production side. 00:04:30.600 --> 00:04:33.420 align:middle line:90% And then yeah, that would be it. 00:04:33.420 --> 00:04:36.420 align:middle line:84% Now, there is a market side if you want to have it separately. 00:04:36.420 --> 00:04:39.340 align:middle line:84% That is how many customers are available. 00:04:39.340 --> 00:04:41.400 align:middle line:84% So I think with those three chunks, 00:04:41.400 --> 00:04:44.380 align:middle line:84% you could isolate each of them relatively smoothly. 00:04:44.380 --> 00:04:50.880 align:middle line:84% So the customer side, there will be a sales force coming into it 00:04:50.880 --> 00:04:55.220 align:middle line:84% or even a sales rate coming into it, 00:04:55.220 --> 00:04:57.120 align:middle line:90% and that will take care of it. 00:04:57.120 --> 00:04:58.960 align:middle line:84% That would be how I would cut it. 00:04:58.960 --> 00:05:02.540 align:middle line:84% But we can go through some of the examples 00:05:02.540 --> 00:05:05.020 align:middle line:84% or how different students have done it 00:05:05.020 --> 00:05:08.280 align:middle line:84% and see how that has been tackled. 00:05:08.280 --> 00:05:08.780 align:middle line:90% Good. 00:05:08.780 --> 00:05:10.040 align:middle line:90% Other challenges? 00:05:10.040 --> 00:05:21.700 align:middle line:90% 00:05:21.700 --> 00:05:27.580 align:middle line:84% So we can-- let me just close this thing 00:05:27.580 --> 00:05:29.860 align:middle line:90% so it doesn't confuse us. 00:05:29.860 --> 00:05:32.600 align:middle line:90% Let's look at the submissions. 00:05:32.600 --> 00:05:36.082 align:middle line:90% 00:05:36.082 --> 00:05:42.940 align:middle line:84% OK, so let's actually do a little bit of a browsing 00:05:42.940 --> 00:05:46.080 align:middle line:90% of models for sector. 00:05:46.080 --> 00:05:49.160 align:middle line:90% So this is Ross's we have. 00:05:49.160 --> 00:05:51.440 align:middle line:90% And let me just-- 00:05:51.440 --> 00:05:54.540 align:middle line:84% so we have the demand side, potential demand. 00:05:54.540 --> 00:05:59.016 align:middle line:90% 00:05:59.016 --> 00:06:00.260 align:middle line:90% Hmm, that's interesting. 00:06:00.260 --> 00:06:05.920 align:middle line:90% So, yeah. 00:06:05.920 --> 00:06:08.880 align:middle line:90% That's interesting. 00:06:08.880 --> 00:06:11.220 align:middle line:90% Yeah, so I like it. 00:06:11.220 --> 00:06:13.580 align:middle line:84% This is actually an interesting way of doing it. 00:06:13.580 --> 00:06:15.620 align:middle line:84% I have to look at the formulations. 00:06:15.620 --> 00:06:19.960 align:middle line:84% But basically, Ross is kind of formulating the customer 00:06:19.960 --> 00:06:23.600 align:middle line:84% sector and the demand part together in a single stock 00:06:23.600 --> 00:06:28.000 align:middle line:84% and flow structure, which should work fine. 00:06:28.000 --> 00:06:32.400 align:middle line:84% Then we have the reputation, we have the sales force, 00:06:32.400 --> 00:06:38.820 align:middle line:84% and we have the factory building part. 00:06:38.820 --> 00:06:43.060 align:middle line:84% So these are the right stocks in a high level view. 00:06:43.060 --> 00:06:43.780 align:middle line:90% That's great. 00:06:43.780 --> 00:06:45.900 align:middle line:90% Let's look at a few more. 00:06:45.900 --> 00:06:50.180 align:middle line:90% 00:06:50.180 --> 00:06:55.660 align:middle line:84% Album has production capacity and production capacity work 00:06:55.660 --> 00:07:03.720 align:middle line:84% in process, backlog, tracking of profits, potential customers, 00:07:03.720 --> 00:07:08.520 align:middle line:84% customers, and then the return rates. 00:07:08.520 --> 00:07:10.900 align:middle line:84% That depends on the differentiation. 00:07:10.900 --> 00:07:16.460 align:middle line:84% And then you can get the sales from that return rate and sales 00:07:16.460 --> 00:07:16.960 align:middle line:90% force. 00:07:16.960 --> 00:07:20.160 align:middle line:84% So pretty much the same key stock variables. 00:07:20.160 --> 00:07:20.660 align:middle line:90% Great. 00:07:20.660 --> 00:07:26.340 align:middle line:90% 00:07:26.340 --> 00:07:27.260 align:middle line:90% Oriana. 00:07:27.260 --> 00:07:31.180 align:middle line:90% We have sales force. 00:07:31.180 --> 00:07:35.740 align:middle line:84% We have reputation for delivery delay. 00:07:35.740 --> 00:07:42.580 align:middle line:84% We have customers who are moving between interested and current. 00:07:42.580 --> 00:07:47.640 align:middle line:84% And there is we should see what that part is done and then 00:07:47.640 --> 00:07:48.860 align:middle line:90% potentially interested. 00:07:48.860 --> 00:07:53.420 align:middle line:84% So there is a little bit of additional diffusion dynamics, 00:07:53.420 --> 00:07:54.840 align:middle line:90% maybe, here. 00:07:54.840 --> 00:07:57.880 align:middle line:90% We should look at that. 00:07:57.880 --> 00:08:01.840 align:middle line:84% So actually, can you tell what is the distinction 00:08:01.840 --> 00:08:03.180 align:middle line:90% between the three categories? 00:08:03.180 --> 00:08:06.840 align:middle line:84% So customers are the ones that currently have the product. 00:08:06.840 --> 00:08:10.640 align:middle line:84% Interested customers are the ones who want to buy it. 00:08:10.640 --> 00:08:12.880 align:middle line:90% So that creates-- 00:08:12.880 --> 00:08:15.600 align:middle line:84% AUDIENCE: So my idea was that there's whatever product there 00:08:15.600 --> 00:08:18.360 align:middle line:84% is, they're going to be like x could potentially 00:08:18.360 --> 00:08:19.260 align:middle line:90% have use for it. 00:08:19.260 --> 00:08:22.120 align:middle line:84% And then the sales force lets those people 00:08:22.120 --> 00:08:23.560 align:middle line:90% know that it exists. 00:08:23.560 --> 00:08:26.420 align:middle line:84% Then out of those people who now know that it exists, 00:08:26.420 --> 00:08:29.080 align:middle line:84% depending on the reputation of the company, 00:08:29.080 --> 00:08:30.760 align:middle line:84% a certain portion are going to buy. 00:08:30.760 --> 00:08:31.760 align:middle line:90% HAZHIR RAHMANDAD: I see. 00:08:31.760 --> 00:08:34.440 align:middle line:84% AUDIENCE: I never-- like, there should be a return back, 00:08:34.440 --> 00:08:37.360 align:middle line:84% like customers and interested customers back potentially, 00:08:37.360 --> 00:08:37.880 align:middle line:90% like-- 00:08:37.880 --> 00:08:38.880 align:middle line:90% HAZHIR RAHMANDAD: I see. 00:08:38.880 --> 00:08:41.520 align:middle line:84% So in a sense here, the sales force 00:08:41.520 --> 00:08:44.920 align:middle line:84% is not needed for repeat purchases, 00:08:44.920 --> 00:08:46.320 align:middle line:90% for replacement purchases? 00:08:46.320 --> 00:08:47.840 align:middle line:90% Is that a correct understanding? 00:08:47.840 --> 00:08:52.140 align:middle line:84% AUDIENCE: Yeah, but if I just added an additional flow pack, 00:08:52.140 --> 00:08:52.798 align:middle line:90% it would-- 00:08:52.798 --> 00:08:54.340 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so either you 00:08:54.340 --> 00:08:56.900 align:middle line:84% could take them here so that now they have to be, 00:08:56.900 --> 00:09:02.260 align:middle line:84% or the rate for purchase here could be bottlenecked 00:09:02.260 --> 00:09:04.220 align:middle line:84% by the availability of sales force 00:09:04.220 --> 00:09:08.417 align:middle line:84% rather than just being kind of a fraction based on reputation. 00:09:08.417 --> 00:09:10.000 align:middle line:84% So that would be another way to do it. 00:09:10.000 --> 00:09:12.860 align:middle line:84% But then you didn't even need the potentially interested 00:09:12.860 --> 00:09:13.440 align:middle line:90% customers. 00:09:13.440 --> 00:09:15.000 align:middle line:90% You could have one. 00:09:15.000 --> 00:09:15.560 align:middle line:90% Got it. 00:09:15.560 --> 00:09:20.300 align:middle line:84% And then you have the view for the backlog. 00:09:20.300 --> 00:09:25.340 align:middle line:84% So you have the backlog, factory, authorized number 00:09:25.340 --> 00:09:26.000 align:middle line:90% of factories. 00:09:26.000 --> 00:09:28.020 align:middle line:84% So this is capturing the delay in the process 00:09:28.020 --> 00:09:31.060 align:middle line:90% of generating factories. 00:09:31.060 --> 00:09:32.360 align:middle line:90% And the units are OK. 00:09:32.360 --> 00:09:33.200 align:middle line:90% That's great. 00:09:33.200 --> 00:09:34.740 align:middle line:84% AUDIENCE: May I just ask a question? 00:09:34.740 --> 00:09:35.698 align:middle line:90% HAZHIR RAHMANDAD: Sure. 00:09:35.698 --> 00:09:39.860 align:middle line:84% AUDIENCE: How did you deal with the depreciation 00:09:39.860 --> 00:09:43.120 align:middle line:90% rate of both vectors? 00:09:43.120 --> 00:09:45.620 align:middle line:84% HAZHIR RAHMANDAD: So this is a first order depreciation 00:09:45.620 --> 00:09:46.540 align:middle line:90% in this case. 00:09:46.540 --> 00:09:49.280 align:middle line:84% AUDIENCE: I meant in deciding how many to construct. 00:09:49.280 --> 00:09:51.280 align:middle line:84% HAZHIR RAHMANDAD: Oh, construction of factories. 00:09:51.280 --> 00:09:55.480 align:middle line:84% So this is coming from pressure to expand factories, which 00:09:55.480 --> 00:09:57.800 align:middle line:90% is coming from delivery delay. 00:09:57.800 --> 00:10:02.640 align:middle line:90% And then there is-- 00:10:02.640 --> 00:10:04.860 align:middle line:84% so we haven't yet gone through these equations. 00:10:04.860 --> 00:10:06.920 align:middle line:84% So I just wanted to get a quick overview 00:10:06.920 --> 00:10:09.760 align:middle line:84% of the structures or the stocks that different students are 00:10:09.760 --> 00:10:10.260 align:middle line:90% using. 00:10:10.260 --> 00:10:11.802 align:middle line:84% Then we can go through the equations. 00:10:11.802 --> 00:10:16.040 align:middle line:84% I think there might be issues that could be improved here. 00:10:16.040 --> 00:10:18.560 align:middle line:90% Let's see if this is-- 00:10:18.560 --> 00:10:19.960 align:middle line:90% searching. 00:10:19.960 --> 00:10:24.480 align:middle line:84% So we have the customers who are interested 00:10:24.480 --> 00:10:27.620 align:middle line:90% and those who have purchased. 00:10:27.620 --> 00:10:31.080 align:middle line:84% So this is pretty similar to the first couple of-- 00:10:31.080 --> 00:10:33.827 align:middle line:84% and then sales force impacts that purchase rate. 00:10:33.827 --> 00:10:35.660 align:middle line:84% It's kind of a little bit complicated there. 00:10:35.660 --> 00:10:37.800 align:middle line:84% We need to look at the equations. 00:10:37.800 --> 00:10:40.180 align:middle line:90% Orders come here. 00:10:40.180 --> 00:10:43.520 align:middle line:84% That creates a delivery delay through shipment, 00:10:43.520 --> 00:10:45.140 align:middle line:90% which depends on the stocks. 00:10:45.140 --> 00:10:46.920 align:middle line:90% Perceived capital stock. 00:10:46.920 --> 00:10:49.580 align:middle line:84% And then perceived delivery delay 00:10:49.580 --> 00:10:53.880 align:middle line:84% is impacting the expansion of capacity. 00:10:53.880 --> 00:10:55.400 align:middle line:90% So that also makes sense. 00:10:55.400 --> 00:11:02.780 align:middle line:84% There is the issue of capacity, capital stock adjustment time, 00:11:02.780 --> 00:11:05.060 align:middle line:84% and how that is captured or kind of work in process 00:11:05.060 --> 00:11:10.740 align:middle line:84% for capital is not explicit here, but otherwise looks good. 00:11:10.740 --> 00:11:14.500 align:middle line:84% Ann's model, let's Zoom in a little bit. 00:11:14.500 --> 00:11:20.420 align:middle line:84% So we have customers who are potential, 00:11:20.420 --> 00:11:24.900 align:middle line:90% then become interested. 00:11:24.900 --> 00:11:27.620 align:middle line:84% Then they're waiting for their orders. 00:11:27.620 --> 00:11:32.860 align:middle line:84% So Ann is accounting for this intermediate stage 00:11:32.860 --> 00:11:35.860 align:middle line:90% of waiting orders. 00:11:35.860 --> 00:11:37.740 align:middle line:84% We'll have to look at this equation 00:11:37.740 --> 00:11:43.760 align:middle line:84% to see how it maps into the backlog story. 00:11:43.760 --> 00:11:45.940 align:middle line:84% This adds a little bit more complexity, 00:11:45.940 --> 00:11:48.940 align:middle line:90% but core concepts are similar. 00:11:48.940 --> 00:11:52.180 align:middle line:90% Sales force is straightforward. 00:11:52.180 --> 00:11:55.300 align:middle line:84% And then cash reserve for sales force wages. 00:11:55.300 --> 00:12:00.520 align:middle line:84% So she's accounting for money in the bank, 00:12:00.520 --> 00:12:03.660 align:middle line:84% rather than directly taking the sales force from revenue. 00:12:03.660 --> 00:12:07.540 align:middle line:84% So that adds one more stock that others didn't have. 00:12:07.540 --> 00:12:10.380 align:middle line:84% It's OK, but may not be necessary. 00:12:10.380 --> 00:12:17.120 align:middle line:84% And then we have the facilities and death of the buildings, 00:12:17.120 --> 00:12:18.660 align:middle line:90% or death to buildings. 00:12:18.660 --> 00:12:20.260 align:middle line:90% That's very revolutionary. 00:12:20.260 --> 00:12:24.400 align:middle line:90% 00:12:24.400 --> 00:12:25.020 align:middle line:90% OK. 00:12:25.020 --> 00:12:30.720 align:middle line:90% 00:12:30.720 --> 00:12:37.300 align:middle line:84% And then we have here Hisam's products that are owned, 00:12:37.300 --> 00:12:40.500 align:middle line:84% orders that are in backlog, potential customers. 00:12:40.500 --> 00:12:42.500 align:middle line:84% So this is kind of capturing backlog 00:12:42.500 --> 00:12:46.540 align:middle line:84% similar to Ross in the same stock and flow 00:12:46.540 --> 00:12:48.640 align:middle line:90% structure as the customers. 00:12:48.640 --> 00:12:50.460 align:middle line:90% And that should work fine. 00:12:50.460 --> 00:12:55.860 align:middle line:84% We have the sales force driving the sales. 00:12:55.860 --> 00:12:59.560 align:middle line:84% It should also have some impact on-- 00:12:59.560 --> 00:13:04.460 align:middle line:90% 00:13:04.460 --> 00:13:07.320 align:middle line:84% yeah, so that kind of takes people back here. 00:13:07.320 --> 00:13:09.000 align:middle line:84% So there is kind of a question there, 00:13:09.000 --> 00:13:12.340 align:middle line:84% whether that's the best way to handle it or not. 00:13:12.340 --> 00:13:15.300 align:middle line:84% Then production capacity under development and production 00:13:15.300 --> 00:13:16.080 align:middle line:90% capacity. 00:13:16.080 --> 00:13:17.940 align:middle line:90% Great. 00:13:17.940 --> 00:13:21.480 align:middle line:90% And then did we check? 00:13:21.480 --> 00:13:21.980 align:middle line:90% Yes. 00:13:21.980 --> 00:13:26.420 align:middle line:90% 00:13:26.420 --> 00:13:30.140 align:middle line:84% Gerard's model has the reputation 00:13:30.140 --> 00:13:36.480 align:middle line:84% for delivery delay, I assume, adopters, and those 00:13:36.480 --> 00:13:38.060 align:middle line:90% who become willing to rebuy. 00:13:38.060 --> 00:13:45.080 align:middle line:84% So that's kind of the purchasing part. 00:13:45.080 --> 00:13:48.560 align:middle line:84% There is more complexity here in terms of potential adopters 00:13:48.560 --> 00:13:49.820 align:middle line:90% becoming interested. 00:13:49.820 --> 00:13:52.600 align:middle line:84% And then you can convince those interested into becoming 00:13:52.600 --> 00:13:54.440 align:middle line:90% adopters. 00:13:54.440 --> 00:13:55.580 align:middle line:90% That's fine. 00:13:55.580 --> 00:14:02.560 align:middle line:84% It's an interesting alternative because you're now 00:14:02.560 --> 00:14:05.360 align:middle line:84% using sales force to drive interest, and then 00:14:05.360 --> 00:14:07.980 align:middle line:84% these people, maybe with a delay, adopt. 00:14:07.980 --> 00:14:10.880 align:middle line:90% 00:14:10.880 --> 00:14:20.720 align:middle line:84% And that delay is regulating how likely it is for them to adopt. 00:14:20.720 --> 00:14:24.860 align:middle line:84% And alternatively, they may decide not to adopt. 00:14:24.860 --> 00:14:27.440 align:middle line:84% So this is a little bit of a different conceptualization, 00:14:27.440 --> 00:14:29.220 align:middle line:90% which might be pretty good. 00:14:29.220 --> 00:14:31.900 align:middle line:84% I have to look at it a little bit more closely. 00:14:31.900 --> 00:14:36.660 align:middle line:84% And then we have the similar production capacity dynamics 00:14:36.660 --> 00:14:37.160 align:middle line:90% here. 00:14:37.160 --> 00:14:39.780 align:middle line:90% 00:14:39.780 --> 00:14:43.380 align:middle line:90% This one is Alexander. 00:14:43.380 --> 00:14:45.340 align:middle line:90% So we-- yes. 00:14:45.340 --> 00:14:48.940 align:middle line:90% 00:14:48.940 --> 00:14:53.860 align:middle line:90% Backlog, installed base. 00:14:53.860 --> 00:14:55.820 align:middle line:84% So you have a little bit of a co-flow 00:14:55.820 --> 00:14:58.940 align:middle line:90% of products versus customers. 00:14:58.940 --> 00:15:00.980 align:middle line:90% And that's also nice to have. 00:15:00.980 --> 00:15:08.660 align:middle line:84% You account for different units, for people versus products. 00:15:08.660 --> 00:15:16.480 align:middle line:84% Sales force, and then you have in the other view the capital, 00:15:16.480 --> 00:15:19.600 align:middle line:84% the production capital as a single stock. 00:15:19.600 --> 00:15:22.460 align:middle line:90% 00:15:22.460 --> 00:15:35.320 align:middle line:84% And then we have Jason's, so potential customers 00:15:35.320 --> 00:15:40.520 align:middle line:84% versus customers with the sales rate. 00:15:40.520 --> 00:15:46.800 align:middle line:84% Order fulfillment has the delivery delay and backlog. 00:15:46.800 --> 00:15:50.800 align:middle line:84% And then delivery delay gets perceived by customers 00:15:50.800 --> 00:15:54.480 align:middle line:90% separately than by the company. 00:15:54.480 --> 00:15:57.360 align:middle line:90% All of that makes sense? 00:15:57.360 --> 00:15:59.180 align:middle line:90% Capacity acquisition. 00:15:59.180 --> 00:16:02.080 align:middle line:90% 00:16:02.080 --> 00:16:06.740 align:middle line:84% So we have the capacity and capacity under development. 00:16:06.740 --> 00:16:11.400 align:middle line:84% So that's also fairly clean and standard. 00:16:11.400 --> 00:16:15.380 align:middle line:84% And then sales force is just the stock sales force, 00:16:15.380 --> 00:16:18.100 align:middle line:84% which makes sense, with some adjustment time. 00:16:18.100 --> 00:16:19.140 align:middle line:90% OK, good. 00:16:19.140 --> 00:16:22.400 align:middle line:90% 00:16:22.400 --> 00:16:24.420 align:middle line:90% This is Cathy's. 00:16:24.420 --> 00:16:25.920 align:middle line:90% Let me see if I missed somebody. 00:16:25.920 --> 00:16:26.880 align:middle line:90% No. 00:16:26.880 --> 00:16:28.200 align:middle line:90% Cathy's. 00:16:28.200 --> 00:16:32.140 align:middle line:90% So accounting is the last group. 00:16:32.140 --> 00:16:39.700 align:middle line:84% So we have customers who are potential with purchase 00:16:39.700 --> 00:16:43.180 align:middle line:90% orders with the product. 00:16:43.180 --> 00:16:44.620 align:middle line:84% So customers with purchase orders 00:16:44.620 --> 00:16:47.840 align:middle line:84% are the ones waiting in the backlog of stock, basically. 00:16:47.840 --> 00:16:48.760 align:middle line:90% So they have the-- 00:16:48.760 --> 00:16:51.360 align:middle line:90% 00:16:51.360 --> 00:16:51.860 align:middle line:90% yeah. 00:16:51.860 --> 00:16:56.460 align:middle line:84% And then customers with the product, 00:16:56.460 --> 00:16:59.660 align:middle line:84% and then customers who don't have the product, 00:16:59.660 --> 00:17:02.480 align:middle line:84% and then they need to decide to come back. 00:17:02.480 --> 00:17:06.040 align:middle line:84% So this takes care of backlog and-- 00:17:06.040 --> 00:17:08.780 align:middle line:90% 00:17:08.780 --> 00:17:11.380 align:middle line:84% yeah, so all of that is accounted for. 00:17:11.380 --> 00:17:13.040 align:middle line:84% It's actually an interesting piece. 00:17:13.040 --> 00:17:17.380 align:middle line:84% So a quick look at this structure. 00:17:17.380 --> 00:17:22.380 align:middle line:90% What it tells me is that-- 00:17:22.380 --> 00:17:30.300 align:middle line:84% so this is a quick deeper dive into to a heuristic you can use. 00:17:30.300 --> 00:17:33.320 align:middle line:84% So I look at this stock and flow structure. 00:17:33.320 --> 00:17:36.800 align:middle line:84% If you put all of the stocks together as a bigger stock, 00:17:36.800 --> 00:17:38.180 align:middle line:90% what are the inflows? 00:17:38.180 --> 00:17:41.600 align:middle line:90% 00:17:41.600 --> 00:17:45.060 align:middle line:84% Do you see any inflow to that full stock and flow structure? 00:17:45.060 --> 00:17:45.720 align:middle line:90% Nothing. 00:17:45.720 --> 00:17:48.840 align:middle line:90% What about an outflow? 00:17:48.840 --> 00:17:50.140 align:middle line:90% There is an outflow. 00:17:50.140 --> 00:17:53.600 align:middle line:84% So as soon as I see that, it tells me 00:17:53.600 --> 00:17:58.480 align:middle line:84% that ultimately, this market will be drained to nothing. 00:17:58.480 --> 00:18:05.560 align:middle line:84% And that is kind of a little bit of an unrealistic setup. 00:18:05.560 --> 00:18:08.800 align:middle line:84% So kind of zooming out and thinking 00:18:08.800 --> 00:18:14.320 align:middle line:84% of a mix of stocks and flows, whether they are together, 00:18:14.320 --> 00:18:18.380 align:middle line:84% conserving whatever you are trying to do or not, 00:18:18.380 --> 00:18:21.760 align:middle line:84% and they potentially lead to an absorbing state that 00:18:21.760 --> 00:18:24.878 align:middle line:84% is nobody or nothing, that's kind of important to have 00:18:24.878 --> 00:18:25.920 align:middle line:90% in the back of your mind. 00:18:25.920 --> 00:18:27.560 align:middle line:84% Now, sometimes that is realistic. 00:18:27.560 --> 00:18:29.460 align:middle line:84% That is, for some things there is actually 00:18:29.460 --> 00:18:31.380 align:middle line:84% an absorbing state where there is nobody left 00:18:31.380 --> 00:18:33.640 align:middle line:90% or no money left, or whatnot. 00:18:33.640 --> 00:18:36.940 align:middle line:84% But in this case, I think that's not, probably, 00:18:36.940 --> 00:18:40.360 align:middle line:90% the most realistic alternative. 00:18:40.360 --> 00:18:44.560 align:middle line:84% And then you have rookie sales force, experienced sales force. 00:18:44.560 --> 00:18:49.400 align:middle line:84% So there is kind of additional dimension of rookie dynamics, 00:18:49.400 --> 00:18:51.820 align:middle line:84% which are pretty relevant, even though it wasn't 00:18:51.820 --> 00:18:53.440 align:middle line:90% part of the problem definition. 00:18:53.440 --> 00:18:55.260 align:middle line:90% That's nice to have. 00:18:55.260 --> 00:19:00.960 align:middle line:84% And then you have the backlog and the manufacturing capacity. 00:19:00.960 --> 00:19:03.240 align:middle line:84% Now, this is also a little bit more complex here. 00:19:03.240 --> 00:19:05.280 align:middle line:84% You have units in production queue. 00:19:05.280 --> 00:19:10.820 align:middle line:84% So this is kind of the back, the normal backlog, the way 00:19:10.820 --> 00:19:12.480 align:middle line:90% I would have thought about it. 00:19:12.480 --> 00:19:14.150 align:middle line:84% But then you have units in production. 00:19:14.150 --> 00:19:16.400 align:middle line:84% So we should look at these in a little bit more depth. 00:19:16.400 --> 00:19:18.660 align:middle line:90% I don't know. 00:19:18.660 --> 00:19:22.600 align:middle line:84% AUDIENCE: I was thinking that the units in production queue, 00:19:22.600 --> 00:19:26.800 align:middle line:84% that's just-- like, that's the list of customers or orders I 00:19:26.800 --> 00:19:27.580 align:middle line:90% need to fill. 00:19:27.580 --> 00:19:29.520 align:middle line:84% And then I'm limited by my capacity 00:19:29.520 --> 00:19:34.440 align:middle line:84% on actually starting the manufacturing process. 00:19:34.440 --> 00:19:35.700 align:middle line:90% So I separated those. 00:19:35.700 --> 00:19:38.960 align:middle line:84% HAZHIR RAHMANDAD: So your manufacturing process kind 00:19:38.960 --> 00:19:43.920 align:middle line:90% of limits the production starts. 00:19:43.920 --> 00:19:49.280 align:middle line:84% And after a fixed time, they end, basically. 00:19:49.280 --> 00:19:51.780 align:middle line:84% Yeah, and your total backlog would be the sum of these two. 00:19:51.780 --> 00:19:54.940 align:middle line:84% So you would have the delivery delay calculated correctly, 00:19:54.940 --> 00:19:57.800 align:middle line:90% that average delivery delay. 00:19:57.800 --> 00:19:59.060 align:middle line:90% Yeah, that's fine. 00:19:59.060 --> 00:20:01.872 align:middle line:90% 00:20:01.872 --> 00:20:03.540 align:middle line:90% It's sounds right. 00:20:03.540 --> 00:20:06.200 align:middle line:84% It's probably overkill, in the sense 00:20:06.200 --> 00:20:10.620 align:middle line:84% that you don't need two stocks to capture this process. 00:20:10.620 --> 00:20:12.600 align:middle line:84% Basically, if you had the outflow 00:20:12.600 --> 00:20:15.520 align:middle line:84% to be regulated by manufacturing capacity, 00:20:15.520 --> 00:20:18.400 align:middle line:84% that would do the same thing, ultimately, for you. 00:20:18.400 --> 00:20:19.175 align:middle line:90% Very similar. 00:20:19.175 --> 00:20:20.300 align:middle line:90% Not exactly the same thing. 00:20:20.300 --> 00:20:22.800 align:middle line:84% So mathematically, there is some difference between the two, 00:20:22.800 --> 00:20:26.520 align:middle line:84% but I wouldn't worry too much about that difference here. 00:20:26.520 --> 00:20:27.280 align:middle line:90% OK, good. 00:20:27.280 --> 00:20:31.220 align:middle line:90% 00:20:31.220 --> 00:20:33.100 align:middle line:90% OK. 00:20:33.100 --> 00:20:42.860 align:middle line:84% So Andreeva, we have sales force, 00:20:42.860 --> 00:20:49.940 align:middle line:84% customers with potential versus current with some return rate. 00:20:49.940 --> 00:20:52.900 align:middle line:84% And here, you are not distinguishing 00:20:52.900 --> 00:20:56.200 align:middle line:84% between first-time customers versus those returning ones. 00:20:56.200 --> 00:21:02.420 align:middle line:84% So that has some implications in terms of formulations, 00:21:02.420 --> 00:21:05.780 align:middle line:90% but that's fine as well. 00:21:05.780 --> 00:21:10.060 align:middle line:90% Backlog is fairly standard. 00:21:10.060 --> 00:21:13.900 align:middle line:84% Factories under construction, constructed. 00:21:13.900 --> 00:21:15.857 align:middle line:84% And you're calling these factories 00:21:15.857 --> 00:21:16.940 align:middle line:90% depreciating, [INAUDIBLE]. 00:21:16.940 --> 00:21:20.620 align:middle line:84% Is it a stock that you want to track? 00:21:20.620 --> 00:21:22.140 align:middle line:90% Do you do anything with it? 00:21:22.140 --> 00:21:23.520 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:21:23.520 --> 00:21:24.780 align:middle line:84% HAZHIR RAHMANDAD: Yeah, it's not being used? 00:21:24.780 --> 00:21:25.280 align:middle line:90% Yeah, OK. 00:21:25.280 --> 00:21:27.460 align:middle line:90% That's fine. 00:21:27.460 --> 00:21:27.960 align:middle line:90% OK. 00:21:27.960 --> 00:21:31.120 align:middle line:90% That all makes sense. 00:21:31.120 --> 00:21:32.380 align:middle line:90% Did you see the sales force? 00:21:32.380 --> 00:21:33.600 align:middle line:90% Yes. 00:21:33.600 --> 00:21:34.220 align:middle line:90% OK. 00:21:34.220 --> 00:21:43.640 align:middle line:84% And then Ali's model has the backlog, delivery delay, 00:21:43.640 --> 00:21:45.960 align:middle line:90% by the market, by the company. 00:21:45.960 --> 00:21:48.960 align:middle line:84% So you are taking delivery delay perceived by the market 00:21:48.960 --> 00:21:52.480 align:middle line:84% from delivery delay perceived by the company. 00:21:52.480 --> 00:21:53.958 align:middle line:90% That's an interesting idea. 00:21:53.958 --> 00:21:54.500 align:middle line:90% I don't know. 00:21:54.500 --> 00:21:57.100 align:middle line:84% I think you could calculate them separately, 00:21:57.100 --> 00:22:00.560 align:middle line:84% I think as [? Katy ?] I think had done. 00:22:00.560 --> 00:22:05.200 align:middle line:84% But whether it is taken from the perception of the company 00:22:05.200 --> 00:22:07.080 align:middle line:84% is a little bit of a harder sell. 00:22:07.080 --> 00:22:10.600 align:middle line:84% Because the company has access to a bunch of information 00:22:10.600 --> 00:22:13.940 align:middle line:84% that it's not going to directly share to inform the customers. 00:22:13.940 --> 00:22:16.720 align:middle line:84% Customers are using their own channels of information. 00:22:16.720 --> 00:22:21.420 align:middle line:90% So as such, I wouldn't take-- 00:22:21.420 --> 00:22:22.900 align:middle line:90% yeah. 00:22:22.900 --> 00:22:25.960 align:middle line:84% Capacity, capacity under construction. 00:22:25.960 --> 00:22:27.780 align:middle line:90% That's a standard. 00:22:27.780 --> 00:22:31.000 align:middle line:84% And then we have the customer structure here. 00:22:31.000 --> 00:22:34.860 align:middle line:90% So this is current customers. 00:22:34.860 --> 00:22:36.820 align:middle line:90% They become potential. 00:22:36.820 --> 00:22:39.640 align:middle line:90% You pitch to customers. 00:22:39.640 --> 00:22:43.780 align:middle line:84% And then they can become current customers. 00:22:43.780 --> 00:22:47.240 align:middle line:84% So you have the pitching process kind of very explicit. 00:22:47.240 --> 00:22:51.100 align:middle line:84% And that is what is coming from the sales force. 00:22:51.100 --> 00:22:55.180 align:middle line:84% So this is allowing you to separate the step of acceptance 00:22:55.180 --> 00:23:02.060 align:middle line:84% that is driven by delivery delay from the step of trying 00:23:02.060 --> 00:23:08.660 align:middle line:90% to influence pitching. 00:23:08.660 --> 00:23:09.240 align:middle line:90% That's good. 00:23:09.240 --> 00:23:12.580 align:middle line:84% That's one way to kind of overcome that potentially 00:23:12.580 --> 00:23:14.540 align:middle line:84% a little bit more challenging formulation 00:23:14.540 --> 00:23:19.360 align:middle line:90% and keeping it very operational. 00:23:19.360 --> 00:23:19.880 align:middle line:90% Good. 00:23:19.880 --> 00:23:25.980 align:middle line:84% Sales force, recent revenue backlog-- 00:23:25.980 --> 00:23:26.940 align:middle line:90% and we saw the rest. 00:23:26.940 --> 00:23:27.440 align:middle line:90% OK. 00:23:27.440 --> 00:23:27.980 align:middle line:90% Great. 00:23:27.980 --> 00:23:37.520 align:middle line:90% 00:23:37.520 --> 00:23:41.800 align:middle line:90% Finally, we have Jake's. 00:23:41.800 --> 00:23:45.280 align:middle line:90% So how come Jake's comes last? 00:23:45.280 --> 00:23:50.580 align:middle line:84% Daniels is not the first letter [INAUDIBLE] of it. 00:23:50.580 --> 00:23:52.400 align:middle line:84% AUDIENCE: I think you opened it first. 00:23:52.400 --> 00:23:53.733 align:middle line:90% HAZHIR RAHMANDAD: Somehow, yeah. 00:23:53.733 --> 00:23:56.360 align:middle line:84% But I thought I had sorted things 00:23:56.360 --> 00:23:59.180 align:middle line:90% based on alphabetical order. 00:23:59.180 --> 00:23:59.680 align:middle line:90% OK. 00:23:59.680 --> 00:24:10.380 align:middle line:84% So we have the backlog, work in progress, 00:24:10.380 --> 00:24:13.420 align:middle line:90% finished goods, delivered goods. 00:24:13.420 --> 00:24:19.180 align:middle line:84% So you have this additional detail on the delivery of goods 00:24:19.180 --> 00:24:21.120 align:middle line:90% just based on some delays. 00:24:21.120 --> 00:24:27.420 align:middle line:84% So the capacity is what does the production, 00:24:27.420 --> 00:24:32.180 align:middle line:84% but these delays introduce some additional complexity 00:24:32.180 --> 00:24:40.200 align:middle line:90% in terms of delivery delays. 00:24:40.200 --> 00:24:41.160 align:middle line:90% That's interesting. 00:24:41.160 --> 00:24:44.420 align:middle line:84% So actually, let's now do a little bit of a deeper dive here 00:24:44.420 --> 00:24:46.280 align:middle line:84% and understand what's happening here. 00:24:46.280 --> 00:24:51.540 align:middle line:90% So units are all fine. 00:24:51.540 --> 00:24:54.260 align:middle line:90% So the story of-- 00:24:54.260 --> 00:24:55.500 align:middle line:90% people purchase. 00:24:55.500 --> 00:24:59.060 align:middle line:84% This goes into outstanding order backlog. 00:24:59.060 --> 00:25:04.300 align:middle line:84% Production capacity allow us to process this. 00:25:04.300 --> 00:25:07.900 align:middle line:84% Now, you are combining kind of probably two things 00:25:07.900 --> 00:25:11.940 align:middle line:90% into production. 00:25:11.940 --> 00:25:17.840 align:middle line:84% AUDIENCE: Yeah, my thought was like, 00:25:17.840 --> 00:25:22.200 align:middle line:84% you're going to be able to start work in progress-- 00:25:22.200 --> 00:25:24.120 align:middle line:84% basically run at full capacity if you need to 00:25:24.120 --> 00:25:26.637 align:middle line:84% just based on the backlog, but you can't go above that. 00:25:26.637 --> 00:25:27.512 align:middle line:90% HAZHIR RAHMANDAD: OK. 00:25:27.512 --> 00:25:29.120 align:middle line:84% So the equation here has a little bit 00:25:29.120 --> 00:25:30.700 align:middle line:90% of a unit problem for me. 00:25:30.700 --> 00:25:35.560 align:middle line:84% So I think of available capacity to be in widgets per year 00:25:35.560 --> 00:25:37.080 align:middle line:90% or units per year. 00:25:37.080 --> 00:25:41.120 align:middle line:84% But it seems you're using some definition of capacity that 00:25:41.120 --> 00:25:44.900 align:middle line:84% is in the units of units, not in the units of units per year. 00:25:44.900 --> 00:25:47.720 align:middle line:90% Can you tell what is-- 00:25:47.720 --> 00:25:48.420 align:middle line:90% AUDIENCE: Yeah. 00:25:48.420 --> 00:25:50.320 align:middle line:90% In my head, capacity is like-- 00:25:50.320 --> 00:25:51.900 align:middle line:84% it's not necessarily units per year, 00:25:51.900 --> 00:25:57.360 align:middle line:84% but you can have x amount of widgets currently 00:25:57.360 --> 00:26:00.440 align:middle line:90% on the product line basically. 00:26:00.440 --> 00:26:05.520 align:middle line:84% Like, at any given time, that is a unit that can be somewhere. 00:26:05.520 --> 00:26:06.520 align:middle line:90% HAZHIR RAHMANDAD: I see. 00:26:06.520 --> 00:26:12.220 align:middle line:84% So you are saying you have a number of factories, which 00:26:12.220 --> 00:26:14.460 align:middle line:90% you just call it as a factory. 00:26:14.460 --> 00:26:17.860 align:middle line:84% Then unit capacity per factory-- you're 00:26:17.860 --> 00:26:20.220 align:middle line:90% calling it unit per factory. 00:26:20.220 --> 00:26:26.120 align:middle line:84% So you are kind of removing time from the production capacity. 00:26:26.120 --> 00:26:31.100 align:middle line:84% You are translating it into work in process capacity in terms 00:26:31.100 --> 00:26:37.700 align:middle line:84% of number of units that are on the factory's floor. 00:26:37.700 --> 00:26:39.077 align:middle line:90% AUDIENCE: Yeah. 00:26:39.077 --> 00:26:41.660 align:middle line:84% HAZHIR RAHMANDAD: And then you have a processing time of sort, 00:26:41.660 --> 00:26:47.420 align:middle line:84% time to process orders, that basically captures how fast you 00:26:47.420 --> 00:26:49.900 align:middle line:90% drain that capacity. 00:26:49.900 --> 00:26:52.200 align:middle line:90% So this could work. 00:26:52.200 --> 00:26:55.340 align:middle line:84% I think technically it is not wrong, 00:26:55.340 --> 00:26:58.080 align:middle line:84% but it is not the standard way we typically formulate these. 00:26:58.080 --> 00:27:01.060 align:middle line:84% So the standard way is to think of capacity 00:27:01.060 --> 00:27:07.180 align:middle line:84% in units per unit of time-- so widget per year or month. 00:27:07.180 --> 00:27:08.980 align:middle line:90% This is our production capacity. 00:27:08.980 --> 00:27:11.400 align:middle line:84% That production capacity then would directly 00:27:11.400 --> 00:27:14.280 align:middle line:84% go into your production start as one of the constraints 00:27:14.280 --> 00:27:15.500 align:middle line:90% of that minimum function. 00:27:15.500 --> 00:27:18.400 align:middle line:84% So it would be minimum of the production capacity 00:27:18.400 --> 00:27:27.160 align:middle line:84% and the outstanding order backlog divided by time 00:27:27.160 --> 00:27:29.720 align:middle line:84% to process order or minimum time to process order. 00:27:29.720 --> 00:27:31.800 align:middle line:84% Because then it would be minimum time 00:27:31.800 --> 00:27:36.760 align:middle line:84% to process order rather than time to process order. 00:27:36.760 --> 00:27:38.720 align:middle line:84% The definition of this thing right 00:27:38.720 --> 00:27:41.720 align:middle line:84% now is kind of doing two very different things. 00:27:41.720 --> 00:27:44.800 align:middle line:84% So time to process orders, on the one hand, 00:27:44.800 --> 00:27:48.600 align:middle line:84% in your current formulation, is acting as minimum time 00:27:48.600 --> 00:27:52.880 align:middle line:84% to process order to put an upper bound on how fast you can drain 00:27:52.880 --> 00:27:57.000 align:middle line:84% or you can produce given how much backlog you have. 00:27:57.000 --> 00:27:59.300 align:middle line:84% So even if you have infinite capacity, 00:27:59.300 --> 00:28:03.240 align:middle line:84% you cannot drain the outstanding order backlog with an infinite 00:28:03.240 --> 00:28:08.780 align:middle line:84% speed because you have a constraint in terms of number 00:28:08.780 --> 00:28:11.660 align:middle line:90% of items in the backlog. 00:28:11.660 --> 00:28:15.460 align:middle line:84% Then there is a separate time constant for how fast 00:28:15.460 --> 00:28:18.980 align:middle line:84% your factory produces the units it has 00:28:18.980 --> 00:28:23.420 align:middle line:84% in its working process, which is the way you were conceptualizing 00:28:23.420 --> 00:28:25.840 align:middle line:84% it in our first part of the conversation, 00:28:25.840 --> 00:28:30.020 align:middle line:84% going back to the available capacity. 00:28:30.020 --> 00:28:32.440 align:middle line:84% So if you're saying available capacity, 00:28:32.440 --> 00:28:35.540 align:middle line:84% it is the total number of units on the factory floor, 00:28:35.540 --> 00:28:39.620 align:middle line:84% then it takes, for example, 10 days, or 10 months, 00:28:39.620 --> 00:28:42.260 align:middle line:90% or whatever to process that. 00:28:42.260 --> 00:28:45.460 align:middle line:84% You would divide that capacity by that processing time 00:28:45.460 --> 00:28:49.580 align:middle line:84% to get your average rate of outflow 00:28:49.580 --> 00:28:51.700 align:middle line:84% from the-- and right now, those two 00:28:51.700 --> 00:28:54.740 align:middle line:84% are conflated into this single parameter. 00:28:54.740 --> 00:28:58.100 align:middle line:84% So that is one of the places where these two time 00:28:58.100 --> 00:29:00.320 align:middle line:84% constants should not be the same but are the same. 00:29:00.320 --> 00:29:00.960 align:middle line:90% So that's-- 00:29:00.960 --> 00:29:04.480 align:middle line:90% 00:29:04.480 --> 00:29:08.040 align:middle line:84% AUDIENCE: Or, yeah, I think the actual time 00:29:08.040 --> 00:29:10.600 align:middle line:84% to complete production is going to be the time 00:29:10.600 --> 00:29:12.943 align:middle line:84% to complete production after work in progress. 00:29:12.943 --> 00:29:14.360 align:middle line:84% It's literally just the time to go 00:29:14.360 --> 00:29:18.922 align:middle line:90% from order to get on the floor. 00:29:18.922 --> 00:29:19.880 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 00:29:19.880 --> 00:29:20.672 align:middle line:90% AUDIENCE: But yeah. 00:29:20.672 --> 00:29:22.420 align:middle line:84% HAZHIR RAHMANDAD: So that would be-- 00:29:22.420 --> 00:29:25.200 align:middle line:90% 00:29:25.200 --> 00:29:28.720 align:middle line:84% so that's a little bit tricky because-- 00:29:28.720 --> 00:29:33.160 align:middle line:84% so now let's assume your production capacity 00:29:33.160 --> 00:29:34.840 align:middle line:90% is very small. 00:29:34.840 --> 00:29:38.940 align:middle line:84% Well, because it is very small, then you don't start production. 00:29:38.940 --> 00:29:42.680 align:middle line:84% But once you start it, it will go through this time 00:29:42.680 --> 00:29:43.860 align:middle line:90% to complete production. 00:29:43.860 --> 00:29:46.520 align:middle line:84% So basically, time to process orders 00:29:46.520 --> 00:29:50.440 align:middle line:84% plus time to complete production is the total time for production 00:29:50.440 --> 00:29:51.700 align:middle line:90% in your conceptualization. 00:29:51.700 --> 00:29:53.180 align:middle line:84% So you have the degrees of freedom 00:29:53.180 --> 00:29:56.480 align:middle line:84% you need to basically get this to work. 00:29:56.480 --> 00:29:59.320 align:middle line:90% That makes sense. 00:29:59.320 --> 00:30:03.780 align:middle line:84% So it's more kind of following the standard formulation would 00:30:03.780 --> 00:30:05.360 align:middle line:90% make it a little bit cleaner. 00:30:05.360 --> 00:30:09.240 align:middle line:84% But also, you don't then need all of these intermediate steps. 00:30:09.240 --> 00:30:11.780 align:middle line:84% Basically, here, you are adding two additional delays 00:30:11.780 --> 00:30:13.180 align:middle line:90% into the delivery. 00:30:13.180 --> 00:30:19.340 align:middle line:84% And so this is kind of another important architectural choice. 00:30:19.340 --> 00:30:21.700 align:middle line:84% What are the time constants for my stocks? 00:30:21.700 --> 00:30:24.460 align:middle line:84% The stocks for finished good delivery, 00:30:24.460 --> 00:30:27.622 align:middle line:84% work in process delivery-- if they are relatively fast, 00:30:27.622 --> 00:30:28.580 align:middle line:90% then I don't need them. 00:30:28.580 --> 00:30:33.527 align:middle line:84% If I think they are fairly long-- so this is 0.2 months-- 00:30:33.527 --> 00:30:34.360 align:middle line:90% AUDIENCE: 0.2 years. 00:30:34.360 --> 00:30:35.818 align:middle line:84% HAZHIR RAHMANDAD: 0.2 years, sorry. 00:30:35.818 --> 00:30:38.140 align:middle line:90% This is 0.125 years. 00:30:38.140 --> 00:30:43.300 align:middle line:84% So these are in the order of a month or two. 00:30:43.300 --> 00:30:49.960 align:middle line:84% If my other time constants are kind of in that order, 00:30:49.960 --> 00:30:53.580 align:middle line:84% then it makes sense to separate them or be explicit about them. 00:30:53.580 --> 00:30:56.560 align:middle line:84% If other time constants are a little bit longer, 00:30:56.560 --> 00:31:01.320 align:middle line:84% then it makes sense to bring all of them 00:31:01.320 --> 00:31:05.680 align:middle line:84% together into a single structure. 00:31:05.680 --> 00:31:08.140 align:middle line:84% So time to perceive delivery delay-- 00:31:08.140 --> 00:31:09.740 align:middle line:90% you have 0.25 years as well. 00:31:09.740 --> 00:31:11.920 align:middle line:90% So it's in the same ballpark. 00:31:11.920 --> 00:31:14.700 align:middle line:84% So maybe this separation is justified-- 00:31:14.700 --> 00:31:22.460 align:middle line:84% or reputation adjustment time a year for the outsiders. 00:31:22.460 --> 00:31:23.150 align:middle line:90% OK. 00:31:23.150 --> 00:31:26.100 align:middle line:84% So I think this part is working fine, 00:31:26.100 --> 00:31:28.780 align:middle line:84% at least in the basic physics of it. 00:31:28.780 --> 00:31:34.940 align:middle line:84% Now you have shipping going into received products. 00:31:34.940 --> 00:31:35.708 align:middle line:90% AUDIENCE: Yeah. 00:31:35.708 --> 00:31:38.000 align:middle line:84% You mentioned earlier it's sort of split between people 00:31:38.000 --> 00:31:39.140 align:middle line:90% and units-- it's not. 00:31:39.140 --> 00:31:40.560 align:middle line:84% These are just in terms of units. 00:31:40.560 --> 00:31:43.060 align:middle line:84% Like, the interest of market is just going to be units here. 00:31:43.060 --> 00:31:45.028 align:middle line:84% It's like a potential holder, not a-- yeah. 00:31:45.028 --> 00:31:45.986 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 00:31:45.986 --> 00:31:48.420 align:middle line:84% So it would have been easy to make that transition, 00:31:48.420 --> 00:31:51.080 align:middle line:84% but you don't also need to if you are assuming 00:31:51.080 --> 00:31:52.560 align:middle line:90% each person is buying one unit. 00:31:52.560 --> 00:31:53.460 align:middle line:90% That's fine. 00:31:53.460 --> 00:31:56.260 align:middle line:84% And I think for this purpose, we are not changing that. 00:31:56.260 --> 00:31:58.300 align:middle line:84% So it's OK to have the same units. 00:31:58.300 --> 00:32:00.660 align:middle line:90% So when you deliver something-- 00:32:00.660 --> 00:32:02.720 align:middle line:84% so this is market waiting for orders. 00:32:02.720 --> 00:32:06.900 align:middle line:84% This is really what I think should be deciding the delivery 00:32:06.900 --> 00:32:08.580 align:middle line:90% delay. 00:32:08.580 --> 00:32:10.600 align:middle line:84% So where does the delivery delay come from? 00:32:10.600 --> 00:32:13.660 align:middle line:84% So right now, you have it from time 00:32:13.660 --> 00:32:18.500 align:middle line:84% in backlog based on orders outstanding in backlog divided 00:32:18.500 --> 00:32:21.380 align:middle line:84% by the production start plus time 00:32:21.380 --> 00:32:24.440 align:middle line:84% to complete production plus time to ship. 00:32:24.440 --> 00:32:28.117 align:middle line:84% So you're adding these three steps together. 00:32:28.117 --> 00:32:28.700 align:middle line:90% AUDIENCE: Yes. 00:32:28.700 --> 00:32:31.620 align:middle line:84% HAZHIR RAHMANDAD: So that's probably 00:32:31.620 --> 00:32:33.740 align:middle line:84% the same as what you would do if you 00:32:33.740 --> 00:32:38.560 align:middle line:84% had market waiting for orders divided by received products. 00:32:38.560 --> 00:32:41.620 align:middle line:84% So that probably would give the same answer. 00:32:41.620 --> 00:32:43.060 align:middle line:90% It's a good thing to test that. 00:32:43.060 --> 00:32:43.840 align:middle line:90% Alexander? 00:32:43.840 --> 00:32:46.100 align:middle line:84% AUDIENCE: [INAUDIBLE] ask-- back to the WIP 00:32:46.100 --> 00:32:49.860 align:middle line:84% and finished goods [? chain. ?] Shouldn't the latter two flows 00:32:49.860 --> 00:32:52.300 align:middle line:84% be constrained by capacity in some way? 00:32:52.300 --> 00:32:54.300 align:middle line:84% Because I'm thinking that-- let's say capacity 00:32:54.300 --> 00:32:56.220 align:middle line:90% is initially super high. 00:32:56.220 --> 00:32:59.420 align:middle line:84% So then you have a lot in the work in progress. 00:32:59.420 --> 00:33:00.882 align:middle line:84% And then, suddenly, capacity drops. 00:33:00.882 --> 00:33:02.340 align:middle line:84% HAZHIR RAHMANDAD: Yeah, absolutely. 00:33:02.340 --> 00:33:06.560 align:middle line:84% So this formulation is not fully consistent 00:33:06.560 --> 00:33:08.760 align:middle line:84% with those scenarios-- that it is assuming 00:33:08.760 --> 00:33:11.960 align:middle line:84% that once you are in the production process, 00:33:11.960 --> 00:33:14.760 align:middle line:84% the production capacity wouldn't change anymore 00:33:14.760 --> 00:33:16.480 align:middle line:90% for those batches. 00:33:16.480 --> 00:33:19.080 align:middle line:84% So that is, basically, the time constant for changing 00:33:19.080 --> 00:33:22.760 align:middle line:84% the production capacity is much longer than the time 00:33:22.760 --> 00:33:25.780 align:middle line:84% constant for changing the units in the production. 00:33:25.780 --> 00:33:29.700 align:middle line:84% But it does create extreme condition problems, as you say. 00:33:29.700 --> 00:33:30.960 align:middle line:90% Exactly. 00:33:30.960 --> 00:33:34.640 align:middle line:84% And that's one of the reasons the standard formulation just 00:33:34.640 --> 00:33:37.560 align:middle line:84% says, have one place where production capacity 00:33:37.560 --> 00:33:41.020 align:middle line:84% is operational, typically at the very outflow of that backlog. 00:33:41.020 --> 00:33:44.920 align:middle line:84% And then don't worry about the intermediate steps. 00:33:44.920 --> 00:33:45.520 align:middle line:90% Yeah. 00:33:45.520 --> 00:33:46.520 align:middle line:90% So you're right. 00:33:46.520 --> 00:33:49.960 align:middle line:84% I think you can come up with extreme conditions like that 00:33:49.960 --> 00:33:51.700 align:middle line:84% that will break this formulation. 00:33:51.700 --> 00:33:52.200 align:middle line:90% OK. 00:33:52.200 --> 00:33:56.000 align:middle line:84% So that delivery delay should work fine. 00:33:56.000 --> 00:33:59.820 align:middle line:84% You get the delivery delay perception here. 00:33:59.820 --> 00:34:03.840 align:middle line:84% Then you have a delay reputation that comes from that perception. 00:34:03.840 --> 00:34:06.580 align:middle line:84% So there is a little bit of a question, 00:34:06.580 --> 00:34:14.380 align:middle line:84% what is this perception doing in factory construction 00:34:14.380 --> 00:34:15.080 align:middle line:90% sensitivity? 00:34:15.080 --> 00:34:17.460 align:middle line:84% So this is similar to what Ali had, 00:34:17.460 --> 00:34:22.280 align:middle line:84% which is there is a perception by company, 00:34:22.280 --> 00:34:24.860 align:middle line:84% and then there is a perception of delivery delay 00:34:24.860 --> 00:34:26.719 align:middle line:90% by the customers. 00:34:26.719 --> 00:34:28.940 align:middle line:84% But you are getting the perception of delivery delay 00:34:28.940 --> 00:34:34.940 align:middle line:84% by customers, which is this delay reputation, from perceived 00:34:34.940 --> 00:34:36.719 align:middle line:90% delivery delay by the company. 00:34:36.719 --> 00:34:39.500 align:middle line:84% And these two, again, are a little bit different pathways 00:34:39.500 --> 00:34:40.920 align:middle line:90% information-wise. 00:34:40.920 --> 00:34:46.080 align:middle line:84% So I would perhaps have separate single delays, 00:34:46.080 --> 00:34:48.340 align:middle line:84% both of them taken from this current delivery delay 00:34:48.340 --> 00:34:52.060 align:middle line:90% but with different times. 00:34:52.060 --> 00:34:54.860 align:middle line:84% I think in terms of the actual behavior of the model, 00:34:54.860 --> 00:34:56.060 align:middle line:90% they would be very similar. 00:34:56.060 --> 00:34:57.600 align:middle line:84% It's basically a second order delay 00:34:57.600 --> 00:34:59.280 align:middle line:90% versus a first order delay. 00:34:59.280 --> 00:35:01.620 align:middle line:84% So it's not going to make a big difference, 00:35:01.620 --> 00:35:05.640 align:middle line:84% but it is sometimes conceptually cleaner that way. 00:35:05.640 --> 00:35:09.600 align:middle line:84% Now, yours has a little bit of additional change, 00:35:09.600 --> 00:35:12.960 align:middle line:84% that you're normalizing the delivery delay that 00:35:12.960 --> 00:35:16.640 align:middle line:84% is perceived by the company versus the desired delivery 00:35:16.640 --> 00:35:21.520 align:middle line:90% delay that they may have. 00:35:21.520 --> 00:35:26.280 align:middle line:84% So it's really kind of the ratio of those two that is input or-- 00:35:26.280 --> 00:35:29.336 align:middle line:90% so it's the minimum of the-- 00:35:29.336 --> 00:35:33.805 align:middle line:90% or wait, so delay reputation-- 00:35:33.805 --> 00:35:36.180 align:middle line:84% AUDIENCE: I just didn't want [? them ?] to go below zero. 00:35:36.180 --> 00:35:37.097 align:middle line:90% HAZHIR RAHMANDAD: Yes. 00:35:37.097 --> 00:35:40.600 align:middle line:90% 00:35:40.600 --> 00:35:46.120 align:middle line:84% Customer desired delivery delay minus perceived delivery delay 00:35:46.120 --> 00:35:51.240 align:middle line:84% divided by reputation adjustment time. 00:35:51.240 --> 00:35:56.420 align:middle line:84% So customer desired delivery delay minus perceived delivery 00:35:56.420 --> 00:35:58.040 align:middle line:90% delay-- that becomes-- 00:35:58.040 --> 00:36:00.860 align:middle line:90% 00:36:00.860 --> 00:36:02.973 align:middle line:90% so this is actually not-- 00:36:02.973 --> 00:36:04.640 align:middle line:84% AUDIENCE: Yeah, that's not right, is it? 00:36:04.640 --> 00:36:06.057 align:middle line:84% HAZHIR RAHMANDAD: It is not right. 00:36:06.057 --> 00:36:08.860 align:middle line:84% So let's throw out the minimum for us in our mind. 00:36:08.860 --> 00:36:10.900 align:middle line:84% This part really is not operational for most 00:36:10.900 --> 00:36:12.440 align:middle line:90% of the behavior. 00:36:12.440 --> 00:36:15.380 align:middle line:84% So what remains is the customer desired delivery delay 00:36:15.380 --> 00:36:17.380 align:middle line:90% minus perceived delivery delay. 00:36:17.380 --> 00:36:20.820 align:middle line:84% So it is closing the gap between perceived delivery 00:36:20.820 --> 00:36:23.160 align:middle line:84% delay and customer desired delivery delay. 00:36:23.160 --> 00:36:26.880 align:middle line:90% 00:36:26.880 --> 00:36:27.810 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:36:27.810 --> 00:36:30.060 align:middle line:84% HAZHIR RAHMANDAD: And then it takes it into this delay 00:36:30.060 --> 00:36:31.240 align:middle line:90% reputation stock. 00:36:31.240 --> 00:36:34.180 align:middle line:84% So it is integrating the difference between those two 00:36:34.180 --> 00:36:37.140 align:middle line:84% things with some adjustment time. 00:36:37.140 --> 00:36:39.400 align:middle line:84% So integrating the difference between-- 00:36:39.400 --> 00:36:43.860 align:middle line:84% so if customer desire is always two months and perceived 00:36:43.860 --> 00:36:46.960 align:middle line:90% delivery delay and the actual-- 00:36:46.960 --> 00:36:50.040 align:middle line:84% or sorry, the perceived delivery delay is three months, 00:36:50.040 --> 00:36:52.860 align:middle line:84% it is integrating minus one months over time. 00:36:52.860 --> 00:36:55.763 align:middle line:84% It's, like, adding and adding to that stock. 00:36:55.763 --> 00:36:57.180 align:middle line:84% So that's not what you want to do. 00:36:57.180 --> 00:37:00.000 align:middle line:90% So this is not correct. 00:37:00.000 --> 00:37:03.640 align:middle line:84% So what you need is simply, if you wanted to normalize, 00:37:03.640 --> 00:37:07.120 align:middle line:84% you would say, I'm looking at perceived delivery delay divided 00:37:07.120 --> 00:37:09.040 align:middle line:84% by customer desired delivery delay, 00:37:09.040 --> 00:37:14.520 align:middle line:84% or, minus that divided by something, as your indicated 00:37:14.520 --> 00:37:15.585 align:middle line:90% reputation. 00:37:15.585 --> 00:37:16.960 align:middle line:84% And then you would have indicated 00:37:16.960 --> 00:37:20.160 align:middle line:84% reputation minus delay reputation 00:37:20.160 --> 00:37:22.770 align:middle line:90% to inform the equation. 00:37:22.770 --> 00:37:23.270 align:middle line:90% OK. 00:37:23.270 --> 00:37:26.120 align:middle line:90% So that's delivery delay. 00:37:26.120 --> 00:37:31.160 align:middle line:84% Factories-- this is kind of a one stock version. 00:37:31.160 --> 00:37:36.760 align:middle line:84% That gets a gap based on the desired factory. 00:37:36.760 --> 00:37:40.460 align:middle line:84% And desired factory is coming from-- 00:37:40.460 --> 00:37:40.960 align:middle line:90% OK. 00:37:40.960 --> 00:37:41.475 align:middle line:90% So this is-- 00:37:41.475 --> 00:37:43.100 align:middle line:84% AUDIENCE: Add together work in progress 00:37:43.100 --> 00:37:45.162 align:middle line:84% and the current backlogs, basically. 00:37:45.162 --> 00:37:46.120 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 00:37:46.120 --> 00:37:49.080 align:middle line:84% So it's looking at the current backlog. 00:37:49.080 --> 00:37:56.620 align:middle line:84% That is the work in progress, outstanding order backlog 00:37:56.620 --> 00:37:59.140 align:middle line:90% together. 00:37:59.140 --> 00:38:01.220 align:middle line:90% Why-- oh, yeah. 00:38:01.220 --> 00:38:05.600 align:middle line:84% So it's a little bit tricky here. 00:38:05.600 --> 00:38:09.780 align:middle line:84% So this is going to make it harder for me to right now think 00:38:09.780 --> 00:38:10.280 align:middle line:90% about it. 00:38:10.280 --> 00:38:13.700 align:middle line:84% Because you're using this capacity only 00:38:13.700 --> 00:38:22.140 align:middle line:84% to drain this outstanding order backlog relatively quickly. 00:38:22.140 --> 00:38:26.500 align:middle line:84% But then you are using the work in progress backlog 00:38:26.500 --> 00:38:31.140 align:middle line:84% to expand the desired factory as well. 00:38:31.140 --> 00:38:34.900 align:middle line:84% And that makes things a little bit harder 00:38:34.900 --> 00:38:37.160 align:middle line:84% to make sure it is fully consistent. 00:38:37.160 --> 00:38:40.580 align:middle line:84% So if for whatever reason we increase this time to complete 00:38:40.580 --> 00:38:45.760 align:middle line:84% production, we increase the work in progress backlog 00:38:45.760 --> 00:38:48.880 align:middle line:84% by a factor of 10, now your desired factory 00:38:48.880 --> 00:38:57.280 align:middle line:84% will go up by a factor of 10 even though you are not 00:38:57.280 --> 00:39:01.720 align:middle line:90% changing the purchase rate. 00:39:01.720 --> 00:39:06.960 align:middle line:84% So now, at one level, that is correct 00:39:06.960 --> 00:39:09.700 align:middle line:84% because your time to complete production has increased. 00:39:09.700 --> 00:39:13.520 align:middle line:84% So you now need a lot more work in process capacity 00:39:13.520 --> 00:39:15.020 align:middle line:90% on the floor of the factory. 00:39:15.020 --> 00:39:17.280 align:middle line:84% So it is not wrong in that sense. 00:39:17.280 --> 00:39:19.740 align:middle line:84% But the way we think about factories, 00:39:19.740 --> 00:39:22.560 align:middle line:90% that's more confusing. 00:39:22.560 --> 00:39:25.160 align:middle line:84% So it is, again, kind of an indicator for using 00:39:25.160 --> 00:39:26.780 align:middle line:90% the standard formulation. 00:39:26.780 --> 00:39:29.520 align:middle line:90% 00:39:29.520 --> 00:39:34.340 align:middle line:84% And then here you would be basically missing, OK, 00:39:34.340 --> 00:39:41.360 align:middle line:84% what is the effect of desired delivery delay, so to speak. 00:39:41.360 --> 00:39:44.600 align:middle line:90% Again, that is now embedded. 00:39:44.600 --> 00:39:47.420 align:middle line:84% So there is the effect of delivery delay on the top. 00:39:47.420 --> 00:39:49.660 align:middle line:84% But then there is the effect of desired delivery 00:39:49.660 --> 00:39:55.260 align:middle line:84% delay on what you want the desired factory to be. 00:39:55.260 --> 00:39:57.520 align:middle line:84% That is now embedded inside these stocks 00:39:57.520 --> 00:40:01.220 align:middle line:84% because they change with the time to complete production 00:40:01.220 --> 00:40:03.120 align:middle line:90% and time to process orders. 00:40:03.120 --> 00:40:05.460 align:middle line:84% And the inherent assumption here is then 00:40:05.460 --> 00:40:10.900 align:middle line:84% that the desired delivery delay is the sum of these two delays. 00:40:10.900 --> 00:40:14.680 align:middle line:84% That's all kind of hidden inside that formulation. 00:40:14.680 --> 00:40:19.500 align:middle line:84% It makes it less transparent for somebody 00:40:19.500 --> 00:40:20.640 align:middle line:90% who is reading them all. 00:40:20.640 --> 00:40:22.680 align:middle line:90% But I don't think it is wrong. 00:40:22.680 --> 00:40:26.400 align:middle line:84% I'm just thinking it can be simplified. 00:40:26.400 --> 00:40:26.900 align:middle line:90% OK. 00:40:26.900 --> 00:40:31.380 align:middle line:84% And then you have capacity replacement plus something 00:40:31.380 --> 00:40:32.480 align:middle line:90% about closing the gap. 00:40:32.480 --> 00:40:34.960 align:middle line:84% But let's go into this formulation. 00:40:34.960 --> 00:40:36.440 align:middle line:90% So you can't remove factories. 00:40:36.440 --> 00:40:37.240 align:middle line:90% That's important. 00:40:37.240 --> 00:40:38.460 align:middle line:90% That's good. 00:40:38.460 --> 00:40:42.080 align:middle line:84% Capacity replacement plus factory gap 00:40:42.080 --> 00:40:45.800 align:middle line:84% multiplied by construction sensitivity divided by time 00:40:45.800 --> 00:40:47.020 align:middle line:90% to build new factories. 00:40:47.020 --> 00:40:53.480 align:middle line:84% So construction sensitivity is coming from your delivery delay. 00:40:53.480 --> 00:40:56.840 align:middle line:84% And you have customer desired delivery delay versus perceived 00:40:56.840 --> 00:40:57.620 align:middle line:90% delivery delay. 00:40:57.620 --> 00:41:13.720 align:middle line:84% Now, this is a ratio-- so the ratio of perceived delivery 00:41:13.720 --> 00:41:17.960 align:middle line:84% delay to customer desired delivery delay is 1. 00:41:17.960 --> 00:41:23.240 align:middle line:84% Then this multiplicative system will give you 00:41:23.240 --> 00:41:28.760 align:middle line:84% a quarter, a value of 0.25, which is your base sensitivity. 00:41:28.760 --> 00:41:35.960 align:middle line:84% You're going to then multiply your factory 00:41:35.960 --> 00:41:39.440 align:middle line:90% gap by that sensitivity. 00:41:39.440 --> 00:41:44.380 align:middle line:84% And then you would divide it by time to build new capacity. 00:41:44.380 --> 00:41:46.240 align:middle line:90% So that is fine. 00:41:46.240 --> 00:41:49.180 align:middle line:84% It is closing the gap with a time constant. 00:41:49.180 --> 00:41:51.540 align:middle line:84% But if you notice, basically, you 00:41:51.540 --> 00:41:54.300 align:middle line:84% have two parameters that are double counting 00:41:54.300 --> 00:41:56.200 align:middle line:90% or just kind of redundant. 00:41:56.200 --> 00:41:58.700 align:middle line:84% So the factory construction sensitivity 00:41:58.700 --> 00:42:01.900 align:middle line:84% has a parameter which is multiplicative. 00:42:01.900 --> 00:42:06.660 align:middle line:84% And then you are dividing that by time to build new capacity. 00:42:06.660 --> 00:42:11.560 align:middle line:84% And therefore, they are fully collinear. 00:42:11.560 --> 00:42:13.980 align:middle line:90% You don't need both of them. 00:42:13.980 --> 00:42:19.460 align:middle line:84% So as such, it's basically redundant. 00:42:19.460 --> 00:42:23.660 align:middle line:84% AUDIENCE: How would you get unit consistency? 00:42:23.660 --> 00:42:28.100 align:middle line:84% [INAUDIBLE] sensitivity effect on the time constant and then 00:42:28.100 --> 00:42:29.020 align:middle line:90% [INAUDIBLE] 00:42:29.020 --> 00:42:29.978 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 00:42:29.978 --> 00:42:36.780 align:middle line:84% So I think conceptually, I see the appeal 00:42:36.780 --> 00:42:39.720 align:middle line:84% of having a parameter that changes 00:42:39.720 --> 00:42:41.600 align:middle line:90% the strength of this feedback. 00:42:41.600 --> 00:42:45.200 align:middle line:84% I probably would have used a power function rather than 00:42:45.200 --> 00:42:47.720 align:middle line:90% a linear multiplication. 00:42:47.720 --> 00:42:51.000 align:middle line:84% And then the two would have had different effects. 00:42:51.000 --> 00:42:57.520 align:middle line:84% So it would have been that ratio of perceived versus customer 00:42:57.520 --> 00:42:59.160 align:middle line:90% desired raised to some power. 00:42:59.160 --> 00:43:02.900 align:middle line:84% That would probably do a better job. 00:43:02.900 --> 00:43:03.400 align:middle line:90% OK. 00:43:03.400 --> 00:43:05.580 align:middle line:84% So I think this should work fine. 00:43:05.580 --> 00:43:08.620 align:middle line:84% And then you have this basic delivery here. 00:43:08.620 --> 00:43:09.120 align:middle line:90% OK. 00:43:09.120 --> 00:43:17.160 align:middle line:84% These are pretty meaningful with a few places 00:43:17.160 --> 00:43:19.240 align:middle line:90% that we could do better. 00:43:19.240 --> 00:43:26.400 align:middle line:84% Sales force is converting people who are interested. 00:43:26.400 --> 00:43:32.620 align:middle line:84% And the conversion rate depends on sales force outreach. 00:43:32.620 --> 00:43:37.602 align:middle line:84% This is kind of how many people they can reach out to. 00:43:37.602 --> 00:43:40.060 align:middle line:84% AUDIENCE: The conversion rate doesn't depend on sales force 00:43:40.060 --> 00:43:41.227 align:middle line:90% average, but the sales does. 00:43:41.227 --> 00:43:42.143 align:middle line:90% HAZHIR RAHMANDAD: Yes. 00:43:42.143 --> 00:43:43.020 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:43:43.020 --> 00:43:45.062 align:middle line:84% HAZHIR RAHMANDAD: Yeah, so you have the outreach. 00:43:45.062 --> 00:43:50.980 align:middle line:84% And then conversion rate depends on maximum, a reputation effect. 00:43:50.980 --> 00:43:53.820 align:middle line:90% And what else does it affect? 00:43:53.820 --> 00:43:57.340 align:middle line:84% So if you don't have any total interested market, 00:43:57.340 --> 00:44:04.340 align:middle line:84% how would you shut off the sales path? 00:44:04.340 --> 00:44:05.760 align:middle line:90% So you have it in here. 00:44:05.760 --> 00:44:07.200 align:middle line:84% No, actually, you do have it here. 00:44:07.200 --> 00:44:10.100 align:middle line:90% 00:44:10.100 --> 00:44:11.180 align:middle line:90% OK. 00:44:11.180 --> 00:44:15.740 align:middle line:84% So the question I was asking is-- the first order control 00:44:15.740 --> 00:44:18.860 align:middle line:90% for the sales outflow. 00:44:18.860 --> 00:44:20.380 align:middle line:84% And you want that to be dependent 00:44:20.380 --> 00:44:22.220 align:middle line:90% on total interested market. 00:44:22.220 --> 00:44:23.320 align:middle line:90% It is coming here. 00:44:23.320 --> 00:44:24.880 align:middle line:90% It is just not as visible. 00:44:24.880 --> 00:44:27.540 align:middle line:84% So there is a maximum yearly outreach per sales person 00:44:27.540 --> 00:44:31.660 align:middle line:84% that is feasible versus your yearly outreach that 00:44:31.660 --> 00:44:32.920 align:middle line:90% comes from these people. 00:44:32.920 --> 00:44:37.160 align:middle line:84% And you are probably using some minimum function here. 00:44:37.160 --> 00:44:39.520 align:middle line:84% Yes, minimum of this yearly outreach 00:44:39.520 --> 00:44:42.680 align:middle line:84% and maximum possible per salesperson. 00:44:42.680 --> 00:44:43.780 align:middle line:90% So that's fine. 00:44:43.780 --> 00:44:45.320 align:middle line:90% That would do the job. 00:44:45.320 --> 00:44:48.040 align:middle line:84% Now, again, the assumption that goes into this equation 00:44:48.040 --> 00:44:48.723 align:middle line:90% is important. 00:44:48.723 --> 00:44:50.640 align:middle line:84% So the assumption that goes into this equation 00:44:50.640 --> 00:44:54.880 align:middle line:84% is that, basically, if you don't have 00:44:54.880 --> 00:44:56.720 align:middle line:84% enough people in the market compared 00:44:56.720 --> 00:45:00.520 align:middle line:84% to the capacity of your sales force, 00:45:00.520 --> 00:45:03.560 align:middle line:84% the utilization of the sales force 00:45:03.560 --> 00:45:08.000 align:middle line:84% goes down linearly with how many people are in the market. 00:45:08.000 --> 00:45:10.680 align:middle line:84% So that assumption basically says 00:45:10.680 --> 00:45:16.520 align:middle line:84% sales force people are not that capable of identifying 00:45:16.520 --> 00:45:18.160 align:middle line:90% open leads. 00:45:18.160 --> 00:45:20.180 align:middle line:84% That is, if the open leads go down, 00:45:20.180 --> 00:45:24.000 align:middle line:84% you linearly are bringing down the effectiveness 00:45:24.000 --> 00:45:26.120 align:middle line:90% of your sales force people. 00:45:26.120 --> 00:45:29.600 align:middle line:84% Whereas, if there are just enough people, 00:45:29.600 --> 00:45:36.740 align:middle line:84% the sales force will be fully utilized at their 100% capacity. 00:45:36.740 --> 00:45:40.020 align:middle line:84% So that utilization function is kind of a little bit 00:45:40.020 --> 00:45:41.520 align:middle line:90% too extreme on both sides. 00:45:41.520 --> 00:45:46.380 align:middle line:84% That is, if I have just enough people who 00:45:46.380 --> 00:45:50.900 align:middle line:84% are interested in the product as my sales force capacity, 00:45:50.900 --> 00:45:52.580 align:middle line:84% it is a little bit unrealistic to think 00:45:52.580 --> 00:45:56.740 align:middle line:84% that the sales force capacity will be used at its maximum. 00:45:56.740 --> 00:46:02.940 align:middle line:84% And on the other hand, if I have 80% of that number, 00:46:02.940 --> 00:46:07.400 align:middle line:84% bringing down the utilization by 20% may also be unrealistic. 00:46:07.400 --> 00:46:12.700 align:middle line:84% So some softer adjustment using one of the utilization functions 00:46:12.700 --> 00:46:15.220 align:middle line:84% that we have seen in other exercises as well 00:46:15.220 --> 00:46:18.310 align:middle line:84% would probably be a better idea here, 00:46:18.310 --> 00:46:19.560 align:middle line:90% even though this is not wrong. 00:46:19.560 --> 00:46:20.880 align:middle line:90% This is a good starting point. 00:46:20.880 --> 00:46:23.600 align:middle line:90% So this would do OK. 00:46:23.600 --> 00:46:26.280 align:middle line:84% But there are better formulations there. 00:46:26.280 --> 00:46:28.080 align:middle line:90% But otherwise, this is fine. 00:46:28.080 --> 00:46:31.560 align:middle line:84% And it takes care of the sales the way 00:46:31.560 --> 00:46:35.160 align:middle line:84% you want it, again, with the caveat that this now 00:46:35.160 --> 00:46:41.960 align:middle line:84% assumes every new sales is very much the same 00:46:41.960 --> 00:46:44.240 align:middle line:90% as every repeat sales. 00:46:44.240 --> 00:46:49.800 align:middle line:84% So that time to identify new leads versus go back 00:46:49.800 --> 00:46:52.580 align:middle line:84% to your old customers is usually very different. 00:46:52.580 --> 00:46:55.680 align:middle line:84% And therefore, it does make sense potentially 00:46:55.680 --> 00:46:57.580 align:middle line:90% to separate those two. 00:46:57.580 --> 00:46:59.620 align:middle line:84% But this is a simpler formulation. 00:46:59.620 --> 00:47:01.460 align:middle line:84% And it's fine to have a simpler formulation. 00:47:01.460 --> 00:47:03.860 align:middle line:84% Just be conscious that you are making that choice. 00:47:03.860 --> 00:47:06.780 align:middle line:84% And I think some of the folks have an alternative here. 00:47:06.780 --> 00:47:07.360 align:middle line:90% OK. 00:47:07.360 --> 00:47:11.440 align:middle line:84% I think formulation-wise this is a good summary of what 00:47:11.440 --> 00:47:12.960 align:middle line:90% is happening here. 00:47:12.960 --> 00:47:16.040 align:middle line:90% I assume you have not-- 00:47:16.040 --> 00:47:19.380 align:middle line:84% you have a fixed sales force fraction. 00:47:19.380 --> 00:47:19.880 align:middle line:90% Yes. 00:47:19.880 --> 00:47:22.780 align:middle line:90% So this fraction is fixed. 00:47:22.780 --> 00:47:25.180 align:middle line:84% And then we have the accounting that takes care of it. 00:47:25.180 --> 00:47:25.680 align:middle line:90% OK. 00:47:25.680 --> 00:47:27.740 align:middle line:84% Let me just go to some random other-- 00:47:27.740 --> 00:47:28.620 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:47:28.620 --> 00:47:29.700 align:middle line:90% HAZHIR RAHMANDAD: Yes. 00:47:29.700 --> 00:47:30.440 align:middle line:90% Go ahead. 00:47:30.440 --> 00:47:35.378 align:middle line:84% AUDIENCE: [INAUDIBLE] Just going back to the capacity [INAUDIBLE] 00:47:35.378 --> 00:47:37.420 align:middle line:84% this is kind of what I was asking about early on. 00:47:37.420 --> 00:47:41.620 align:middle line:90% 00:47:41.620 --> 00:47:44.127 align:middle line:84% Yeah, so I had this issue where just 00:47:44.127 --> 00:47:46.460 align:middle line:84% something that I was considering was that factories take 00:47:46.460 --> 00:47:48.220 align:middle line:90% two or three years to build. 00:47:48.220 --> 00:47:52.120 align:middle line:84% And so if a factory is deprecated, 00:47:52.120 --> 00:47:53.903 align:middle line:84% you can't just immediately replace it. 00:47:53.903 --> 00:47:54.820 align:middle line:90% HAZHIR RAHMANDAD: Yes. 00:47:54.820 --> 00:47:55.220 align:middle line:90% AUDIENCE: Right? 00:47:55.220 --> 00:47:55.860 align:middle line:90% So-- 00:47:55.860 --> 00:47:56.320 align:middle line:90% HAZHIR RAHMANDAD: Yes. 00:47:56.320 --> 00:47:57.240 align:middle line:90% So that's right. 00:47:57.240 --> 00:47:59.700 align:middle line:84% So this is definitely a simplification 00:47:59.700 --> 00:48:02.580 align:middle line:84% that, in case of this exercise, I would have-- 00:48:02.580 --> 00:48:04.240 align:middle line:84% I think in the problem definition, 00:48:04.240 --> 00:48:06.140 align:middle line:84% there is a conversation about the time it 00:48:06.140 --> 00:48:07.400 align:middle line:90% takes to build the factories. 00:48:07.400 --> 00:48:10.300 align:middle line:84% So I think a better formulation would include the factory 00:48:10.300 --> 00:48:11.900 align:middle line:90% under development stock. 00:48:11.900 --> 00:48:15.420 align:middle line:84% And in fact, I would also use higher order delays 00:48:15.420 --> 00:48:17.780 align:middle line:90% for depreciation of factories. 00:48:17.780 --> 00:48:22.445 align:middle line:84% First order depreciation is kind of a weird thing for a factory. 00:48:22.445 --> 00:48:23.820 align:middle line:84% The year you built a factory, you 00:48:23.820 --> 00:48:27.300 align:middle line:84% don't start throwing it out of your capacity. 00:48:27.300 --> 00:48:30.440 align:middle line:84% So at least a third order delay is due there. 00:48:30.440 --> 00:48:34.820 align:middle line:84% Even for completion of factories out of the work in process, 00:48:34.820 --> 00:48:38.200 align:middle line:84% that also makes sense to have a higher order delay. 00:48:38.200 --> 00:48:43.620 align:middle line:84% So that would be probably a not cleaner necessarily, 00:48:43.620 --> 00:48:46.420 align:middle line:90% but a more realistic solution. 00:48:46.420 --> 00:48:46.920 align:middle line:90% All right. 00:48:46.920 --> 00:48:50.140 align:middle line:90% So which is this one? 00:48:50.140 --> 00:48:53.360 align:middle line:90% So this is Kathy's. 00:48:53.360 --> 00:48:57.680 align:middle line:84% Let's take-- so manufacturing here, 00:48:57.680 --> 00:49:02.100 align:middle line:84% we had these units that are in the queue versus in production. 00:49:02.100 --> 00:49:07.000 align:middle line:84% We talked about-- it's a little bit similar to Jake's part, 00:49:07.000 --> 00:49:11.340 align:middle line:84% that you probably don't need all of that detail. 00:49:11.340 --> 00:49:14.080 align:middle line:90% But it could work just fine. 00:49:14.080 --> 00:49:15.760 align:middle line:84% So the way you are doing it is also 00:49:15.760 --> 00:49:19.840 align:middle line:84% similar to Jake in the sense that the production start 00:49:19.840 --> 00:49:23.960 align:middle line:90% is gated by the capacity. 00:49:23.960 --> 00:49:27.780 align:middle line:84% And so this is one of those places 00:49:27.780 --> 00:49:30.960 align:middle line:84% that I'm now not sure what is happening. 00:49:30.960 --> 00:49:33.740 align:middle line:84% So the manufacturing capacity, you also 00:49:33.740 --> 00:49:36.960 align:middle line:84% have it in units rather than units per unit of time. 00:49:36.960 --> 00:49:40.800 align:middle line:84% So that's exactly the same issue that we had with Jake's point. 00:49:40.800 --> 00:49:44.460 align:middle line:84% And again, I think that's deviating 00:49:44.460 --> 00:49:47.745 align:middle line:84% from standard formulations that will make some challenges. 00:49:47.745 --> 00:49:49.120 align:middle line:84% Now, it doesn't mean it is wrong. 00:49:49.120 --> 00:49:51.660 align:middle line:84% So let's look at what is happening with the-- so you 00:49:51.660 --> 00:49:54.380 align:middle line:84% are saying the units in production 00:49:54.380 --> 00:50:01.940 align:middle line:84% are stopping or kind of limiting what capacity remains. 00:50:01.940 --> 00:50:04.620 align:middle line:84% Whatever is not in the production opens up 00:50:04.620 --> 00:50:07.240 align:middle line:90% a gate for starting production. 00:50:07.240 --> 00:50:09.940 align:middle line:84% And that gate is controlled by some minimum time 00:50:09.940 --> 00:50:12.100 align:middle line:90% to start production. 00:50:12.100 --> 00:50:15.220 align:middle line:84% And that minimum time to start production also 00:50:15.220 --> 00:50:18.560 align:middle line:84% gates the desired production start. 00:50:18.560 --> 00:50:22.040 align:middle line:84% This is actually very similar to Jake's formulation, I think, 00:50:22.040 --> 00:50:23.480 align:middle line:90% mathematically. 00:50:23.480 --> 00:50:26.160 align:middle line:90% It has the same problems with-- 00:50:26.160 --> 00:50:29.120 align:middle line:90% yeah, I think it will work. 00:50:29.120 --> 00:50:34.760 align:middle line:84% It wouldn't generate any wrong results, but it is-- 00:50:34.760 --> 00:50:40.040 align:middle line:84% AUDIENCE: My question is, what is including this [INAUDIBLE]? 00:50:40.040 --> 00:50:41.920 align:middle line:84% Because I think the problem statement 00:50:41.920 --> 00:50:47.720 align:middle line:84% said, assume just a backlog, no work in progress. 00:50:47.720 --> 00:50:51.280 align:middle line:84% So [INAUDIBLE] just right on backlog. 00:50:51.280 --> 00:50:55.020 align:middle line:84% And there was no capacity in use and non-working capacity. 00:50:55.020 --> 00:50:56.722 align:middle line:84% So I didn't even think to do that. 00:50:56.722 --> 00:50:57.680 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 00:50:57.680 --> 00:50:58.222 align:middle line:90% I don't know. 00:50:58.222 --> 00:50:59.740 align:middle line:90% What were you trying to do? 00:50:59.740 --> 00:51:02.600 align:middle line:84% What were you trying to get out of this formulation, Kathy? 00:51:02.600 --> 00:51:06.240 align:middle line:84% Do you remember when you were doing this? 00:51:06.240 --> 00:51:10.200 align:middle line:84% AUDIENCE: Wow, [INAUDIBLE] yeah, I 00:51:10.200 --> 00:51:16.780 align:middle line:84% think I was just picturing in my mind what is happening. 00:51:16.780 --> 00:51:20.360 align:middle line:84% Units are kind of-- oh, these are the orders we have to fill. 00:51:20.360 --> 00:51:20.860 align:middle line:90% OK. 00:51:20.860 --> 00:51:21.940 align:middle line:90% That's different. 00:51:21.940 --> 00:51:23.800 align:middle line:84% When do we get to start something? 00:51:23.800 --> 00:51:27.140 align:middle line:84% When we have an available slot for it? 00:51:27.140 --> 00:51:29.640 align:middle line:84% What are the slots determined [? by, ?] the capacity? 00:51:29.640 --> 00:51:32.820 align:middle line:84% So I think I was just really like taking it literally 00:51:32.820 --> 00:51:34.760 align:middle line:90% from that image in my mind. 00:51:34.760 --> 00:51:36.940 align:middle line:90% I wasn't, like, abstracting. 00:51:36.940 --> 00:51:39.260 align:middle line:90% And then the capacity thing-- 00:51:39.260 --> 00:51:45.680 align:middle line:84% I think it was just a product of my brain. 00:51:45.680 --> 00:51:51.700 align:middle line:84% But I had it originally in units per month and more so standard. 00:51:51.700 --> 00:51:55.140 align:middle line:84% And then I started thinking about the slots idea 00:51:55.140 --> 00:51:57.120 align:middle line:84% and how that would kind be easy to implement. 00:51:57.120 --> 00:52:00.180 align:middle line:84% And so I changed it all last minute to units and just carry 00:52:00.180 --> 00:52:01.000 align:middle line:90% [INAUDIBLE]. 00:52:01.000 --> 00:52:02.250 align:middle line:90% HAZHIR RAHMANDAD: Interesting. 00:52:02.250 --> 00:52:04.300 align:middle line:84% So that makes it a little bit more clear for me 00:52:04.300 --> 00:52:05.880 align:middle line:90% how you thought through it. 00:52:05.880 --> 00:52:09.880 align:middle line:84% And I think the process lesson out of that is, on the one hand, 00:52:09.880 --> 00:52:12.560 align:middle line:84% being operational-- just you try to be operational. 00:52:12.560 --> 00:52:15.700 align:middle line:84% Being operational is usually a good skill and good thing 00:52:15.700 --> 00:52:18.760 align:middle line:84% to do in the modeling because you have a tangible thing 00:52:18.760 --> 00:52:21.160 align:middle line:90% that you're trying to track. 00:52:21.160 --> 00:52:24.980 align:middle line:84% But then it sometimes gets you too much into the weeds. 00:52:24.980 --> 00:52:28.060 align:middle line:84% And you may get lost in a model that is out of balance. 00:52:28.060 --> 00:52:30.200 align:middle line:84% That is, it has too much detail in some sector 00:52:30.200 --> 00:52:35.000 align:middle line:84% and is kind of missing other feedback, complexity parts that 00:52:35.000 --> 00:52:35.917 align:middle line:90% may be more important. 00:52:35.917 --> 00:52:37.958 align:middle line:84% Now, I don't think you are missing something here 00:52:37.958 --> 00:52:39.260 align:middle line:90% on the other side, necessarily. 00:52:39.260 --> 00:52:43.800 align:middle line:84% But just that creates a trade off. 00:52:43.800 --> 00:52:46.860 align:middle line:84% There is really no easy solution to that. 00:52:46.860 --> 00:52:51.120 align:middle line:84% This is kind of a skill you have to build of, OK, let me go deep, 00:52:51.120 --> 00:52:53.660 align:middle line:84% have an operational story, but then ask myself, 00:52:53.660 --> 00:52:55.600 align:middle line:90% how can I summarize? 00:52:55.600 --> 00:52:57.540 align:middle line:84% Do I need all of that complexity? 00:52:57.540 --> 00:53:00.200 align:middle line:84% And I think the time constants for these 00:53:00.200 --> 00:53:03.360 align:middle line:84% and what it buys you-- do I use, for example, 00:53:03.360 --> 00:53:05.280 align:middle line:84% these units in production differently 00:53:05.280 --> 00:53:06.940 align:middle line:90% from units in production queue? 00:53:06.940 --> 00:53:07.440 align:middle line:90% You're not. 00:53:07.440 --> 00:53:10.560 align:middle line:84% You're just basically adding up both of those 00:53:10.560 --> 00:53:12.480 align:middle line:90% to calculate a delivery delay. 00:53:12.480 --> 00:53:14.640 align:middle line:90% So why do I need-- 00:53:14.640 --> 00:53:18.620 align:middle line:84% AUDIENCE: Well, so I'm using them differently in that one 00:53:18.620 --> 00:53:22.080 align:middle line:84% of them is capacity constrained and one of them isn't. 00:53:22.080 --> 00:53:26.180 align:middle line:84% Like, there's no constraint on me taking orders. 00:53:26.180 --> 00:53:30.180 align:middle line:84% And so I didn't want to meter the process of taking orders. 00:53:30.180 --> 00:53:34.660 align:middle line:84% I wanted to meter the process of starting production. 00:53:34.660 --> 00:53:37.215 align:middle line:84% AUDIENCE: Why are we taking orders [INAUDIBLE]? 00:53:37.215 --> 00:53:38.340 align:middle line:90% AUDIENCE: I can't hear you. 00:53:38.340 --> 00:53:38.820 align:middle line:90% What are you-- 00:53:38.820 --> 00:53:40.237 align:middle line:84% HAZHIR RAHMANDAD: So taking orders 00:53:40.237 --> 00:53:41.680 align:middle line:84% is units entering queue, correct? 00:53:41.680 --> 00:53:42.555 align:middle line:90% AUDIENCE: Yeah, yeah. 00:53:42.555 --> 00:53:45.300 align:middle line:84% And I didn't want that to be capacity constrained because-- 00:53:45.300 --> 00:53:46.967 align:middle line:84% HAZHIR RAHMANDAD: So that's not capacity 00:53:46.967 --> 00:53:48.420 align:middle line:90% constrained in a single stock. 00:53:48.420 --> 00:53:51.020 align:middle line:84% So think about a situation where, 00:53:51.020 --> 00:53:52.600 align:middle line:84% rather than having these two stocks, 00:53:52.600 --> 00:53:54.220 align:middle line:90% you just didn't have this flow. 00:53:54.220 --> 00:53:56.380 align:middle line:90% This was the one single stock. 00:53:56.380 --> 00:54:00.040 align:middle line:84% And the outflow here was constrained by the capacity, 00:54:00.040 --> 00:54:01.040 align:middle line:90% your total capacity. 00:54:01.040 --> 00:54:05.440 align:middle line:84% That's the solution that would sidestep these problems. 00:54:05.440 --> 00:54:05.940 align:middle line:90% OK. 00:54:05.940 --> 00:54:08.380 align:middle line:84% Your average delivery delay is now 00:54:08.380 --> 00:54:10.860 align:middle line:84% calculated as units in production plus units 00:54:10.860 --> 00:54:17.400 align:middle line:84% in production queue divided by unit delivery. 00:54:17.400 --> 00:54:21.360 align:middle line:84% And then, what is the customer delay? 00:54:21.360 --> 00:54:22.800 align:middle line:90% Oh, that's just the [INAUDIBLE]. 00:54:22.800 --> 00:54:23.780 align:middle line:90% So that's the x. 00:54:23.780 --> 00:54:24.280 align:middle line:90% OK. 00:54:24.280 --> 00:54:26.280 align:middle line:90% That's fine. 00:54:26.280 --> 00:54:30.100 align:middle line:84% If it goes 0, it becomes expected [? delivery. ?] 00:54:30.100 --> 00:54:31.600 align:middle line:84% AUDIENCE: Yeah, I didn't really know 00:54:31.600 --> 00:54:35.460 align:middle line:84% what was the best thing to make the x in that case, but 00:54:35.460 --> 00:54:35.960 align:middle line:90% [INAUDIBLE]. 00:54:35.960 --> 00:54:37.320 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 00:54:37.320 --> 00:54:41.560 align:middle line:84% So if your delivery rate goes to 0, 00:54:41.560 --> 00:54:46.200 align:middle line:84% basically the average delivery delay is a very large number. 00:54:46.200 --> 00:54:49.140 align:middle line:84% That basically says things are not good. 00:54:49.140 --> 00:54:51.120 align:middle line:84% So if you use expected delivery delay, 00:54:51.120 --> 00:54:52.260 align:middle line:90% that's not the best choice. 00:54:52.260 --> 00:54:54.800 align:middle line:84% It is actually saying, this is the level 00:54:54.800 --> 00:54:56.460 align:middle line:90% that people will be happy with. 00:54:56.460 --> 00:54:59.960 align:middle line:84% So if ever this model touches the situation when 00:54:59.960 --> 00:55:02.480 align:middle line:84% there is no delivery, it will create 00:55:02.480 --> 00:55:04.980 align:middle line:84% delivery delays that are very attractive to customers. 00:55:04.980 --> 00:55:07.480 align:middle line:84% And as such, that's not the right choice. 00:55:07.480 --> 00:55:11.220 align:middle line:84% Now, in your simulation, you may never see that. 00:55:11.220 --> 00:55:14.660 align:middle line:84% So you may not understand or see it 00:55:14.660 --> 00:55:16.840 align:middle line:90% in terms of the actual behavior. 00:55:16.840 --> 00:55:18.600 align:middle line:84% But when you are putting that equation, 00:55:18.600 --> 00:55:21.820 align:middle line:84% this is the thinking that you have to do? 00:55:21.820 --> 00:55:23.700 align:middle line:84% AUDIENCE: But what if there are no units 00:55:23.700 --> 00:55:25.365 align:middle line:90% in production [INAUDIBLE]? 00:55:25.365 --> 00:55:26.740 align:middle line:84% AUDIENCE: Yeah, so I was thinking 00:55:26.740 --> 00:55:28.360 align:middle line:84% if there are no units in production, 00:55:28.360 --> 00:55:32.360 align:middle line:84% it's because there are no orders. 00:55:32.360 --> 00:55:33.440 align:middle line:90% There's nothing going on. 00:55:33.440 --> 00:55:38.100 align:middle line:84% Therefore, there's no reason for the perceived delay 00:55:38.100 --> 00:55:39.953 align:middle line:84% to really deviate from the expectation. 00:55:39.953 --> 00:55:41.620 align:middle line:84% HAZHIR RAHMANDAD: So basically, then you 00:55:41.620 --> 00:55:47.515 align:middle line:84% have a 0 numerator and a 0 denominator. 00:55:47.515 --> 00:55:48.140 align:middle line:90% AUDIENCE: Yeah. 00:55:48.140 --> 00:55:50.820 align:middle line:84% HAZHIR RAHMANDAD: So that is a different situation. 00:55:50.820 --> 00:55:55.000 align:middle line:84% That happens rarely in this model because as long as-- 00:55:55.000 --> 00:55:56.500 align:middle line:84% AUDIENCE: It's just that the first-- 00:55:56.500 --> 00:55:58.417 align:middle line:84% HAZHIR RAHMANDAD: Getting these stocks to be 0 00:55:58.417 --> 00:56:00.700 align:middle line:84% is very hard because it requires somehow 00:56:00.700 --> 00:56:04.520 align:middle line:84% emptying anything that was in the stocks in the first place. 00:56:04.520 --> 00:56:07.380 align:middle line:84% Whereas, getting the flow to be 0-- that's basically, 00:56:07.380 --> 00:56:11.180 align:middle line:84% if you shut off production, it goes to 0-- now, in your model, 00:56:11.180 --> 00:56:11.680 align:middle line:90% not really. 00:56:11.680 --> 00:56:13.640 align:middle line:84% Because this also depends on the stock. 00:56:13.640 --> 00:56:17.160 align:middle line:84% So this part of the delivery delay 00:56:17.160 --> 00:56:21.800 align:middle line:84% is always average completion time unless units in production 00:56:21.800 --> 00:56:22.900 align:middle line:90% goes to 0. 00:56:22.900 --> 00:56:25.520 align:middle line:84% But the way units in production goes to 0 00:56:25.520 --> 00:56:29.800 align:middle line:84% is when you shut off this flow through, for example, not having 00:56:29.800 --> 00:56:31.720 align:middle line:90% capacity or something. 00:56:31.720 --> 00:56:33.120 align:middle line:90% Anyway, this is OK. 00:56:33.120 --> 00:56:35.200 align:middle line:90% There's nothing bad here. 00:56:35.200 --> 00:56:37.320 align:middle line:84% And perceived delivery delay-- fine, 00:56:37.320 --> 00:56:40.160 align:middle line:84% you are using some smooth function, I assume. 00:56:40.160 --> 00:56:45.180 align:middle line:84% And smooth 3 might be an overkill-- just a simple smooth. 00:56:45.180 --> 00:56:50.280 align:middle line:84% So a smooth 3 is information delay third order. 00:56:50.280 --> 00:56:55.080 align:middle line:84% You usually don't use it that much because it requires 00:56:55.080 --> 00:56:59.080 align:middle line:84% a measurement system that has multiple steps 00:56:59.080 --> 00:57:01.200 align:middle line:84% and then kind of maybe cobbled together 00:57:01.200 --> 00:57:04.440 align:middle line:84% with some additional perception of stakeholders. 00:57:04.440 --> 00:57:07.080 align:middle line:84% It is only in those situations where a higher order 00:57:07.080 --> 00:57:08.900 align:middle line:90% smooth function is justified. 00:57:08.900 --> 00:57:10.980 align:middle line:84% Typical perception processes are very much 00:57:10.980 --> 00:57:15.260 align:middle line:84% first order delays, first order smoothness-- but not a big deal. 00:57:15.260 --> 00:57:20.180 align:middle line:84% I would put a box around things that have a smooth or delays so 00:57:20.180 --> 00:57:22.160 align:middle line:84% that, when I look at the diagram, 00:57:22.160 --> 00:57:24.140 align:middle line:84% I quickly see what are the hidden 00:57:24.140 --> 00:57:29.180 align:middle line:90% stocks in the formulation. 00:57:29.180 --> 00:57:34.300 align:middle line:84% And then you have the delivery based reputation 00:57:34.300 --> 00:57:38.180 align:middle line:84% as just the ratio of these two things. 00:57:38.180 --> 00:57:44.300 align:middle line:84% And you are saying it doesn't go above 1. 00:57:44.300 --> 00:57:47.540 align:middle line:90% So that is, if-- 00:57:47.540 --> 00:57:50.690 align:middle line:90% it doesn't go above 1? 00:57:50.690 --> 00:57:53.020 align:middle line:90% Yes, doesn't go above 1. 00:57:53.020 --> 00:57:57.740 align:middle line:84% So if this ratio is customer delivery delay expectation is 00:57:57.740 --> 00:58:00.620 align:middle line:84% much larger than perceived delivery delay, 00:58:00.620 --> 00:58:03.060 align:middle line:84% you want to filter out those situations 00:58:03.060 --> 00:58:05.840 align:middle line:84% and say, delivery delay based reputation is OK. 00:58:05.840 --> 00:58:07.200 align:middle line:90% It's 1. 00:58:07.200 --> 00:58:08.160 align:middle line:90% That's OK. 00:58:08.160 --> 00:58:09.520 align:middle line:90% Nothing wrong there. 00:58:09.520 --> 00:58:11.600 align:middle line:84% But when I see that formulation, I also 00:58:11.600 --> 00:58:14.440 align:middle line:84% have in the back of my mind that you're basically 00:58:14.440 --> 00:58:19.680 align:middle line:84% removing information about situations where delivery 00:58:19.680 --> 00:58:22.200 align:middle line:84% delay is even better than what the customers want 00:58:22.200 --> 00:58:24.660 align:middle line:84% for downstream usage of this variable. 00:58:24.660 --> 00:58:27.000 align:middle line:90% And do I want to do that? 00:58:27.000 --> 00:58:28.560 align:middle line:90% There is always a question. 00:58:28.560 --> 00:58:33.840 align:middle line:84% Now, I may want to do it because it simplifies my life 00:58:33.840 --> 00:58:35.640 align:middle line:90% to worry about the range. 00:58:35.640 --> 00:58:38.420 align:middle line:84% I know now that this is always between 0 and 1. 00:58:38.420 --> 00:58:41.220 align:middle line:84% So in all other places, I just use it with that assumption. 00:58:41.220 --> 00:58:42.800 align:middle line:90% I don't think about it. 00:58:42.800 --> 00:58:46.000 align:middle line:84% But on the other hand, I'm stopping myself 00:58:46.000 --> 00:58:48.880 align:middle line:84% from having more nuanced formulations that 00:58:48.880 --> 00:58:53.360 align:middle line:84% may benefit from the fact that, OK, I now have a situation where 00:58:53.360 --> 00:58:55.980 align:middle line:84% I'm doing really well compared to what the customer is-- 00:58:55.980 --> 00:58:59.440 align:middle line:84% but a little bit of a trade off there. 00:58:59.440 --> 00:58:59.980 align:middle line:90% OK. 00:58:59.980 --> 00:59:05.060 align:middle line:84% So manufacturing capacity fairly similar to Jake's, I 00:59:05.060 --> 00:59:09.220 align:middle line:84% think, in terms of having a single stock. 00:59:09.220 --> 00:59:13.680 align:middle line:84% Lifetime-- again, first order delay not the best formulation. 00:59:13.680 --> 00:59:15.420 align:middle line:90% Missing work in process-- 00:59:15.420 --> 00:59:18.700 align:middle line:84% not the best formulation, but OK. 00:59:18.700 --> 00:59:21.960 align:middle line:84% And then indicated manufacturing capacity-- 00:59:21.960 --> 00:59:25.180 align:middle line:84% you are getting it from this units in production queue. 00:59:25.180 --> 00:59:27.420 align:middle line:84% So this is one of these places that things 00:59:27.420 --> 00:59:29.420 align:middle line:84% start to get messy because you're 00:59:29.420 --> 00:59:31.280 align:middle line:90% ignoring units in production. 00:59:31.280 --> 00:59:33.720 align:middle line:84% You're just getting it from units in production queue. 00:59:33.720 --> 00:59:36.260 align:middle line:90% 00:59:36.260 --> 00:59:39.360 align:middle line:84% And then you're just generating the gap between the two. 00:59:39.360 --> 00:59:41.060 align:middle line:84% So you are not getting the feedback 00:59:41.060 --> 00:59:45.620 align:middle line:84% from the delivery delay that is perceived by the firm. 00:59:45.620 --> 00:59:51.620 align:middle line:84% So that's an important missing part of this whole thing. 00:59:51.620 --> 00:59:53.940 align:middle line:84% Now, it is not that it is completely missing. 00:59:53.940 --> 00:59:55.960 align:middle line:84% Because when your delivery delay is high, 00:59:55.960 --> 00:59:57.980 align:middle line:84% these units in production queue go up. 00:59:57.980 --> 01:00:02.100 align:middle line:84% So as such, it is implicitly there. 01:00:02.100 --> 01:00:05.380 align:middle line:84% But this is just not the cleanest formulation. 01:00:05.380 --> 01:00:10.320 align:middle line:84% It is also missing the replacement pathway. 01:00:10.320 --> 01:00:12.940 align:middle line:84% So this will create a steady state error. 01:00:12.940 --> 01:00:14.523 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:00:14.523 --> 01:00:16.440 align:middle line:84% HAZHIR RAHMANDAD: So if manufacturing capacity 01:00:16.440 --> 01:00:20.280 align:middle line:84% decommissioning is missing from the replacement 01:00:20.280 --> 01:00:24.605 align:middle line:84% or from the new builds, in the steady state, 01:00:24.605 --> 01:00:26.480 align:middle line:84% you always have an error-- error in the sense 01:00:26.480 --> 01:00:30.340 align:middle line:84% that you are not getting what you have desired. 01:00:30.340 --> 01:00:34.440 align:middle line:84% So manufacturing capacity is always 01:00:34.440 --> 01:00:37.340 align:middle line:84% above your actual manufacturing capacity. 01:00:37.340 --> 01:00:41.520 align:middle line:84% So that's bad stock management in general. 01:00:41.520 --> 01:00:43.028 align:middle line:90% Is that part clear? 01:00:43.028 --> 01:00:44.320 align:middle line:90% Anybody has a question on that? 01:00:44.320 --> 01:00:47.940 align:middle line:84% Because that is not a trivial, obvious thing. 01:00:47.940 --> 01:00:51.740 align:middle line:84% But if you have built stock management structures, 01:00:51.740 --> 01:00:53.500 align:middle line:90% you have come across that thing. 01:00:53.500 --> 01:00:55.560 align:middle line:84% That is, for the stock management 01:00:55.560 --> 01:00:59.360 align:middle line:84% to be internally consistent and give the desired values, 01:00:59.360 --> 01:01:02.720 align:middle line:84% you always need two different components. 01:01:02.720 --> 01:01:07.260 align:middle line:84% One is replacement of anything that is going out of the stock. 01:01:07.260 --> 01:01:10.420 align:middle line:84% The other is closing the gap between the desired 01:01:10.420 --> 01:01:12.620 align:middle line:90% level and the current level. 01:01:12.620 --> 01:01:14.700 align:middle line:84% If you only have the closing the gap 01:01:14.700 --> 01:01:17.940 align:middle line:84% between desired and current level, 01:01:17.940 --> 01:01:21.300 align:middle line:84% you will have what we say is a steady state error. 01:01:21.300 --> 01:01:28.200 align:middle line:84% It means in the steady state, the outflow is not replaced. 01:01:28.200 --> 01:01:32.700 align:middle line:84% And therefore, there should be a gap between the desired 01:01:32.700 --> 01:01:38.580 align:middle line:84% and actual to generate enough inflow to equate the outflow. 01:01:38.580 --> 01:01:42.220 align:middle line:84% So therefore, by definition, your desired level 01:01:42.220 --> 01:01:45.860 align:middle line:84% will always be above your current stock 01:01:45.860 --> 01:01:46.880 align:middle line:90% in the equilibrium. 01:01:46.880 --> 01:01:48.640 align:middle line:90% And that is just not right. 01:01:48.640 --> 01:01:50.920 align:middle line:90% So that's important. 01:01:50.920 --> 01:01:51.500 align:middle line:90% Hassan? 01:01:51.500 --> 01:01:54.240 align:middle line:84% AUDIENCE: In the text of the question, 01:01:54.240 --> 01:01:58.980 align:middle line:84% you first suggested that we look more into the pressure that 01:01:58.980 --> 01:02:02.420 align:middle line:84% these [INAUDIBLE] disclaimer, I didn't. 01:02:02.420 --> 01:02:04.320 align:middle line:84% I went for the [INAUDIBLE] and I knew 01:02:04.320 --> 01:02:09.280 align:middle line:84% that I didn't [INAUDIBLE] but that doesn't 01:02:09.280 --> 01:02:11.400 align:middle line:84% apply to when we're formulating based 01:02:11.400 --> 01:02:12.692 align:middle line:90% on the pressure in the company. 01:02:12.692 --> 01:02:13.650 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:02:13.650 --> 01:02:14.760 align:middle line:90% So the pressure based-- 01:02:14.760 --> 01:02:17.840 align:middle line:84% kind of the intuition there is that the firm 01:02:17.840 --> 01:02:25.220 align:middle line:84% is creating an anchor based on their current level of capacity 01:02:25.220 --> 01:02:28.480 align:middle line:84% and then decide whether to expand or contract. 01:02:28.480 --> 01:02:32.080 align:middle line:84% And if it is expanding, it is coming from the pressure 01:02:32.080 --> 01:02:32.860 align:middle line:90% to expand. 01:02:32.860 --> 01:02:36.200 align:middle line:84% So that anchor and adjustment is a different formulation 01:02:36.200 --> 01:02:40.760 align:middle line:84% for stock management compared to replacing what is coming in 01:02:40.760 --> 01:02:43.583 align:middle line:90% and adding the-- 01:02:43.583 --> 01:02:45.000 align:middle line:84% AUDIENCE: The pressure formulation 01:02:45.000 --> 01:02:50.040 align:middle line:84% based on the intensity of change that the company actually 01:02:50.040 --> 01:02:52.880 align:middle line:90% does based on that pressure. 01:02:52.880 --> 01:02:55.157 align:middle line:84% There is a chance that they will have-- 01:02:55.157 --> 01:02:56.740 align:middle line:84% HAZHIR RAHMANDAD: They will overshoot. 01:02:56.740 --> 01:02:58.980 align:middle line:84% AUDIENCE: Yeah, or they can't keep up with [INAUDIBLE]. 01:02:58.980 --> 01:03:01.147 align:middle line:84% HAZHIR RAHMANDAD: Or they can't keep up, absolutely. 01:03:01.147 --> 01:03:04.900 align:middle line:84% So that pressure-based formulation alternative-- first 01:03:04.900 --> 01:03:10.980 align:middle line:84% of all, even in that one, you do need the replacement piece. 01:03:10.980 --> 01:03:12.980 align:middle line:84% If you don't have the replacement piece, 01:03:12.980 --> 01:03:15.460 align:middle line:84% it will create a steady state problem. 01:03:15.460 --> 01:03:17.660 align:middle line:84% Because there's always kind of an outflow. 01:03:17.660 --> 01:03:20.120 align:middle line:84% And the inflow is just always trying to catch up. 01:03:20.120 --> 01:03:23.020 align:middle line:84% So there will be a gap between the desired, which, 01:03:23.020 --> 01:03:26.400 align:middle line:84% the desired can come from your current level with the pressure. 01:03:26.400 --> 01:03:27.800 align:middle line:90% But still, there will be a gap. 01:03:27.800 --> 01:03:32.780 align:middle line:84% So as such, that is also having the same problem. 01:03:32.780 --> 01:03:35.200 align:middle line:84% It's just that, where does the desire come from? 01:03:35.200 --> 01:03:37.900 align:middle line:84% Does it come from the customer orders or something 01:03:37.900 --> 01:03:38.880 align:middle line:90% along those lines? 01:03:38.880 --> 01:03:44.140 align:middle line:84% Or does it come from the delivery delay perceived 01:03:44.140 --> 01:03:49.500 align:middle line:84% and the anchor being current capacity? 01:03:49.500 --> 01:03:51.600 align:middle line:84% I think both of them are pretty legitimate. 01:03:51.600 --> 01:03:55.960 align:middle line:84% And in fact, the behaviors may not be that different. 01:03:55.960 --> 01:03:59.240 align:middle line:84% In fact, anchoring to what is the current backlog 01:03:59.240 --> 01:04:03.440 align:middle line:84% or current recent customer order rates 01:04:03.440 --> 01:04:07.760 align:middle line:90% is less prone to big swings. 01:04:07.760 --> 01:04:10.120 align:middle line:84% Because if you have a very strong or very weak 01:04:10.120 --> 01:04:15.040 align:middle line:84% response to that pressure, then your desired levels 01:04:15.040 --> 01:04:19.000 align:middle line:84% will not be matching or will be overshooting 01:04:19.000 --> 01:04:20.420 align:middle line:90% what the customers are asking. 01:04:20.420 --> 01:04:22.420 align:middle line:90% And that creates instability. 01:04:22.420 --> 01:04:24.840 align:middle line:84% But I asked for that formulation exactly 01:04:24.840 --> 01:04:26.380 align:middle line:90% because of that challenge. 01:04:26.380 --> 01:04:30.400 align:middle line:84% That is how Forrester formulated it. 01:04:30.400 --> 01:04:33.160 align:middle line:84% And that is one of the reasons you 01:04:33.160 --> 01:04:37.440 align:middle line:84% get kind of very unstable behaviors in the market growth 01:04:37.440 --> 01:04:37.940 align:middle line:90% models. 01:04:37.940 --> 01:04:41.460 align:middle line:84% If you went with the more standard stock management, 01:04:41.460 --> 01:04:43.260 align:middle line:84% it is actually a little bit more stable. 01:04:43.260 --> 01:04:48.105 align:middle line:84% Now, I haven't done the full comparison there. 01:04:48.105 --> 01:04:49.480 align:middle line:84% AUDIENCE: Could you explain again 01:04:49.480 --> 01:04:52.640 align:middle line:84% why it's more unstable if you do it with delivery [INAUDIBLE]? 01:04:52.640 --> 01:04:55.980 align:middle line:84% HAZHIR RAHMANDAD: So when you are doing it with the-- 01:04:55.980 --> 01:04:58.400 align:middle line:84% there is an impact of delivery delay on expansion. 01:04:58.400 --> 01:05:03.860 align:middle line:84% So your desired factory level would be your current factory 01:05:03.860 --> 01:05:09.140 align:middle line:84% level multiplied by some effective delivery on capacity 01:05:09.140 --> 01:05:10.220 align:middle line:90% expansion. 01:05:10.220 --> 01:05:14.580 align:middle line:84% That effective delivery on capacity expansion 01:05:14.580 --> 01:05:17.320 align:middle line:84% is dependent on some parameter or some table function. 01:05:17.320 --> 01:05:18.520 align:middle line:90% But let's say a parameter. 01:05:18.520 --> 01:05:23.100 align:middle line:84% And it can be sharper response or slower response. 01:05:23.100 --> 01:05:27.860 align:middle line:84% And that sharpness is not controlled in any direct way 01:05:27.860 --> 01:05:31.060 align:middle line:84% by how big is your delivery delay 01:05:31.060 --> 01:05:35.100 align:middle line:84% or how big is your customer demand currently 01:05:35.100 --> 01:05:36.600 align:middle line:90% compared to your capacity. 01:05:36.600 --> 01:05:40.260 align:middle line:84% So ultimately, if your customer demand is in widgets per year 01:05:40.260 --> 01:05:44.460 align:middle line:84% or widgets per month, and your capacity is also 01:05:44.460 --> 01:05:46.340 align:middle line:84% in widgets per month, you want the two 01:05:46.340 --> 01:05:50.300 align:middle line:84% to be matching with the potential extrapolation 01:05:50.300 --> 01:05:51.440 align:middle line:90% that demand is growing. 01:05:51.440 --> 01:05:54.920 align:middle line:84% So let me look forward a few months based 01:05:54.920 --> 01:05:58.720 align:middle line:84% on how long it takes to create capacity so that I'm matching 01:05:58.720 --> 01:06:02.000 align:middle line:84% the demand I expect to have six months or two years 01:06:02.000 --> 01:06:06.800 align:middle line:84% from now, whatever is your work in process delay. 01:06:06.800 --> 01:06:14.440 align:middle line:84% That is kind of the more clean, logical approach, 01:06:14.440 --> 01:06:18.920 align:middle line:84% in the sense that you have a anchor that directly takes you 01:06:18.920 --> 01:06:21.880 align:middle line:84% to serving the customers exactly at the rate 01:06:21.880 --> 01:06:24.160 align:middle line:90% that they are going to come in. 01:06:24.160 --> 01:06:26.060 align:middle line:84% You can add a buffer or whatever you want. 01:06:26.060 --> 01:06:30.540 align:middle line:84% But the alternative doesn't provide that direct mapping. 01:06:30.540 --> 01:06:35.360 align:middle line:84% And therefore, you can be too fast in expansion or too slow 01:06:35.360 --> 01:06:37.188 align:middle line:90% in expansion. 01:06:37.188 --> 01:06:37.980 align:middle line:90% AUDIENCE: Oh, wait. 01:06:37.980 --> 01:06:38.480 align:middle line:90% Hold on. 01:06:38.480 --> 01:06:41.160 align:middle line:84% I thought you were saying that the delay formulation's 01:06:41.160 --> 01:06:45.040 align:middle line:84% anchoring is what may lead to more instability. 01:06:45.040 --> 01:06:47.820 align:middle line:84% Whereas, the alternative that you're saying is, 01:06:47.820 --> 01:06:49.200 align:middle line:90% this is more stable, yeah. 01:06:49.200 --> 01:06:51.180 align:middle line:84% HAZHIR RAHMANDAD: The one that is more stable-- 01:06:51.180 --> 01:06:55.020 align:middle line:84% or not necessarily stable, but cleaner in the sense 01:06:55.020 --> 01:06:56.780 align:middle line:90% that it is more rational-- 01:06:56.780 --> 01:07:00.740 align:middle line:84% is the one that gets the desired production 01:07:00.740 --> 01:07:07.100 align:middle line:84% rate from the expected customer order rate 01:07:07.100 --> 01:07:12.840 align:middle line:84% and then from that calculates the desired production capacity. 01:07:12.840 --> 01:07:14.715 align:middle line:84% AUDIENCE: Which is what Kathy's doing, right? 01:07:14.715 --> 01:07:16.257 align:middle line:84% HAZHIR RAHMANDAD: Which is what Kathy 01:07:16.257 --> 01:07:18.380 align:middle line:84% is doing, with a caveat that Kathy is not 01:07:18.380 --> 01:07:22.060 align:middle line:84% taking it from the units entering 01:07:22.060 --> 01:07:24.840 align:middle line:90% queue, which is the demand. 01:07:24.840 --> 01:07:28.740 align:middle line:84% She's getting it from this stock here and then turning it 01:07:28.740 --> 01:07:30.160 align:middle line:90% into a stock thing. 01:07:30.160 --> 01:07:31.900 align:middle line:84% Because this manufacturing delivery 01:07:31.900 --> 01:07:35.840 align:middle line:84% is also not in the units of widgets per month. 01:07:35.840 --> 01:07:38.300 align:middle line:84% But it is in the units of units or widgets. 01:07:38.300 --> 01:07:42.780 align:middle line:84% So that creates a little bit additional messiness here 01:07:42.780 --> 01:07:47.320 align:middle line:84% that I can't mentally simulate fully. 01:07:47.320 --> 01:07:49.460 align:middle line:84% AUDIENCE: So two more technical questions. 01:07:49.460 --> 01:07:53.175 align:middle line:84% So first is that, because this firm is producing to a backlog, 01:07:53.175 --> 01:07:54.800 align:middle line:84% they have some desired shipments, which 01:07:54.800 --> 01:07:57.760 align:middle line:84% is the backlog divided by some desired shipment time-- again, 01:07:57.760 --> 01:07:59.720 align:middle line:90% with a single stock [INAUDIBLE]. 01:07:59.720 --> 01:08:01.640 align:middle line:84% So what is the difference between using 01:08:01.640 --> 01:08:04.123 align:middle line:90% that versus the order rate? 01:08:04.123 --> 01:08:06.040 align:middle line:84% HAZHIR RAHMANDAD: It should be about the same. 01:08:06.040 --> 01:08:10.700 align:middle line:84% So basically, you do the backlog divided by delivery delay. 01:08:10.700 --> 01:08:14.540 align:middle line:84% You get the average level of the orders coming in. 01:08:14.540 --> 01:08:17.300 align:middle line:84% So as such, it is a little bit more stable a signal. 01:08:17.300 --> 01:08:19.100 align:middle line:84% So if you have a lot of ups and downs, 01:08:19.100 --> 01:08:21.040 align:middle line:90% that's a better formulation. 01:08:21.040 --> 01:08:23.800 align:middle line:84% For a deterministic model that doesn't kind of 01:08:23.800 --> 01:08:26.817 align:middle line:84% have noisy things, then it doesn't make a difference. 01:08:26.817 --> 01:08:28.359 align:middle line:84% AUDIENCE: And then my second question 01:08:28.359 --> 01:08:30.680 align:middle line:84% was that Jake kind of did both, actually. 01:08:30.680 --> 01:08:34.439 align:middle line:84% Like, he determined the indicated capacity level 01:08:34.439 --> 01:08:37.180 align:middle line:84% by the backlog level or something. 01:08:37.180 --> 01:08:40.040 align:middle line:84% But then the time was adjusted by this pressure 01:08:40.040 --> 01:08:41.800 align:middle line:90% due to delay essentially. 01:08:41.800 --> 01:08:43.462 align:middle line:90% Because he had this [INAUDIBLE]. 01:08:43.462 --> 01:08:44.379 align:middle line:90% HAZHIR RAHMANDAD: Yes. 01:08:44.379 --> 01:08:47.939 align:middle line:84% So basically, he was thinking of closing the gap 01:08:47.939 --> 01:08:50.939 align:middle line:84% to be more sensitive or less sensitive 01:08:50.939 --> 01:08:54.380 align:middle line:84% based on the delivery delay that is observed. 01:08:54.380 --> 01:08:57.740 align:middle line:84% And that is actually a nice formulation in the sense 01:08:57.740 --> 01:09:03.460 align:middle line:84% that, you are closing the gap no matter what, but how fast you're 01:09:03.460 --> 01:09:09.140 align:middle line:84% closing the gap is partly driven by how bad the delivery delay 01:09:09.140 --> 01:09:10.279 align:middle line:90% situation looks like. 01:09:10.279 --> 01:09:13.220 align:middle line:84% So I think I like that, as it kind of conceptually 01:09:13.220 --> 01:09:17.700 align:middle line:84% captures a bunch of these problems we are trying to solve. 01:09:17.700 --> 01:09:18.200 align:middle line:90% OK. 01:09:18.200 --> 01:09:21.580 align:middle line:84% I think this is good conversation in this model. 01:09:21.580 --> 01:09:25.819 align:middle line:84% Just a quick thing-- so I assume you are using a fixed sales 01:09:25.819 --> 01:09:28.380 align:middle line:90% force-- 01:09:28.380 --> 01:09:30.316 align:middle line:84% not fixed, fractional sales force. 01:09:30.316 --> 01:09:30.899 align:middle line:90% AUDIENCE: Yes. 01:09:30.899 --> 01:09:31.774 align:middle line:90% HAZHIR RAHMANDAD: OK. 01:09:31.774 --> 01:09:33.460 align:middle line:90% So that's fine. 01:09:33.460 --> 01:09:35.859 align:middle line:90% That's fraction of revenue. 01:09:35.859 --> 01:09:39.960 align:middle line:84% I'm not going into through the rookie dynamics. 01:09:39.960 --> 01:09:42.340 align:middle line:84% I assume they are coming from the standard formulations 01:09:42.340 --> 01:09:43.160 align:middle line:90% in Stearman. 01:09:43.160 --> 01:09:45.242 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:09:45.242 --> 01:09:46.200 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:09:46.200 --> 01:09:47.479 align:middle line:90% Good. 01:09:47.479 --> 01:09:51.000 align:middle line:90% Let's do a couple more models. 01:09:51.000 --> 01:09:51.500 align:middle line:90% Let's see. 01:09:51.500 --> 01:09:52.680 align:middle line:90% Who's is this one? 01:09:52.680 --> 01:09:54.400 align:middle line:90% Ann. 01:09:54.400 --> 01:09:55.320 align:middle line:90% OK. 01:09:55.320 --> 01:10:01.040 align:middle line:84% So again, we have the facilities, first order 01:10:01.040 --> 01:10:02.120 align:middle line:90% draining-- 01:10:02.120 --> 01:10:05.280 align:middle line:84% not the best alternative, but OK-- 01:10:05.280 --> 01:10:12.500 align:middle line:84% new construction, coming from gap from threshold. 01:10:12.500 --> 01:10:16.280 align:middle line:90% 01:10:16.280 --> 01:10:16.780 align:middle line:90% OK. 01:10:16.780 --> 01:10:19.580 align:middle line:84% So the new construction-- let's look at the equation. 01:10:19.580 --> 01:10:22.320 align:middle line:90% This is a little bit-- 01:10:22.320 --> 01:10:25.160 align:middle line:84% so you have the total facilities divided by time 01:10:25.160 --> 01:10:29.000 align:middle line:84% to expand multiplied by effective delivery 01:10:29.000 --> 01:10:30.700 align:middle line:90% delay on capacity expansion. 01:10:30.700 --> 01:10:32.560 align:middle line:84% Ann, can you explain what you were thinking 01:10:32.560 --> 01:10:34.700 align:middle line:90% in formulating it this way? 01:10:34.700 --> 01:10:39.400 align:middle line:90% 01:10:39.400 --> 01:10:40.220 align:middle line:90% AUDIENCE: Yeah. 01:10:40.220 --> 01:10:43.040 align:middle line:90% So I misunderstood the prompt. 01:10:43.040 --> 01:10:48.180 align:middle line:84% But I was thinking that just because there's a-- 01:10:48.180 --> 01:10:51.060 align:middle line:84% I was thinking that the decision to build new facilities 01:10:51.060 --> 01:10:57.500 align:middle line:84% depends on your particular threshold of customer 01:10:57.500 --> 01:11:01.700 align:middle line:84% dissatisfaction because of the cost in building new facilities. 01:11:01.700 --> 01:11:04.940 align:middle line:84% So even though you're not meeting the customer 01:11:04.940 --> 01:11:07.340 align:middle line:84% expectations, there's still that delay 01:11:07.340 --> 01:11:11.240 align:middle line:84% until a certain threshold has been exceeded. 01:11:11.240 --> 01:11:13.540 align:middle line:84% And then they decide to create new buildings. 01:11:13.540 --> 01:11:18.340 align:middle line:84% And they just increase by, like, a percentage 01:11:18.340 --> 01:11:20.760 align:middle line:84% based on the current facilities that they have. 01:11:20.760 --> 01:11:21.718 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:11:21.718 --> 01:11:25.180 align:middle line:84% So this is not really doing what you were thinking 01:11:25.180 --> 01:11:27.240 align:middle line:90% or what you need to do here. 01:11:27.240 --> 01:11:29.060 align:middle line:84% Because basically, this one is creating 01:11:29.060 --> 01:11:32.000 align:middle line:84% a very strong first order reinforcing loop that says, 01:11:32.000 --> 01:11:33.500 align:middle line:84% the more facility you have, the more 01:11:33.500 --> 01:11:38.380 align:middle line:84% you will build, with some attenuating 01:11:38.380 --> 01:11:40.480 align:middle line:84% effect coming from effective delivery 01:11:40.480 --> 01:11:41.980 align:middle line:90% delay on capacity expansion. 01:11:41.980 --> 01:11:46.840 align:middle line:84% So if your total capacity is very high, 01:11:46.840 --> 01:11:51.520 align:middle line:84% then probably this delivery delay effect gets smaller. 01:11:51.520 --> 01:11:55.880 align:middle line:84% And at some point, it may slow down or even stop. 01:11:55.880 --> 01:11:59.760 align:middle line:84% So this is a little bit additional complexity here 01:11:59.760 --> 01:12:00.500 align:middle line:90% to understand. 01:12:00.500 --> 01:12:03.740 align:middle line:84% But this is kind of 1 minus e to the power of something. 01:12:03.740 --> 01:12:11.820 align:middle line:84% So it's probably going down with more gap in delivery delay. 01:12:11.820 --> 01:12:16.000 align:middle line:90% 01:12:16.000 --> 01:12:21.120 align:middle line:84% 1 minus e to the power of minus-- so it goes up like this. 01:12:21.120 --> 01:12:22.240 align:middle line:90% Yeah. 01:12:22.240 --> 01:12:26.100 align:middle line:84% So if you don't have any delivery delay, it can go to 0. 01:12:26.100 --> 01:12:29.680 align:middle line:84% If this goes to 0, then you wouldn't 01:12:29.680 --> 01:12:31.260 align:middle line:90% be building new capacity. 01:12:31.260 --> 01:12:33.020 align:middle line:84% So it may look like it is working, 01:12:33.020 --> 01:12:36.520 align:middle line:84% but it kind of works in a very weird way that doesn't directly 01:12:36.520 --> 01:12:40.900 align:middle line:84% map into useful real-world constructs. 01:12:40.900 --> 01:12:43.460 align:middle line:84% So the solution, again, is basically 01:12:43.460 --> 01:12:47.540 align:middle line:84% a factor for replacement of death to buildings 01:12:47.540 --> 01:12:52.020 align:middle line:84% and a factor for adjusting toward your desired level. 01:12:52.020 --> 01:12:56.220 align:middle line:84% So there is a desired level of capacity minus facility 01:12:56.220 --> 01:12:57.640 align:middle line:90% minus current facility. 01:12:57.640 --> 01:13:02.620 align:middle line:84% And that desired facility could come from your current facility 01:13:02.620 --> 01:13:05.260 align:middle line:84% as well multiplied by some factor that 01:13:05.260 --> 01:13:08.160 align:middle line:84% is expanding the facility based on delivery delay. 01:13:08.160 --> 01:13:10.540 align:middle line:84% So something along those lines might 01:13:10.540 --> 01:13:14.780 align:middle line:90% be what you need to have here. 01:13:14.780 --> 01:13:20.860 align:middle line:84% Backlog part fairly simple-- and simple is good here. 01:13:20.860 --> 01:13:24.300 align:middle line:90% Just have new orders coming in. 01:13:24.300 --> 01:13:26.720 align:middle line:84% And you have the orders versus customers. 01:13:26.720 --> 01:13:28.320 align:middle line:84% So you have the change of units here. 01:13:28.320 --> 01:13:29.940 align:middle line:90% That's fine. 01:13:29.940 --> 01:13:35.320 align:middle line:84% And then the delivery of units is determined by what? 01:13:35.320 --> 01:13:38.040 align:middle line:90% 01:13:38.040 --> 01:13:40.700 align:middle line:84% Backlog divided by delivery delay? 01:13:40.700 --> 01:13:43.680 align:middle line:90% Oh. 01:13:43.680 --> 01:13:45.900 align:middle line:90% Oh. 01:13:45.900 --> 01:13:46.400 align:middle line:90% Huh. 01:13:46.400 --> 01:13:47.080 align:middle line:90% OK. 01:13:47.080 --> 01:13:48.420 align:middle line:90% Yeah, this works. 01:13:48.420 --> 01:13:51.160 align:middle line:84% Ann, this is an interesting piece. 01:13:51.160 --> 01:13:54.320 align:middle line:84% What you have is right, but it's not following 01:13:54.320 --> 01:13:56.380 align:middle line:90% the classical formulations. 01:13:56.380 --> 01:13:58.920 align:middle line:84% And it creates a little bit of a confusion for me 01:13:58.920 --> 01:14:00.160 align:middle line:90% when I first saw it. 01:14:00.160 --> 01:14:03.760 align:middle line:84% So typically, we would have the delivery 01:14:03.760 --> 01:14:07.320 align:middle line:84% to be regulated by the production capacity. 01:14:07.320 --> 01:14:11.240 align:middle line:84% So it would be the minimum of production capacity 01:14:11.240 --> 01:14:14.820 align:middle line:90% and maximum delivery rate. 01:14:14.820 --> 01:14:16.440 align:middle line:84% And that maximum delivery rate would 01:14:16.440 --> 01:14:20.360 align:middle line:84% come from backlog of orders divided by a minimum time 01:14:20.360 --> 01:14:21.520 align:middle line:90% to deliver. 01:14:21.520 --> 01:14:23.800 align:middle line:84% So that would be your order delivery. 01:14:23.800 --> 01:14:26.520 align:middle line:84% Then backlog divided by order delivery 01:14:26.520 --> 01:14:28.440 align:middle line:90% gives you the delivery delay. 01:14:28.440 --> 01:14:30.660 align:middle line:84% The way you have calculated it is the other way. 01:14:30.660 --> 01:14:33.080 align:middle line:84% That is, you have calculated the delivery delay 01:14:33.080 --> 01:14:36.660 align:middle line:90% using production versus backlog. 01:14:36.660 --> 01:14:40.820 align:middle line:84% And you have basically divided the backlog 01:14:40.820 --> 01:14:43.820 align:middle line:90% by production capacity. 01:14:43.820 --> 01:14:47.740 align:middle line:84% AUDIENCE: [INAUDIBLE] idea was like there's a fixed 01:14:47.740 --> 01:14:50.280 align:middle line:90% capacity for each facility. 01:14:50.280 --> 01:14:54.740 align:middle line:84% And as the backlog increases, then there's 01:14:54.740 --> 01:14:57.038 align:middle line:84% a delay in delivery to customers. 01:14:57.038 --> 01:14:59.080 align:middle line:84% HAZHIR RAHMANDAD: No, I'm not saying it is wrong. 01:14:59.080 --> 01:15:01.980 align:middle line:84% I'm saying this is kind of not the typical way we formulate 01:15:01.980 --> 01:15:03.060 align:middle line:90% this. 01:15:03.060 --> 01:15:06.180 align:middle line:84% And the way you have formulated it also kind of 01:15:06.180 --> 01:15:11.260 align:middle line:84% takes away this constraint of maximum delivery 01:15:11.260 --> 01:15:14.720 align:middle line:84% rate coming from not having enough in the backlog. 01:15:14.720 --> 01:15:20.780 align:middle line:90% 01:15:20.780 --> 01:15:21.940 align:middle line:90% Yeah. 01:15:21.940 --> 01:15:24.520 align:middle line:84% And again, it doesn't make a big difference here, 01:15:24.520 --> 01:15:29.280 align:middle line:84% but it is just a little bit non-standard formulation. 01:15:29.280 --> 01:15:30.940 align:middle line:84% AUDIENCE: Isn't deliveries in this case 01:15:30.940 --> 01:15:33.320 align:middle line:84% always equal to production capacity [INAUDIBLE]? 01:15:33.320 --> 01:15:37.520 align:middle line:84% HAZHIR RAHMANDAD: So, Yes, deliveries will all-- 01:15:37.520 --> 01:15:38.020 align:middle line:90% no. 01:15:38.020 --> 01:15:40.300 align:middle line:90% So when the backlog goes to 0-- 01:15:40.300 --> 01:15:43.640 align:middle line:90% 01:15:43.640 --> 01:15:47.740 align:middle line:84% but then you're dividing it by this delivery delay that is 0. 01:15:47.740 --> 01:15:49.740 align:middle line:84% Then it is always production capacity. 01:15:49.740 --> 01:15:52.860 align:middle line:84% And therefore, you don't have first order control. 01:15:52.860 --> 01:15:55.420 align:middle line:84% You will get problematic behavior. 01:15:55.420 --> 01:15:56.100 align:middle line:90% You're right. 01:15:56.100 --> 01:15:59.140 align:middle line:90% 01:15:59.140 --> 01:16:05.880 align:middle line:84% Ann, what Alexander noted here is that when 01:16:05.880 --> 01:16:09.640 align:middle line:84% you multiply your backlog or divide your backlog 01:16:09.640 --> 01:16:13.320 align:middle line:84% by your delivery delay, remember that delivery delay is already 01:16:13.320 --> 01:16:15.940 align:middle line:84% your backlog divided by production capacity. 01:16:15.940 --> 01:16:18.620 align:middle line:84% So the backlogs cancel out each other. 01:16:18.620 --> 01:16:20.160 align:middle line:84% And basically, your order delivery 01:16:20.160 --> 01:16:22.440 align:middle line:90% becomes production capacity. 01:16:22.440 --> 01:16:26.260 align:middle line:84% And therefore, you wouldn't have the first order control. 01:16:26.260 --> 01:16:30.000 align:middle line:84% That is, you wouldn't have a limit that avoids 01:16:30.000 --> 01:16:33.060 align:middle line:90% your backlog going negative. 01:16:33.060 --> 01:16:36.460 align:middle line:84% Now, in your formulation, basically the model 01:16:36.460 --> 01:16:39.240 align:middle line:84% will crash because you will get to a backlog of 0. 01:16:39.240 --> 01:16:41.860 align:middle line:84% And then you will have a division by 0. 01:16:41.860 --> 01:16:45.060 align:middle line:90% That will be what will happen. 01:16:45.060 --> 01:16:48.920 align:middle line:84% So, yes, use the standard formulation. 01:16:48.920 --> 01:16:53.420 align:middle line:90% That's the summary of this case. 01:16:53.420 --> 01:17:01.580 align:middle line:84% Then we have potential customers becoming likely customers. 01:17:01.580 --> 01:17:03.460 align:middle line:84% And there is kind of effective delivery 01:17:03.460 --> 01:17:05.140 align:middle line:90% delay on conversion rate. 01:17:05.140 --> 01:17:11.060 align:middle line:90% 01:17:11.060 --> 01:17:12.500 align:middle line:84% So the conversion rate-- these are 01:17:12.500 --> 01:17:17.460 align:middle line:84% people who have learned about the product 01:17:17.460 --> 01:17:20.340 align:middle line:90% through the sales channel. 01:17:20.340 --> 01:17:20.840 align:middle line:90% OK. 01:17:20.840 --> 01:17:23.860 align:middle line:84% So the conceptualization here, Ann, 01:17:23.860 --> 01:17:25.800 align:middle line:84% if I'm understanding it correctly, 01:17:25.800 --> 01:17:28.300 align:middle line:84% is that the main effect is happening 01:17:28.300 --> 01:17:30.100 align:middle line:90% through this conversion rate. 01:17:30.100 --> 01:17:32.880 align:middle line:84% And once people have converted, then with the time 01:17:32.880 --> 01:17:34.980 align:middle line:90% constant they order. 01:17:34.980 --> 01:17:38.040 align:middle line:84% So this is your order rate, customers placing orders. 01:17:38.040 --> 01:17:43.780 align:middle line:90% And then orders are delivered. 01:17:43.780 --> 01:17:45.460 align:middle line:84% They become customers with products. 01:17:45.460 --> 01:17:48.860 align:middle line:84% And after a while, they go back and buy it again. 01:17:48.860 --> 01:17:49.360 align:middle line:90% So-- 01:17:49.360 --> 01:17:49.943 align:middle line:90% AUDIENCE: Yep. 01:17:49.943 --> 01:17:52.840 align:middle line:84% HAZHIR RAHMANDAD: --you are not using the sales force 01:17:52.840 --> 01:17:57.940 align:middle line:90% to basically to do two things. 01:17:57.940 --> 01:18:01.320 align:middle line:84% One is to convince these people that 01:18:01.320 --> 01:18:06.260 align:middle line:84% have stopped using the product to come back and buy it again. 01:18:06.260 --> 01:18:08.880 align:middle line:84% So this step is happening automatically. 01:18:08.880 --> 01:18:11.840 align:middle line:84% Secondly, you are adding this delay 01:18:11.840 --> 01:18:15.560 align:middle line:84% between the time people are contacted and convinced 01:18:15.560 --> 01:18:19.920 align:middle line:84% by the sales force until they decide to put an order. 01:18:19.920 --> 01:18:22.300 align:middle line:90% Now, that second one is fine. 01:18:22.300 --> 01:18:24.680 align:middle line:84% I think conceptually that's not a problem. 01:18:24.680 --> 01:18:26.300 align:middle line:84% The first one is more problematic. 01:18:26.300 --> 01:18:28.980 align:middle line:84% You do need sales force to also convince 01:18:28.980 --> 01:18:34.740 align:middle line:84% people who need to do a replacement purchase. 01:18:34.740 --> 01:18:36.580 align:middle line:90% AUDIENCE: OK. 01:18:36.580 --> 01:18:38.780 align:middle line:90% HAZHIR RAHMANDAD: OK. 01:18:38.780 --> 01:18:39.340 align:middle line:90% Let's see. 01:18:39.340 --> 01:18:40.640 align:middle line:90% Anything else here? 01:18:40.640 --> 01:18:45.060 align:middle line:84% So sales force-- you have the replacement time. 01:18:45.060 --> 01:18:48.460 align:middle line:84% Oh, time sales force to be fired. 01:18:48.460 --> 01:18:49.980 align:middle line:84% AUDIENCE: So this part [INAUDIBLE] 01:18:49.980 --> 01:18:55.980 align:middle line:84% but I think the idea was that there's 01:18:55.980 --> 01:18:59.898 align:middle line:84% a percentage of revenue reserved to pay the wages of sales force. 01:18:59.898 --> 01:19:00.940 align:middle line:90% HAZHIR RAHMANDAD: Uh-huh. 01:19:00.940 --> 01:19:03.460 align:middle line:90% 01:19:03.460 --> 01:19:11.280 align:middle line:84% AUDIENCE: And if the cash available is too small, 01:19:11.280 --> 01:19:14.720 align:middle line:84% then they would fire some members of the sales 01:19:14.720 --> 01:19:17.820 align:middle line:84% force so that they always have that money to pay them. 01:19:17.820 --> 01:19:20.093 align:middle line:84% And then if the cash builds up again, 01:19:20.093 --> 01:19:21.260 align:middle line:90% they would hire [INAUDIBLE]. 01:19:21.260 --> 01:19:23.028 align:middle line:84% I didn't have time to formulate that. 01:19:23.028 --> 01:19:24.320 align:middle line:90% HAZHIR RAHMANDAD: I see, I see. 01:19:24.320 --> 01:19:27.080 align:middle line:84% So basically, you're kind of having money that 01:19:27.080 --> 01:19:29.640 align:middle line:90% is put aside for sales force. 01:19:29.640 --> 01:19:30.365 align:middle line:90% And if you have-- 01:19:30.365 --> 01:19:31.240 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:19:31.240 --> 01:19:32.198 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:19:32.198 --> 01:19:35.240 align:middle line:84% And if you have enough money, you 01:19:35.240 --> 01:19:36.900 align:middle line:90% are keeping your sales force. 01:19:36.900 --> 01:19:39.095 align:middle line:90% Otherwise, you're firing them. 01:19:39.095 --> 01:19:39.720 align:middle line:90% AUDIENCE: Yeah. 01:19:39.720 --> 01:19:42.262 align:middle line:84% HAZHIR RAHMANDAD: And then you have an exogenous hiring rate. 01:19:42.262 --> 01:19:43.720 align:middle line:90% You're just hiring people. 01:19:43.720 --> 01:19:45.080 align:middle line:90% So-- 01:19:45.080 --> 01:19:47.560 align:middle line:84% AUDIENCE: No, that part is not complete. 01:19:47.560 --> 01:19:52.600 align:middle line:84% I was supposed to replace the people fired 01:19:52.600 --> 01:19:54.960 align:middle line:84% as well when the cash builds up, but I 01:19:54.960 --> 01:19:56.500 align:middle line:90% didn't have time to put it in. 01:19:56.500 --> 01:19:57.380 align:middle line:90% But that's the idea. 01:19:57.380 --> 01:20:01.260 align:middle line:84% HAZHIR RAHMANDAD: So it is somewhat convoluted. 01:20:01.260 --> 01:20:04.320 align:middle line:84% Mostly, people control their stock of employees 01:20:04.320 --> 01:20:06.120 align:middle line:90% through hiring. 01:20:06.120 --> 01:20:07.500 align:middle line:90% Firing does happen. 01:20:07.500 --> 01:20:10.200 align:middle line:84% So that's OK to have a firing that is also determined 01:20:10.200 --> 01:20:14.760 align:middle line:90% based on money you have. 01:20:14.760 --> 01:20:19.120 align:middle line:84% But adding this stock of cash that is reserved for salaries 01:20:19.120 --> 01:20:23.240 align:middle line:84% is kind of not adding that much value in terms 01:20:23.240 --> 01:20:26.620 align:middle line:84% of, it kind of just adds a bunch of parameters 01:20:26.620 --> 01:20:29.860 align:middle line:84% and additional structure you need to manage. 01:20:29.860 --> 01:20:32.460 align:middle line:84% Even having explicit hiring and firing 01:20:32.460 --> 01:20:34.840 align:middle line:84% may not be the most effective way. 01:20:34.840 --> 01:20:37.580 align:middle line:84% You can just have an adjustment of sales force using 01:20:37.580 --> 01:20:41.460 align:middle line:84% a first order adjustment process. 01:20:41.460 --> 01:20:43.620 align:middle line:84% And then the desired sales force can easily 01:20:43.620 --> 01:20:46.760 align:middle line:84% come from your budget for the sales force. 01:20:46.760 --> 01:20:49.680 align:middle line:84% I think that's what others also had in this formulation. 01:20:49.680 --> 01:20:51.780 align:middle line:90% That's much simpler. 01:20:51.780 --> 01:20:59.060 align:middle line:84% If you think people are less eager to fire people 01:20:59.060 --> 01:21:00.940 align:middle line:84% than to hire people, which is part 01:21:00.940 --> 01:21:03.440 align:middle line:84% of the real organizational dynamics, 01:21:03.440 --> 01:21:06.520 align:middle line:84% you can make it an asymmetric delay and adjustment 01:21:06.520 --> 01:21:10.340 align:middle line:84% so that, when the desired level of sales force 01:21:10.340 --> 01:21:13.920 align:middle line:84% is higher than the current level, 01:21:13.920 --> 01:21:15.440 align:middle line:90% you adjust it much faster. 01:21:15.440 --> 01:21:19.620 align:middle line:84% When the desired level is lower, then you adjust it more slowly. 01:21:19.620 --> 01:21:22.640 align:middle line:84% In the extreme, it would be how long people typically 01:21:22.640 --> 01:21:25.740 align:middle line:84% stay in your company, which is for the downward adjustment. 01:21:25.740 --> 01:21:28.940 align:middle line:84% That is, you just let it happen through natural turnover, 01:21:28.940 --> 01:21:31.980 align:middle line:84% not through actually firing people. 01:21:31.980 --> 01:21:36.820 align:middle line:84% So that's a simple way to take care of the hiring and firing, 01:21:36.820 --> 01:21:40.143 align:middle line:84% even if they are asymmetric in terms of adjustment. 01:21:40.143 --> 01:21:41.560 align:middle line:84% AUDIENCE: I think my confusion was 01:21:41.560 --> 01:21:44.920 align:middle line:84% like, if there's a fixed amount that is allocated to the sales 01:21:44.920 --> 01:21:50.097 align:middle line:84% force, then how do you control if that amount is [INAUDIBLE] 01:21:50.097 --> 01:21:52.180 align:middle line:84% HAZHIR RAHMANDAD: So what's the fixed amount unit? 01:21:52.180 --> 01:21:56.520 align:middle line:84% So what's the unit of what is allocated to sales force? 01:21:56.520 --> 01:21:59.920 align:middle line:84% AUDIENCE: Maybe a dollar value per time period 01:21:59.920 --> 01:22:01.580 align:middle line:90% as a percentage of the revenue. 01:22:01.580 --> 01:22:02.580 align:middle line:90% HAZHIR RAHMANDAD: Great. 01:22:02.580 --> 01:22:04.480 align:middle line:90% So let's say dollar per month. 01:22:04.480 --> 01:22:07.500 align:middle line:90% So I have $1,000 per month. 01:22:07.500 --> 01:22:10.420 align:middle line:84% I mean, $1,000-- who comes to work for me for that price? 01:22:10.420 --> 01:22:16.160 align:middle line:84% But let's say $1,000 per month of budget for my sales force. 01:22:16.160 --> 01:22:21.020 align:middle line:84% And then each salesperson, now that we are so cheap, 01:22:21.020 --> 01:22:24.020 align:middle line:84% is willing to work for $50 a month. 01:22:24.020 --> 01:22:30.020 align:middle line:84% So what's the number of sales force I can have? 01:22:30.020 --> 01:22:31.700 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:22:31.700 --> 01:22:36.040 align:middle line:84% HAZHIR RAHMANDAD: So that's 20 sales force. 01:22:36.040 --> 01:22:38.560 align:middle line:84% You're adjusting your current stock towards that desired. 01:22:38.560 --> 01:22:39.768 align:middle line:90% You take care of all of that. 01:22:39.768 --> 01:22:42.420 align:middle line:84% You don't need to worry about another stock that accumulates 01:22:42.420 --> 01:22:46.900 align:middle line:84% that $1,000 and then tries to spend it. 01:22:46.900 --> 01:22:48.900 align:middle line:90% It's just too many things. 01:22:48.900 --> 01:22:49.400 align:middle line:90% OK. 01:22:49.400 --> 01:22:49.900 align:middle line:90% Good. 01:22:49.900 --> 01:22:51.940 align:middle line:84% I think it's a good time to take a break-- 01:22:51.940 --> 01:22:53.060 align:middle line:90% 10 minutes. 01:22:53.060 --> 01:22:55.560 align:middle line:84% And then we come back, look at a few more models. 01:22:55.560 --> 01:22:58.580 align:middle line:90% 01:22:58.580 --> 01:23:03.140 align:middle line:84% So one very minor thing here is, you 01:23:03.140 --> 01:23:06.640 align:middle line:84% are starting the DELAY3 function without initial. 01:23:06.640 --> 01:23:10.540 align:middle line:84% So it is initializing it in equilibrium. 01:23:10.540 --> 01:23:14.860 align:middle line:84% But the stock here of production capacity, PC, 01:23:14.860 --> 01:23:17.880 align:middle line:90% is starting with initial PC. 01:23:17.880 --> 01:23:21.880 align:middle line:84% So this creates an internal inconsistency. 01:23:21.880 --> 01:23:24.760 align:middle line:84% So this is a little bit kind of subtle there. 01:23:24.760 --> 01:23:30.760 align:middle line:84% Because in this link from PC completion to PC depreciation, 01:23:30.760 --> 01:23:35.560 align:middle line:84% there are three stock variables hidden behind the scenes. 01:23:35.560 --> 01:23:37.420 align:middle line:84% And they have their own initial value. 01:23:37.420 --> 01:23:43.160 align:middle line:84% The initial value used there is basically the PC completion 01:23:43.160 --> 01:23:49.240 align:middle line:84% rate at the beginning multiplied by facility life divided 01:23:49.240 --> 01:23:52.960 align:middle line:90% by 3 for each of those stocks. 01:23:52.960 --> 01:23:58.320 align:middle line:84% That's what is being used in the back end. 01:23:58.320 --> 01:24:01.200 align:middle line:84% But what you're using as the sum of those three 01:24:01.200 --> 01:24:04.620 align:middle line:84% that should be the production capacity is initial PC. 01:24:04.620 --> 01:24:08.400 align:middle line:84% And there is no reason here to automatically expect 01:24:08.400 --> 01:24:10.020 align:middle line:90% these two to be the same. 01:24:10.020 --> 01:24:14.400 align:middle line:84% And that can create weird, unexpected behaviors. 01:24:14.400 --> 01:24:20.640 align:middle line:84% So I would have it as a DELAY3 EI with initial condition 01:24:20.640 --> 01:24:25.260 align:middle line:84% and set the initial condition to be your initial PC divided 01:24:25.260 --> 01:24:27.440 align:middle line:90% by facility life. 01:24:27.440 --> 01:24:28.992 align:middle line:90% Then you're all set. 01:24:28.992 --> 01:24:31.200 align:middle line:84% AUDIENCE: Can I ask how that takes care of the issue? 01:24:31.200 --> 01:24:34.580 align:middle line:84% Because PC divided by facility life 01:24:34.580 --> 01:24:36.660 align:middle line:84% only happens for the initial time step. 01:24:36.660 --> 01:24:39.920 align:middle line:84% So after you drain whatever fraction is there, 01:24:39.920 --> 01:24:42.580 align:middle line:90% then you go back to just only-- 01:24:42.580 --> 01:24:44.540 align:middle line:90% like, do you see what I mean? 01:24:44.540 --> 01:24:46.540 align:middle line:90% You're still draining-- 01:24:46.540 --> 01:24:47.860 align:middle line:90% I guess I'm just asking is-- 01:24:47.860 --> 01:24:50.660 align:middle line:84% HAZHIR RAHMANDAD: So I'm trying to get the production 01:24:50.660 --> 01:24:53.940 align:middle line:84% capacity, PC, to be exactly the same as some 01:24:53.940 --> 01:24:57.660 align:middle line:84% of those three initial stocks that 01:24:57.660 --> 01:25:00.120 align:middle line:90% are hidden behind this link. 01:25:00.120 --> 01:25:00.620 align:middle line:90% OK? 01:25:00.620 --> 01:25:03.580 align:middle line:84% Some of those three initial stocks 01:25:03.580 --> 01:25:07.700 align:middle line:84% would be whatever I give it as the initial outflow, 01:25:07.700 --> 01:25:10.100 align:middle line:90% initial value of the delay. 01:25:10.100 --> 01:25:12.200 align:middle line:84% The initial value of delay is the outflow, 01:25:12.200 --> 01:25:15.960 align:middle line:90% in this case, PC depreciation. 01:25:15.960 --> 01:25:19.160 align:middle line:84% So it would be that initial outflow 01:25:19.160 --> 01:25:23.520 align:middle line:90% multiplied by facility life. 01:25:23.520 --> 01:25:27.920 align:middle line:84% So by setting that initial value to be 01:25:27.920 --> 01:25:30.320 align:middle line:84% initial PC divided by facility life, 01:25:30.320 --> 01:25:32.960 align:middle line:90% I achieve this equivalence. 01:25:32.960 --> 01:25:35.900 align:middle line:84% And once I started from the equivalent value, 01:25:35.900 --> 01:25:37.760 align:middle line:84% they will remain equivalent all the way 01:25:37.760 --> 01:25:42.160 align:middle line:84% through because the inflow and outflow of production capacity 01:25:42.160 --> 01:25:43.780 align:middle line:90% PC is coming from this-- 01:25:43.780 --> 01:25:46.080 align:middle line:84% I mean, the outflow is coming from this delay. 01:25:46.080 --> 01:25:48.980 align:middle line:84% That inflow is also this initial of the delay. 01:25:48.980 --> 01:25:54.420 align:middle line:84% So then they are always going to be the same moving forward. 01:25:54.420 --> 01:25:54.920 align:middle line:90% OK. 01:25:54.920 --> 01:26:00.640 align:middle line:84% So WIP first order delay-- that's fine. 01:26:00.640 --> 01:26:03.300 align:middle line:90% PC building is coming from-- 01:26:03.300 --> 01:26:03.800 align:middle line:90% OK. 01:26:03.800 --> 01:26:06.080 align:middle line:90% Let's see what is happening. 01:26:06.080 --> 01:26:08.760 align:middle line:84% Desired PC building-- this is probably max of 0. 01:26:08.760 --> 01:26:11.600 align:middle line:84% And desired PC building-- something that you always 01:26:11.600 --> 01:26:14.140 align:middle line:90% want to make sure is in place. 01:26:14.140 --> 01:26:16.340 align:middle line:90% You have an adjustment for WIP. 01:26:16.340 --> 01:26:21.140 align:middle line:84% You have indicated expansion to meet demand 01:26:21.140 --> 01:26:28.360 align:middle line:84% based on the current production capacity. 01:26:28.360 --> 01:26:31.340 align:middle line:90% 01:26:31.340 --> 01:26:31.840 align:middle line:90% OK. 01:26:31.840 --> 01:26:33.360 align:middle line:84% So let's see what is happening here. 01:26:33.360 --> 01:26:37.260 align:middle line:90% 01:26:37.260 --> 01:26:43.680 align:middle line:84% So you're looking at the total need divided by current value. 01:26:43.680 --> 01:26:46.360 align:middle line:84% So this is desired stock minus stock. 01:26:46.360 --> 01:26:50.180 align:middle line:90% So this is a stock adjustment. 01:26:50.180 --> 01:27:01.900 align:middle line:84% This stock adjustment plus the replacement-- 01:27:01.900 --> 01:27:05.780 align:middle line:84% and what is the expansion commitment delay 01:27:05.780 --> 01:27:08.000 align:middle line:90% you have here, Alban? 01:27:08.000 --> 01:27:10.840 align:middle line:84% Do you have a kind of a story there? 01:27:10.840 --> 01:27:17.160 align:middle line:84% AUDIENCE: So I was trying to capture, 01:27:17.160 --> 01:27:18.660 align:middle line:90% you have decision making time. 01:27:18.660 --> 01:27:21.940 align:middle line:84% Like, you don't come into inspection immediately 01:27:21.940 --> 01:27:22.440 align:middle line:90% [INAUDIBLE]. 01:27:22.440 --> 01:27:23.240 align:middle line:90% HAZHIR RAHMANDAD: OK. 01:27:23.240 --> 01:27:23.780 align:middle line:90% OK. 01:27:23.780 --> 01:27:25.240 align:middle line:90% That's fine. 01:27:25.240 --> 01:27:29.060 align:middle line:84% That's conceptually correct-- maybe an overkill here, 01:27:29.060 --> 01:27:30.920 align:middle line:90% but that's OK. 01:27:30.920 --> 01:27:32.780 align:middle line:90% And you were missing that unit. 01:27:32.780 --> 01:27:33.440 align:middle line:90% OK. 01:27:33.440 --> 01:27:35.020 align:middle line:90% Or maybe I deleted it somehow. 01:27:35.020 --> 01:27:38.080 align:middle line:90% 01:27:38.080 --> 01:27:43.480 align:middle line:84% Then desired completion is just the sum of these two. 01:27:43.480 --> 01:27:46.060 align:middle line:84% And then you add the adjustment for WIP. 01:27:46.060 --> 01:27:47.380 align:middle line:90% So this is fairly standard. 01:27:47.380 --> 01:27:50.560 align:middle line:84% So the key question is, where does the indicated total PC need 01:27:50.560 --> 01:27:52.040 align:middle line:90% come from. 01:27:52.040 --> 01:27:55.380 align:middle line:84% You are looking at the excess delay and the current. 01:27:55.380 --> 01:27:57.960 align:middle line:90% 01:27:57.960 --> 01:28:00.460 align:middle line:84% AUDIENCE: So you're just scaling up your production capacity 01:28:00.460 --> 01:28:01.400 align:middle line:90% but the [INAUDIBLE]. 01:28:01.400 --> 01:28:05.320 align:middle line:84% HAZHIR RAHMANDAD: Yes, with some delay effect. 01:28:05.320 --> 01:28:08.980 align:middle line:84% And this excess delay effect is coming 01:28:08.980 --> 01:28:15.300 align:middle line:84% from the actual delivery delay versus desired delivery delay. 01:28:15.300 --> 01:28:19.500 align:middle line:84% So this is a fairly rapid perception 01:28:19.500 --> 01:28:21.140 align:middle line:90% of the delivery delay. 01:28:21.140 --> 01:28:24.620 align:middle line:84% But because you have this additional delay for expansion 01:28:24.620 --> 01:28:28.460 align:middle line:84% commitment, that delay is kind of captured there in a sense. 01:28:28.460 --> 01:28:33.700 align:middle line:84% So we didn't have an explicit separate delay for perception 01:28:33.700 --> 01:28:35.320 align:middle line:90% of delivery delay by company. 01:28:35.320 --> 01:28:39.140 align:middle line:84% But instead, we have an adjustment there. 01:28:39.140 --> 01:28:43.260 align:middle line:84% And this excess delay is just a ratio of these two. 01:28:43.260 --> 01:28:46.620 align:middle line:84% Similar to Kathy, you are limiting it 01:28:46.620 --> 01:28:51.060 align:middle line:84% to a floor in this case because you are using 01:28:51.060 --> 01:28:53.040 align:middle line:90% the actual versus desired. 01:28:53.040 --> 01:28:56.700 align:middle line:90% 01:28:56.700 --> 01:29:00.060 align:middle line:84% And you only care when the situation is 01:29:00.060 --> 01:29:03.020 align:middle line:90% that things are getting bigger. 01:29:03.020 --> 01:29:04.860 align:middle line:84% So this is kind of an important question 01:29:04.860 --> 01:29:09.240 align:middle line:84% here-- whether you want to allow for reductions in the capacity. 01:29:09.240 --> 01:29:13.840 align:middle line:84% And in this model, I would want to see reductions in capacity, 01:29:13.840 --> 01:29:17.520 align:middle line:84% because you could have a situation where 01:29:17.520 --> 01:29:22.080 align:middle line:84% the steady state, long-term replacement is 01:29:22.080 --> 01:29:29.400 align:middle line:84% much lower than the growth phase replacement plus new purchase. 01:29:29.400 --> 01:29:32.360 align:middle line:84% And therefore, you want to include a pathway 01:29:32.360 --> 01:29:35.160 align:middle line:90% for reductions in capacity. 01:29:35.160 --> 01:29:39.040 align:middle line:84% Whereas, if I'm correct, this model right 01:29:39.040 --> 01:29:42.533 align:middle line:84% now doesn't allow for reductions in capacity. 01:29:42.533 --> 01:29:43.200 align:middle line:90% AUDIENCE: Right. 01:29:43.200 --> 01:29:46.800 align:middle line:84% HAZHIR RAHMANDAD: So is the logic clear 01:29:46.800 --> 01:29:48.620 align:middle line:90% how I made that conclusion? 01:29:48.620 --> 01:29:52.400 align:middle line:90% So let me backtrack. 01:29:52.400 --> 01:29:58.680 align:middle line:84% So the indicated total PC need is current production capacity 01:29:58.680 --> 01:30:02.240 align:middle line:84% multiplied by this expansion factor coming from excess delay. 01:30:02.240 --> 01:30:05.140 align:middle line:84% If excess delay never goes below 1, 01:30:05.140 --> 01:30:08.500 align:middle line:84% this indicates that total production capacity never goes 01:30:08.500 --> 01:30:10.280 align:middle line:84% below the current production capacity 01:30:10.280 --> 01:30:15.240 align:middle line:84% and therefore never can go down in terms of the desired value. 01:30:15.240 --> 01:30:18.660 align:middle line:84% So that's why I inferred that, by this formulation, 01:30:18.660 --> 01:30:23.060 align:middle line:84% you are making it impossible for the production capacity 01:30:23.060 --> 01:30:24.960 align:middle line:90% to go down. 01:30:24.960 --> 01:30:26.460 align:middle line:84% And that's not what you really want. 01:30:26.460 --> 01:30:31.240 align:middle line:84% That takes away some flexibility you do want to have. 01:30:31.240 --> 01:30:31.740 align:middle line:90% OK. 01:30:31.740 --> 01:30:35.860 align:middle line:84% So that's the capacity management backlog. 01:30:35.860 --> 01:30:37.840 align:middle line:90% Pretty straightforward-- sales. 01:30:37.840 --> 01:30:41.460 align:middle line:84% And then you have the delivery taken out. 01:30:41.460 --> 01:30:45.880 align:middle line:84% And the delivery is dependent on the capacity utilization. 01:30:45.880 --> 01:30:49.820 align:middle line:84% You're using a very similar formulation 01:30:49.820 --> 01:30:55.500 align:middle line:84% as we had in the widgets model with a little bit of more 01:30:55.500 --> 01:30:59.460 align:middle line:90% complex formulation here. 01:30:59.460 --> 01:31:01.795 align:middle line:84% AUDIENCE: It's a fuzzy mean [INAUDIBLE] 01:31:01.795 --> 01:31:03.920 align:middle line:84% HAZHIR RAHMANDAD: Yeah, that's the fuzzy mean thing 01:31:03.920 --> 01:31:07.460 align:middle line:84% that Hassan proposed a few weeks ago. 01:31:07.460 --> 01:31:09.880 align:middle line:90% Yeah. 01:31:09.880 --> 01:31:11.760 align:middle line:90% Cool. 01:31:11.760 --> 01:31:13.920 align:middle line:84% AUDIENCE: The only downside of that 01:31:13.920 --> 01:31:16.840 align:middle line:84% is-- and this is what I was struggling 01:31:16.840 --> 01:31:18.560 align:middle line:90% with that formulation. 01:31:18.560 --> 01:31:21.320 align:middle line:84% Let's say, if the desired production 01:31:21.320 --> 01:31:25.760 align:middle line:84% rate is smaller than the capacity, 01:31:25.760 --> 01:31:32.400 align:middle line:84% the company can decide to deliver faster than it's desired 01:31:32.400 --> 01:31:36.340 align:middle line:90% delivery to gain reputation. 01:31:36.340 --> 01:31:37.920 align:middle line:84% But that formula doesn't allow that. 01:31:37.920 --> 01:31:40.560 align:middle line:84% AUDIENCE: If it goes above that 45-degree line, [INAUDIBLE]. 01:31:40.560 --> 01:31:41.518 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:31:41.518 --> 01:31:45.580 align:middle line:84% So that formula by design wasn't going above 45-degree line. 01:31:45.580 --> 01:31:48.680 align:middle line:84% It is actually easy to have alternatives that go above 01:31:48.680 --> 01:31:50.120 align:middle line:90% 45-degree line. 01:31:50.120 --> 01:31:52.360 align:middle line:84% But ultimately, what that will do 01:31:52.360 --> 01:31:55.720 align:middle line:84% is that the actual delivery delay will never 01:31:55.720 --> 01:32:00.900 align:middle line:84% go below desired delivery delay in this formulation. 01:32:00.900 --> 01:32:04.380 align:middle line:84% And therefore, this excess delay actually 01:32:04.380 --> 01:32:09.980 align:middle line:84% would never go below 1, even if you didn't put the max function. 01:32:09.980 --> 01:32:15.820 align:middle line:84% And then that creates this issue of not tracking 01:32:15.820 --> 01:32:18.580 align:middle line:90% when the demand is falling. 01:32:18.580 --> 01:32:22.020 align:middle line:84% It wouldn't be able to bring down the capacity 01:32:22.020 --> 01:32:25.460 align:middle line:84% even if you removed that maximum that we talked about. 01:32:25.460 --> 01:32:28.260 align:middle line:84% So you see this is kind of a little bit, again, subtle. 01:32:28.260 --> 01:32:30.060 align:middle line:84% But this goes back to the conversation 01:32:30.060 --> 01:32:32.860 align:middle line:84% we had about whether using a pressure 01:32:32.860 --> 01:32:35.640 align:middle line:84% to drive capacity-- so you are here using a pressure. 01:32:35.640 --> 01:32:39.100 align:middle line:84% The pressure is coming from that delivery delay-- 01:32:39.100 --> 01:32:45.580 align:middle line:84% or using the recent order rate as the driver 01:32:45.580 --> 01:32:49.560 align:middle line:90% of desired capacity. 01:32:49.560 --> 01:32:52.500 align:middle line:84% Now, I told you to use pressure just so that we 01:32:52.500 --> 01:32:53.800 align:middle line:90% can have these conversations. 01:32:53.800 --> 01:32:56.380 align:middle line:84% But it is also showing the failure modes 01:32:56.380 --> 01:32:59.240 align:middle line:90% built into that pressure. 01:32:59.240 --> 01:33:01.900 align:middle line:84% Now, the Forrester version does bring it down as well. 01:33:01.900 --> 01:33:04.820 align:middle line:90% So he doesn't use this function. 01:33:04.820 --> 01:33:06.820 align:middle line:90% He uses a table function. 01:33:06.820 --> 01:33:11.320 align:middle line:84% And that table function allows for reducing 01:33:11.320 --> 01:33:15.600 align:middle line:84% the current capacity, which this one doesn't allow 01:33:15.600 --> 01:33:17.700 align:middle line:90% for just by these formulations. 01:33:17.700 --> 01:33:21.680 align:middle line:84% But I want you to be sensitive to how 01:33:21.680 --> 01:33:23.400 align:middle line:84% these assumptions in each equation 01:33:23.400 --> 01:33:25.100 align:middle line:84% has reverberations down the road. 01:33:25.100 --> 01:33:28.100 align:middle line:84% So think about those as you walk through your formulations. 01:33:28.100 --> 01:33:28.600 align:middle line:90% OK. 01:33:28.600 --> 01:33:29.840 align:middle line:90% Good. 01:33:29.840 --> 01:33:32.012 align:middle line:90% So we have the cells. 01:33:32.012 --> 01:33:33.720 align:middle line:84% And now you have the potential customers. 01:33:33.720 --> 01:33:37.100 align:middle line:90% They become new customers. 01:33:37.100 --> 01:33:39.940 align:middle line:90% And then you have the return. 01:33:39.940 --> 01:33:43.840 align:middle line:84% And depreciation and return is the same. 01:33:43.840 --> 01:33:46.080 align:middle line:90% So that creates sales. 01:33:46.080 --> 01:33:48.960 align:middle line:90% This also creates sales. 01:33:48.960 --> 01:33:55.000 align:middle line:84% But here, again, similar to Ann's model, 01:33:55.000 --> 01:33:58.140 align:middle line:84% you are not accounting for sales force 01:33:58.140 --> 01:34:01.920 align:middle line:84% being needed for attracting returning customers. 01:34:01.920 --> 01:34:06.260 align:middle line:84% So that's an assumption that might be a bit strong here. 01:34:06.260 --> 01:34:09.540 align:middle line:84% Now let's look at the capturing of new customers using 01:34:09.540 --> 01:34:11.580 align:middle line:90% the current sales force. 01:34:11.580 --> 01:34:15.500 align:middle line:90% So you have a number of pitches. 01:34:15.500 --> 01:34:18.640 align:middle line:84% And that number of pitches can come from productivity. 01:34:18.640 --> 01:34:21.900 align:middle line:84% That productivity is normal productivity and some max 01:34:21.900 --> 01:34:23.380 align:middle line:90% productivity. 01:34:23.380 --> 01:34:30.540 align:middle line:84% And max productivity depends on the ratio of current customers 01:34:30.540 --> 01:34:32.900 align:middle line:90% to the sales force. 01:34:32.900 --> 01:34:36.120 align:middle line:90% So that's-- 01:34:36.120 --> 01:34:42.220 align:middle line:90% 01:34:42.220 --> 01:34:43.800 align:middle line:90% AUDIENCE: Sales force times-- 01:34:43.800 --> 01:34:47.660 align:middle line:90% 01:34:47.660 --> 01:34:50.900 align:middle line:90% HAZHIR RAHMANDAD: Time to pitch. 01:34:50.900 --> 01:34:51.902 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:34:51.902 --> 01:34:52.860 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:34:52.860 --> 01:34:54.580 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:34:54.580 --> 01:34:56.920 align:middle line:84% HAZHIR RAHMANDAD: So it's similar to that-- yeah. 01:34:56.920 --> 01:34:58.740 align:middle line:90% So this is logical. 01:34:58.740 --> 01:35:01.700 align:middle line:84% The constraint works, creates the effect. 01:35:01.700 --> 01:35:04.680 align:middle line:84% But it is not a flexible effect because you 01:35:04.680 --> 01:35:07.040 align:middle line:84% don't have free parameters to account 01:35:07.040 --> 01:35:08.820 align:middle line:90% for non-linear responses. 01:35:08.820 --> 01:35:10.360 align:middle line:90% Yeah. 01:35:10.360 --> 01:35:13.360 align:middle line:84% And then productivity is the minimum 01:35:13.360 --> 01:35:18.760 align:middle line:84% of these two guys, which goes into number of pitches. 01:35:18.760 --> 01:35:22.360 align:middle line:84% And then the conversion rate, given the number 01:35:22.360 --> 01:35:26.320 align:middle line:84% of customers you are pitching, is determined 01:35:26.320 --> 01:35:30.240 align:middle line:90% by this effect of excess delay. 01:35:30.240 --> 01:35:33.640 align:middle line:84% So right now, you don't have any delay perception 01:35:33.640 --> 01:35:36.200 align:middle line:84% for the customers here, unless there 01:35:36.200 --> 01:35:39.220 align:middle line:84% is a delay built into this equation, which shouldn't be. 01:35:39.220 --> 01:35:39.720 align:middle line:90% Yeah. 01:35:39.720 --> 01:35:43.780 align:middle line:84% So I would include an explicit delay. 01:35:43.780 --> 01:35:46.440 align:middle line:84% So right now, the excess delay is coming directly 01:35:46.440 --> 01:35:51.120 align:middle line:90% from the customer service side. 01:35:51.120 --> 01:35:55.980 align:middle line:84% And that takes time to be perceived in the marketplace. 01:35:55.980 --> 01:35:59.080 align:middle line:84% So that perception delay is important to have. 01:35:59.080 --> 01:36:01.500 align:middle line:90% 01:36:01.500 --> 01:36:02.000 align:middle line:90% Yeah. 01:36:02.000 --> 01:36:06.960 align:middle line:84% And then the power function here does fine. 01:36:06.960 --> 01:36:09.800 align:middle line:84% So you have higher excess delays. 01:36:09.800 --> 01:36:12.480 align:middle line:84% You will get lower conversion rates. 01:36:12.480 --> 01:36:15.620 align:middle line:90% And that should be OK. 01:36:15.620 --> 01:36:18.900 align:middle line:84% If you had lower excess delays, which you can't here 01:36:18.900 --> 01:36:21.940 align:middle line:84% because of two different reasons we talked about, 01:36:21.940 --> 01:36:27.340 align:middle line:84% then faster conversion rates become feasible here. 01:36:27.340 --> 01:36:31.580 align:middle line:84% But you have the maximum function of these-- 01:36:31.580 --> 01:36:33.660 align:middle line:90% no, you are multiplying them. 01:36:33.660 --> 01:36:36.160 align:middle line:84% So that's a little bit of a question. 01:36:36.160 --> 01:36:40.940 align:middle line:84% That is, if you have close to 0 delivery delay-- that is, 01:36:40.940 --> 01:36:44.700 align:middle line:84% you remove those constraints that make it hard to reduce 01:36:44.700 --> 01:36:49.200 align:middle line:84% your delivery delay below current levels, 01:36:49.200 --> 01:36:52.520 align:middle line:90% if you have very fast delivery-- 01:36:52.520 --> 01:36:55.100 align:middle line:84% your conversion rate can go to infinity. 01:36:55.100 --> 01:36:58.200 align:middle line:84% It can go very high in this formulation. 01:36:58.200 --> 01:37:00.380 align:middle line:90% And that's just not realistic. 01:37:00.380 --> 01:37:02.880 align:middle line:84% So in fact, conversion rate should 01:37:02.880 --> 01:37:05.760 align:middle line:90% be conversion probability. 01:37:05.760 --> 01:37:08.860 align:middle line:84% So it is really a matter of a probability. 01:37:08.860 --> 01:37:11.720 align:middle line:84% You want it to also be dimensionless, I think. 01:37:11.720 --> 01:37:13.320 align:middle line:90% Yes, it is dimensionless. 01:37:13.320 --> 01:37:16.200 align:middle line:84% So if you call it probability, then it 01:37:16.200 --> 01:37:18.600 align:middle line:84% should never go above 1, which means 01:37:18.600 --> 01:37:22.520 align:middle line:84% you need a more nonlinear function here 01:37:22.520 --> 01:37:25.040 align:middle line:90% to account for that. 01:37:25.040 --> 01:37:26.540 align:middle line:90% But then it works fine. 01:37:26.540 --> 01:37:29.800 align:middle line:90% 01:37:29.800 --> 01:37:30.400 align:middle line:90% Good. 01:37:30.400 --> 01:37:32.220 align:middle line:90% So then we have the sales force. 01:37:32.220 --> 01:37:35.640 align:middle line:84% You have the desired sales force, adjustment time 01:37:35.640 --> 01:37:37.340 align:middle line:90% based on wages and budget. 01:37:37.340 --> 01:37:41.080 align:middle line:90% Budget comes from revenue. 01:37:41.080 --> 01:37:42.920 align:middle line:90% It's all good. 01:37:42.920 --> 01:37:47.680 align:middle line:84% If you want to have a more serious cost-benefit part, 01:37:47.680 --> 01:37:51.360 align:middle line:84% you would account for the cost of your operations. 01:37:51.360 --> 01:37:56.060 align:middle line:84% That is, your production capacity may have fixed costs. 01:37:56.060 --> 01:38:00.580 align:middle line:84% You can think about these sales to be just not product 01:38:00.580 --> 01:38:04.540 align:middle line:84% price but product profit so that this 01:38:04.540 --> 01:38:07.040 align:middle line:84% accounts for the variable cost of production. 01:38:07.040 --> 01:38:09.380 align:middle line:84% But still, you would have the fixed cost of production 01:38:09.380 --> 01:38:13.060 align:middle line:84% and the cost of building of the work in process-- 01:38:13.060 --> 01:38:18.700 align:middle line:90% 01:38:18.700 --> 01:38:19.340 align:middle line:90% yes. 01:38:19.340 --> 01:38:21.480 align:middle line:90% So where does that cost go? 01:38:21.480 --> 01:38:22.980 align:middle line:90% Yes, this is here. 01:38:22.980 --> 01:38:25.100 align:middle line:84% So you have it in the profitability. 01:38:25.100 --> 01:38:27.080 align:middle line:84% And you're getting it out of the revenue. 01:38:27.080 --> 01:38:27.720 align:middle line:90% That's OK. 01:38:27.720 --> 01:38:30.140 align:middle line:90% That's OK. 01:38:30.140 --> 01:38:30.740 align:middle line:90% Yeah. 01:38:30.740 --> 01:38:34.060 align:middle line:90% And then-- yeah, [? Sally? ?] 01:38:34.060 --> 01:38:37.900 align:middle line:84% AUDIENCE: So how do we decide with these types of models 01:38:37.900 --> 01:38:40.500 align:middle line:84% if the decision to expand the capacity 01:38:40.500 --> 01:38:46.140 align:middle line:84% should come from the delivery delay or the production capacity 01:38:46.140 --> 01:38:48.820 align:middle line:84% that we have for the profitability? 01:38:48.820 --> 01:38:52.600 align:middle line:90% 01:38:52.600 --> 01:38:55.160 align:middle line:84% HAZHIR RAHMANDAD: So typically, profitability 01:38:55.160 --> 01:38:58.160 align:middle line:90% is a constraint on expansion. 01:38:58.160 --> 01:39:03.820 align:middle line:84% So you have a desired level of production capacity. 01:39:03.820 --> 01:39:06.100 align:middle line:84% And you go through whatever formulation you're going. 01:39:06.100 --> 01:39:09.760 align:middle line:84% And then you can have another constraint 01:39:09.760 --> 01:39:12.400 align:middle line:84% that says, I don't have enough money to build, 01:39:12.400 --> 01:39:17.320 align:middle line:84% or, I can't convince the banks to loan 01:39:17.320 --> 01:39:20.680 align:middle line:84% me enough money because I'm not very profitable. 01:39:20.680 --> 01:39:22.980 align:middle line:84% So that would be the way I would capture it. 01:39:22.980 --> 01:39:25.560 align:middle line:84% That is, my total production rate 01:39:25.560 --> 01:39:29.720 align:middle line:84% could get affected by availability of financing 01:39:29.720 --> 01:39:31.420 align:middle line:90% based on a function of profits. 01:39:31.420 --> 01:39:34.575 align:middle line:90% 01:39:34.575 --> 01:39:35.200 align:middle line:90% AUDIENCE: Yeah. 01:39:35.200 --> 01:39:37.425 align:middle line:90% HAZHIR RAHMANDAD: Thank you. 01:39:37.425 --> 01:39:38.800 align:middle line:84% AUDIENCE: Could you go over again 01:39:38.800 --> 01:39:40.722 align:middle line:84% why you need that additional value? 01:39:40.722 --> 01:39:41.680 align:middle line:90% HAZHIR RAHMANDAD: Sure. 01:39:41.680 --> 01:39:44.320 align:middle line:84% AUDIENCE: Because I don't quite-- 01:39:44.320 --> 01:39:46.820 align:middle line:84% HAZHIR RAHMANDAD: So let me give you an extreme condition. 01:39:46.820 --> 01:39:50.700 align:middle line:84% So let's say initial PC, I put it as 0. 01:39:50.700 --> 01:39:52.840 align:middle line:90% OK? 01:39:52.840 --> 01:39:58.140 align:middle line:84% Now, PC completion at the beginning is a positive number. 01:39:58.140 --> 01:40:00.780 align:middle line:84% And therefore, the initial value of PC depreciation 01:40:00.780 --> 01:40:02.400 align:middle line:90% is also a positive number. 01:40:02.400 --> 01:40:02.900 align:middle line:90% OK? 01:40:02.900 --> 01:40:10.160 align:middle line:84% Because this formulation is just simply DELAY3 of PC completion. 01:40:10.160 --> 01:40:12.600 align:middle line:84% And when it is DELAY3, it starts from equilibrium, 01:40:12.600 --> 01:40:16.740 align:middle line:84% which means it starts from the initial value of PC depreciation 01:40:16.740 --> 01:40:19.380 align:middle line:90% being equal to PC completion. 01:40:19.380 --> 01:40:20.100 align:middle line:90% OK? 01:40:20.100 --> 01:40:23.660 align:middle line:84% So now I have zero production capacity, 01:40:23.660 --> 01:40:26.840 align:middle line:84% but I have positive capacity depreciation. 01:40:26.840 --> 01:40:28.635 align:middle line:90% That's just not right. 01:40:28.635 --> 01:40:29.260 align:middle line:90% AUDIENCE: Yeah. 01:40:29.260 --> 01:40:32.260 align:middle line:84% So I see why this past [INAUDIBLE] but I'm 01:40:32.260 --> 01:40:36.900 align:middle line:84% a bit confused as to why the value for the initial 01:40:36.900 --> 01:40:39.720 align:middle line:84% is PC divided by facility length. 01:40:39.720 --> 01:40:45.120 align:middle line:84% Because even if you were to have all of these stocks laid out, 01:40:45.120 --> 01:40:51.620 align:middle line:84% that outflow at the end is that figure divided by 3, right? 01:40:51.620 --> 01:40:53.520 align:middle line:90% So that's-- yeah. 01:40:53.520 --> 01:40:56.440 align:middle line:84% HAZHIR RAHMANDAD: So let's for now 01:40:56.440 --> 01:40:58.800 align:middle line:90% ignore the order of the thing. 01:40:58.800 --> 01:41:05.040 align:middle line:84% So in a steady state, the PC depreciation 01:41:05.040 --> 01:41:09.780 align:middle line:84% would be PC, or initial PC, divided by facility life. 01:41:09.780 --> 01:41:10.280 align:middle line:90% Correct? 01:41:10.280 --> 01:41:13.160 align:middle line:90% 01:41:13.160 --> 01:41:14.960 align:middle line:90% That's Little's law. 01:41:14.960 --> 01:41:18.600 align:middle line:84% So stock divided by the time constant 01:41:18.600 --> 01:41:22.200 align:middle line:90% is the flow in equilibrium. 01:41:22.200 --> 01:41:24.040 align:middle line:84% And that's all we are using here. 01:41:24.040 --> 01:41:26.800 align:middle line:84% That is, the initial value of the outflow 01:41:26.800 --> 01:41:29.400 align:middle line:84% should be the equilibrium value so 01:41:29.400 --> 01:41:34.320 align:middle line:84% that then that's how Vensim will operationalize it 01:41:34.320 --> 01:41:38.320 align:middle line:84% inside its hidden three stocks, or 10 stocks, or one stock 01:41:38.320 --> 01:41:40.840 align:middle line:84% depending on what the order you're using-- 01:41:40.840 --> 01:41:44.720 align:middle line:84% will be that total divided by the order of the stocks. 01:41:44.720 --> 01:41:47.400 align:middle line:84% You don't need to worry about that order of the stocks. 01:41:47.400 --> 01:41:48.920 align:middle line:90% Vensim will take care of that. 01:41:48.920 --> 01:41:50.780 align:middle line:84% The sum of those stocks will always 01:41:50.780 --> 01:41:54.400 align:middle line:84% be what you get in the initial PC 01:41:54.400 --> 01:41:58.820 align:middle line:84% once you make the initial outflow 01:41:58.820 --> 01:42:01.285 align:middle line:90% to be what it needs to be here. 01:42:01.285 --> 01:42:02.660 align:middle line:84% AUDIENCE: So Vensim actually uses 01:42:02.660 --> 01:42:05.148 align:middle line:84% that outflow to calculate the value of all of those stocks? 01:42:05.148 --> 01:42:06.940 align:middle line:84% HAZHIR RAHMANDAD: Yes, that initial outflow 01:42:06.940 --> 01:42:10.425 align:middle line:84% is what it uses to calculate all of those stocks. 01:42:10.425 --> 01:42:11.300 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:42:11.300 --> 01:42:12.967 align:middle line:84% HAZHIR RAHMANDAD: Yeah, and it does that 01:42:12.967 --> 01:42:15.380 align:middle line:84% by equally distributing it across all 01:42:15.380 --> 01:42:17.440 align:middle line:90% of the initial stocks. 01:42:17.440 --> 01:42:17.940 align:middle line:90% OK. 01:42:17.940 --> 01:42:23.780 align:middle line:84% So I think the basics are all in place 01:42:23.780 --> 01:42:29.580 align:middle line:84% with all of the minor caveats that we talked about here. 01:42:29.580 --> 01:42:32.220 align:middle line:84% Who likes us to look at their model? 01:42:32.220 --> 01:42:35.680 align:middle line:84% Let's kind of open it up for volunteers. 01:42:35.680 --> 01:42:38.380 align:middle line:90% 01:42:38.380 --> 01:42:39.080 align:middle line:90% Sure. 01:42:39.080 --> 01:42:40.080 align:middle line:90% AUDIENCE: It's not good. 01:42:40.080 --> 01:42:41.163 align:middle line:90% HAZHIR RAHMANDAD: It's OK. 01:42:41.163 --> 01:42:41.760 align:middle line:90% We are not-- 01:42:41.760 --> 01:42:43.200 align:middle line:84% AUDIENCE: [INAUDIBLE] equilibrium. 01:42:43.200 --> 01:42:46.430 align:middle line:84% There's a lot of really bad areas, but-- 01:42:46.430 --> 01:42:48.180 align:middle line:84% HAZHIR RAHMANDAD: That makes it more fun-- 01:42:48.180 --> 01:42:49.976 align:middle line:90% [LAUGHTER] 01:42:49.976 --> 01:42:50.840 align:middle line:90% 01:42:50.840 --> 01:42:51.700 align:middle line:90% --to critique. 01:42:51.700 --> 01:42:54.720 align:middle line:90% 01:42:54.720 --> 01:42:56.300 align:middle line:90% Not fun, more educational. 01:42:56.300 --> 01:42:58.720 align:middle line:90% OK. 01:42:58.720 --> 01:43:01.055 align:middle line:84% So you have the facilities, factories. 01:43:01.055 --> 01:43:02.180 align:middle line:90% You have first order delay. 01:43:02.180 --> 01:43:03.640 align:middle line:90% We talked about that. 01:43:03.640 --> 01:43:06.000 align:middle line:84% And then you have the construction rate 01:43:06.000 --> 01:43:09.040 align:middle line:84% based on the factory adjustment, which is comparing 01:43:09.040 --> 01:43:11.020 align:middle line:90% the authorized or desired. 01:43:11.020 --> 01:43:14.720 align:middle line:84% And the desired is coming from something we'll get to-- 01:43:14.720 --> 01:43:18.360 align:middle line:84% versus the current level with an adjustment time. 01:43:18.360 --> 01:43:21.240 align:middle line:84% And then you add the factory construction rate 01:43:21.240 --> 01:43:24.040 align:middle line:84% so that you don't have the steady state 01:43:24.040 --> 01:43:25.740 align:middle line:90% problem that we talked about. 01:43:25.740 --> 01:43:27.840 align:middle line:90% That's all good. 01:43:27.840 --> 01:43:31.720 align:middle line:84% Again, this is ignoring the work in process inventory 01:43:31.720 --> 01:43:33.020 align:middle line:90% of factories. 01:43:33.020 --> 01:43:36.847 align:middle line:84% And the units for factories are in-- 01:43:36.847 --> 01:43:37.680 align:middle line:90% AUDIENCE: Factories. 01:43:37.680 --> 01:43:38.260 align:middle line:90% HAZHIR RAHMANDAD: --factories. 01:43:38.260 --> 01:43:38.760 align:middle line:90% OK. 01:43:38.760 --> 01:43:43.100 align:middle line:90% 01:43:43.100 --> 01:43:47.420 align:middle line:84% So typically better to make it in the units of production 01:43:47.420 --> 01:43:51.700 align:middle line:84% capacity so that it is easier to think 01:43:51.700 --> 01:43:55.560 align:middle line:84% without worrying about the parameter that converts them, 01:43:55.560 --> 01:43:56.380 align:middle line:90% but it's all right. 01:43:56.380 --> 01:43:57.005 align:middle line:90% AUDIENCE: Yeah. 01:43:57.005 --> 01:44:02.580 align:middle line:84% [INAUDIBLE] obviously discrete amounts. 01:44:02.580 --> 01:44:05.060 align:middle line:84% So then I did integer factory, which is wrong now 01:44:05.060 --> 01:44:07.140 align:middle line:90% that I think about it more. 01:44:07.140 --> 01:44:09.200 align:middle line:84% The idea was, I was like, oh, well, 01:44:09.200 --> 01:44:12.000 align:middle line:84% when I'm building capacity, if I have 1.5 factories, 01:44:12.000 --> 01:44:13.780 align:middle line:90% I only have 1 factory. 01:44:13.780 --> 01:44:15.340 align:middle line:84% Then I wasn't taking into account, 01:44:15.340 --> 01:44:17.700 align:middle line:84% as the factories depreciate, I still 01:44:17.700 --> 01:44:20.702 align:middle line:84% have the [? local ?] factory [INAUDIBLE] integer down to 0. 01:44:20.702 --> 01:44:21.660 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:44:21.660 --> 01:44:25.480 align:middle line:84% So integers are hard to deal with. 01:44:25.480 --> 01:44:28.018 align:middle line:84% Don't get into the integer stuff in Vensim. 01:44:28.018 --> 01:44:29.560 align:middle line:84% I mean, Vensim is not the right tool. 01:44:29.560 --> 01:44:30.280 align:middle line:90% You can do it. 01:44:30.280 --> 01:44:32.180 align:middle line:90% It's just kind of a headache. 01:44:32.180 --> 01:44:36.680 align:middle line:84% So I wouldn't touch integer constraints because then you 01:44:36.680 --> 01:44:38.260 align:middle line:90% have to be really consistent. 01:44:38.260 --> 01:44:40.060 align:middle line:84% Like, the inflow needs to be integer. 01:44:40.060 --> 01:44:41.400 align:middle line:90% The outflow needs to be integer. 01:44:41.400 --> 01:44:42.420 align:middle line:90% And it's not worth it. 01:44:42.420 --> 01:44:44.800 align:middle line:90% So don't worry about that. 01:44:44.800 --> 01:44:48.020 align:middle line:84% But let's for now put that aside. 01:44:48.020 --> 01:44:50.360 align:middle line:84% So now the authorized number of factories-- 01:44:50.360 --> 01:44:52.060 align:middle line:90% you're starting at equilibrium. 01:44:52.060 --> 01:44:55.040 align:middle line:84% So the initial number is the same for both of them. 01:44:55.040 --> 01:45:00.960 align:middle line:84% That authorization is coming from some pressure to expand 01:45:00.960 --> 01:45:03.580 align:middle line:90% and some time to authorize. 01:45:03.580 --> 01:45:04.080 align:middle line:90% OK. 01:45:04.080 --> 01:45:07.440 align:middle line:90% I'm not following this yet. 01:45:07.440 --> 01:45:10.320 align:middle line:84% AUDIENCE: So the idea there was like, 01:45:10.320 --> 01:45:12.840 align:middle line:90% here's some delivery delay. 01:45:12.840 --> 01:45:18.030 align:middle line:84% They have the promised delivery delay and then some-- 01:45:18.030 --> 01:45:19.307 align:middle line:90% the reference delivery delay-- 01:45:19.307 --> 01:45:21.140 align:middle line:84% I don't actually know if it makes any sense, 01:45:21.140 --> 01:45:24.400 align:middle line:84% but the idea of the company itself 01:45:24.400 --> 01:45:26.200 align:middle line:84% is OK with the delivery delay actually 01:45:26.200 --> 01:45:29.200 align:middle line:84% getting this much higher than the promised delivery delay. 01:45:29.200 --> 01:45:36.540 align:middle line:84% And so then there's going to be like some magical pressure 01:45:36.540 --> 01:45:37.948 align:middle line:90% variables to come out of that. 01:45:37.948 --> 01:45:40.240 align:middle line:84% HAZHIR RAHMANDAD: But how do you get the units for now? 01:45:40.240 --> 01:45:41.480 align:middle line:90% So let's focus on that. 01:45:41.480 --> 01:45:44.120 align:middle line:84% So the authorized number of factories in factory-- 01:45:44.120 --> 01:45:49.060 align:middle line:84% so changing authorization is factory per month or year, but-- 01:45:49.060 --> 01:45:49.960 align:middle line:90% per month. 01:45:49.960 --> 01:45:53.020 align:middle line:84% And this one is pressure to expand 01:45:53.020 --> 01:46:00.960 align:middle line:84% capacity multiplied by pressure to factory conversion. 01:46:00.960 --> 01:46:03.220 align:middle line:84% AUDIENCE: The idea was just that there's 01:46:03.220 --> 01:46:06.460 align:middle line:84% some number that would decide how many factories 01:46:06.460 --> 01:46:09.740 align:middle line:84% they want to [INAUDIBLE] but the actual pressure number 01:46:09.740 --> 01:46:10.920 align:middle line:90% doesn't make any sense. 01:46:10.920 --> 01:46:14.180 align:middle line:84% This isn't based in any real idea. 01:46:14.180 --> 01:46:18.060 align:middle line:84% My idea was just that, oh, some delivery delay is happening. 01:46:18.060 --> 01:46:22.140 align:middle line:84% And it's somewhere within or without the reference of how OK 01:46:22.140 --> 01:46:24.300 align:middle line:84% they are with the delivery delay being, 01:46:24.300 --> 01:46:28.140 align:middle line:84% and that that whether it would be actually through a table 01:46:28.140 --> 01:46:30.260 align:middle line:84% or through something else would be converted 01:46:30.260 --> 01:46:32.995 align:middle line:84% from that pressure into how many factories they want to build. 01:46:32.995 --> 01:46:34.620 align:middle line:84% And then from there, you can actually-- 01:46:34.620 --> 01:46:36.870 align:middle line:84% HAZHIR RAHMANDAD: So the thing that you are doing here 01:46:36.870 --> 01:46:38.700 align:middle line:90% is a little bit tough. 01:46:38.700 --> 01:46:42.000 align:middle line:84% Because you're trying to start with the right number 01:46:42.000 --> 01:46:46.400 align:middle line:84% of factories and then just expand that, accumulate on that 01:46:46.400 --> 01:46:47.885 align:middle line:90% or potentially even reduce-- 01:46:47.885 --> 01:46:49.260 align:middle line:84% I haven't looked at the pressure, 01:46:49.260 --> 01:46:51.840 align:middle line:84% whether it goes negative or not-- 01:46:51.840 --> 01:46:56.880 align:middle line:84% without any response or any feedback about how many 01:46:56.880 --> 01:46:59.080 align:middle line:90% factories already you have. 01:46:59.080 --> 01:47:02.920 align:middle line:84% Because there is nowhere here a feedback from the current number 01:47:02.920 --> 01:47:03.700 align:middle line:90% of factories. 01:47:03.700 --> 01:47:06.480 align:middle line:84% So it is not anchored in the current state. 01:47:06.480 --> 01:47:15.000 align:middle line:84% It is only accumulating over the starting point of the factories. 01:47:15.000 --> 01:47:16.920 align:middle line:90% You see what I'm saying? 01:47:16.920 --> 01:47:18.210 align:middle line:90% So-- 01:47:18.210 --> 01:47:20.920 align:middle line:84% AUDIENCE: Oh, when the factories depreciate, [INAUDIBLE] 01:47:20.920 --> 01:47:23.260 align:middle line:84% HAZHIR RAHMANDAD: It is not taking that into account. 01:47:23.260 --> 01:47:27.540 align:middle line:84% If the need for factories goes down significantly-- 01:47:27.540 --> 01:47:30.900 align:middle line:90% it is basically kind of-- 01:47:30.900 --> 01:47:35.120 align:middle line:84% does it call it dead something navigation? 01:47:35.120 --> 01:47:36.618 align:middle line:84% That is, you start from a point-- 01:47:36.618 --> 01:47:37.660 align:middle line:90% AUDIENCE: Dead reckoning. 01:47:37.660 --> 01:47:38.480 align:middle line:84% HAZHIR RAHMANDAD: Dead reckoning. 01:47:38.480 --> 01:47:39.647 align:middle line:90% It's kind of dead reckoning. 01:47:39.647 --> 01:47:43.280 align:middle line:84% You start from a point, which is the current level of factories. 01:47:43.280 --> 01:47:44.140 align:middle line:90% That's OK. 01:47:44.140 --> 01:47:46.140 align:middle line:84% And then you just accumulate into that 01:47:46.140 --> 01:47:47.520 align:middle line:90% based on some pressure. 01:47:47.520 --> 01:47:49.620 align:middle line:84% But you're never getting feedback 01:47:49.620 --> 01:47:53.100 align:middle line:84% whether this number is close to some underlying desired 01:47:53.100 --> 01:47:58.460 align:middle line:84% value that should be driving it or how far you 01:47:58.460 --> 01:48:00.040 align:middle line:90% are from the current factory. 01:48:00.040 --> 01:48:03.240 align:middle line:90% None of that comes into play. 01:48:03.240 --> 01:48:03.740 align:middle line:90% So-- 01:48:03.740 --> 01:48:06.100 align:middle line:84% AUDIENCE: But isn't the desired amount this-- 01:48:06.100 --> 01:48:07.980 align:middle line:84% HAZHIR RAHMANDAD: So the desired amount 01:48:07.980 --> 01:48:09.920 align:middle line:84% is going to change the number of factories. 01:48:09.920 --> 01:48:14.340 align:middle line:84% So there will be connection between the number 01:48:14.340 --> 01:48:17.520 align:middle line:84% of factories that you have and this authorized number. 01:48:17.520 --> 01:48:20.280 align:middle line:84% So those two will remain tied to each other. 01:48:20.280 --> 01:48:23.500 align:middle line:84% But this authorized number is not tied to anything. 01:48:23.500 --> 01:48:28.520 align:middle line:84% It is only being accumulated using some external pressure. 01:48:28.520 --> 01:48:32.320 align:middle line:84% And therefore, it is a little bit tough 01:48:32.320 --> 01:48:35.140 align:middle line:90% to keep it well grounded. 01:48:35.140 --> 01:48:38.640 align:middle line:84% That's the core challenge you're facing. 01:48:38.640 --> 01:48:41.640 align:middle line:84% Now, it may work just fine in the sense 01:48:41.640 --> 01:48:44.760 align:middle line:84% that, if the signal from pressure to expand 01:48:44.760 --> 01:48:48.440 align:middle line:84% can go both positive and negative, if it overshoots, 01:48:48.440 --> 01:48:49.580 align:middle line:90% it brings it down. 01:48:49.580 --> 01:48:52.760 align:middle line:84% If it goes down, it brings it up. 01:48:52.760 --> 01:48:56.240 align:middle line:84% But typical formulations will make this kind 01:48:56.240 --> 01:48:57.860 align:middle line:90% of a fractional effect. 01:48:57.860 --> 01:49:01.340 align:middle line:84% That is, OK, if I have a number of factories, 01:49:01.340 --> 01:49:04.620 align:middle line:84% I will multiply that by this expansion factor 01:49:04.620 --> 01:49:07.720 align:middle line:84% so that I can grow fractionally, so that if I start 01:49:07.720 --> 01:49:11.120 align:middle line:84% with one factory and my rate of change 01:49:11.120 --> 01:49:12.800 align:middle line:84% in this change in authorization is 01:49:12.800 --> 01:49:17.740 align:middle line:84% in the order of 0.5 factory per year, 10 years from now, 01:49:17.740 --> 01:49:19.380 align:middle line:90% now I have 100 factories. 01:49:19.380 --> 01:49:23.080 align:middle line:84% It doesn't make sense to change my factory by 0.5 per year. 01:49:23.080 --> 01:49:26.560 align:middle line:84% I now want to actually be able to change it by 50 per year 01:49:26.560 --> 01:49:28.020 align:middle line:90% maybe. 01:49:28.020 --> 01:49:31.460 align:middle line:84% That is not possible in this formulation 01:49:31.460 --> 01:49:34.420 align:middle line:84% because the scale of these three things coming together 01:49:34.420 --> 01:49:39.520 align:middle line:84% is always constant based on some dimensionless thing from here. 01:49:39.520 --> 01:49:40.830 align:middle line:90% OK. 01:49:40.830 --> 01:49:42.497 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:49:42.497 --> 01:49:44.580 align:middle line:84% HAZHIR RAHMANDAD: Could you speak up a little bit? 01:49:44.580 --> 01:49:44.686 align:middle line:90% AUDIENCE: Sorry. 01:49:44.686 --> 01:49:46.980 align:middle line:84% So isn't the pressure to expand capacity exactly 01:49:46.980 --> 01:49:48.480 align:middle line:90% the variable we're looking at? 01:49:48.480 --> 01:49:51.672 align:middle line:84% Doesn't that depend on delivery delay [INAUDIBLE]? 01:49:51.672 --> 01:49:52.880 align:middle line:90% HAZHIR RAHMANDAD: Sure, sure. 01:49:52.880 --> 01:49:54.000 align:middle line:90% All of that is correct. 01:49:54.000 --> 01:49:57.055 align:middle line:84% And therefore, there will be a feedback, but-- 01:49:57.055 --> 01:49:59.180 align:middle line:84% AUDIENCE: There will be feedback [INAUDIBLE] right? 01:49:59.180 --> 01:50:02.240 align:middle line:84% HAZHIR RAHMANDAD: It is not proportional-- 01:50:02.240 --> 01:50:05.080 align:middle line:84% so the adjustment is not proportional. 01:50:05.080 --> 01:50:08.300 align:middle line:84% And therefore, the feedback remains very linear-ish. 01:50:08.300 --> 01:50:10.760 align:middle line:84% So if the scale of operation is not changing, 01:50:10.760 --> 01:50:13.200 align:middle line:90% this could work OK. 01:50:13.200 --> 01:50:15.520 align:middle line:84% But it is blind to the scale of operation. 01:50:15.520 --> 01:50:18.500 align:middle line:84% So if you end up with a larger scale because you are dealing 01:50:18.500 --> 01:50:20.940 align:middle line:84% with, for example, an exponential growth, 01:50:20.940 --> 01:50:23.300 align:middle line:84% it just cannot catch up with that. 01:50:23.300 --> 01:50:27.680 align:middle line:84% That's the challenge, especially with an exponentially growing 01:50:27.680 --> 01:50:29.520 align:middle line:90% demand, which is the case here. 01:50:29.520 --> 01:50:31.680 align:middle line:84% This becomes a very tough bottleneck. 01:50:31.680 --> 01:50:34.220 align:middle line:84% It just doesn't allow the capacity to catch up. 01:50:34.220 --> 01:50:36.280 align:middle line:84% And then you fall behind and behind. 01:50:36.280 --> 01:50:39.160 align:middle line:84% Now, you don't fall super behind because your delivery delay now 01:50:39.160 --> 01:50:42.060 align:middle line:84% goes up so much that you can't grow exponentially. 01:50:42.060 --> 01:50:46.180 align:middle line:84% That creates a kind of a linear growth type of a dynamic. 01:50:46.180 --> 01:50:48.520 align:middle line:84% That would be my guess, that this kind of creates that. 01:50:48.520 --> 01:50:50.200 align:middle line:90% We can simulate it. 01:50:50.200 --> 01:50:54.195 align:middle line:84% AUDIENCE: [INAUDIBLE] so what's the lesson? 01:50:54.195 --> 01:50:56.320 align:middle line:84% The lesson is you want comparable units when you're 01:50:56.320 --> 01:50:57.680 align:middle line:90% doing this type of [INAUDIBLE]? 01:50:57.680 --> 01:50:59.520 align:middle line:84% HAZHIR RAHMANDAD: So the real lesson 01:50:59.520 --> 01:51:03.760 align:middle line:84% is, if you are creating a desired level for anything, 01:51:03.760 --> 01:51:08.520 align:middle line:84% you want it to be anchored to something current, 01:51:08.520 --> 01:51:10.440 align:middle line:84% whether it is the current capacity that you 01:51:10.440 --> 01:51:12.640 align:middle line:84% want to anchor to or current demand that you 01:51:12.640 --> 01:51:13.780 align:middle line:90% want to anchor to. 01:51:13.780 --> 01:51:16.240 align:middle line:84% But anchor that desire to something current. 01:51:16.240 --> 01:51:20.320 align:middle line:84% Otherwise, expecting the dead process 01:51:20.320 --> 01:51:22.540 align:middle line:84% to get that thing to the right place 01:51:22.540 --> 01:51:25.380 align:middle line:84% is too much to ask from your models. 01:51:25.380 --> 01:51:28.200 align:middle line:90% That's from your formulations. 01:51:28.200 --> 01:51:31.900 align:middle line:84% Because the scope of these adjustments 01:51:31.900 --> 01:51:35.560 align:middle line:84% are not tuned for different scales of operation. 01:51:35.560 --> 01:51:40.340 align:middle line:84% You just multiply everything in this model by a factor of 100. 01:51:40.340 --> 01:51:43.420 align:middle line:84% The model will not scale because this part of the formulation 01:51:43.420 --> 01:51:45.920 align:middle line:84% for changing authorization-- nothing changes there. 01:51:45.920 --> 01:51:50.660 align:middle line:84% It's still in the units of half a unit or one unit per month 01:51:50.660 --> 01:51:52.680 align:middle line:84% no matter what the scale of operation is. 01:51:52.680 --> 01:51:56.360 align:middle line:90% That's kind of the challenge. 01:51:56.360 --> 01:51:56.860 align:middle line:90% OK. 01:51:56.860 --> 01:51:59.340 align:middle line:84% Delivery delay-- fairly standard. 01:51:59.340 --> 01:52:02.700 align:middle line:84% We have the max completion versus the potential completion 01:52:02.700 --> 01:52:04.600 align:middle line:90% coming from your factories. 01:52:04.600 --> 01:52:07.580 align:middle line:84% Let's just skip the integer stuff here. 01:52:07.580 --> 01:52:09.780 align:middle line:90% Then you have the purchase rate. 01:52:09.780 --> 01:52:13.825 align:middle line:84% And then you-- purchase to tasks. 01:52:13.825 --> 01:52:15.700 align:middle line:84% AUDIENCE: Well, it's just assuming that the-- 01:52:15.700 --> 01:52:17.560 align:middle line:84% per customer they're buying [INAUDIBLE]. 01:52:17.560 --> 01:52:18.977 align:middle line:84% HAZHIR RAHMANDAD: Yeah, that's OK. 01:52:18.977 --> 01:52:22.480 align:middle line:84% You're just taking care of the units somehow. 01:52:22.480 --> 01:52:25.460 align:middle line:84% Then you have the sales force and the customers. 01:52:25.460 --> 01:52:26.580 align:middle line:90% So you have customers. 01:52:26.580 --> 01:52:27.740 align:middle line:90% You have purchase rate. 01:52:27.740 --> 01:52:29.160 align:middle line:90% You have depreciation. 01:52:29.160 --> 01:52:32.380 align:middle line:84% This purchase rate is coming from reputation. 01:52:32.380 --> 01:52:34.840 align:middle line:84% So you are kind of doing similar to a few others 01:52:34.840 --> 01:52:37.760 align:middle line:84% that we saw, that there is a process to get customers 01:52:37.760 --> 01:52:41.220 align:middle line:90% into your orbit of customers. 01:52:41.220 --> 01:52:45.320 align:middle line:84% And then once they are in, they will continue to buy from you. 01:52:45.320 --> 01:52:48.220 align:middle line:84% As soon as their product life ends, 01:52:48.220 --> 01:52:51.500 align:middle line:84% they go into this interested customers with a delay. 01:52:51.500 --> 01:52:56.780 align:middle line:84% They will purchase, but that is bounded by your reputation. 01:52:56.780 --> 01:52:59.820 align:middle line:90% So this should be OK. 01:52:59.820 --> 01:53:03.960 align:middle line:90% 01:53:03.960 --> 01:53:05.880 align:middle line:84% It's a little bit tricky kind of how 01:53:05.880 --> 01:53:07.740 align:middle line:90% the reputation is working here. 01:53:07.740 --> 01:53:13.503 align:middle line:84% So if you have zero reputation, yes, this shuts off. 01:53:13.503 --> 01:53:14.920 align:middle line:84% AUDIENCE: [INAUDIBLE] no interest. 01:53:14.920 --> 01:53:15.760 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:53:15.760 --> 01:53:17.620 align:middle line:84% AUDIENCE: But the formulation was incorrect. 01:53:17.620 --> 01:53:19.302 align:middle line:90% But that was the idea. 01:53:19.302 --> 01:53:20.260 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:53:20.260 --> 01:53:23.420 align:middle line:84% So the formulation-- the bigger part that is missing here 01:53:23.420 --> 01:53:26.540 align:middle line:84% is the interested customers going into it. 01:53:26.540 --> 01:53:31.380 align:middle line:84% So the purchase rate should not be 01:53:31.380 --> 01:53:33.280 align:middle line:84% independent of interested customers. 01:53:33.280 --> 01:53:36.735 align:middle line:84% This is interested customers divided by time to purchase. 01:53:36.735 --> 01:53:38.360 align:middle line:84% And it should be a short time constant. 01:53:38.360 --> 01:53:42.760 align:middle line:84% Then you can adjust around that with the reputation. 01:53:42.760 --> 01:53:43.260 align:middle line:90% OK. 01:53:43.260 --> 01:53:44.900 align:middle line:84% And then gaining interest-- there 01:53:44.900 --> 01:53:49.900 align:middle line:90% is a max customers visited. 01:53:49.900 --> 01:53:54.940 align:middle line:84% That's kind of the same idea that we had in other models. 01:53:54.940 --> 01:53:58.020 align:middle line:84% AUDIENCE: This is where I got confused, one, because 01:53:58.020 --> 01:54:00.120 align:middle line:90% of the revenue constraint. 01:54:00.120 --> 01:54:04.573 align:middle line:84% But then, secondly, I think it wasn't something I needed to do. 01:54:04.573 --> 01:54:06.740 align:middle line:84% But it felt like there shouldn't be more sales force 01:54:06.740 --> 01:54:08.845 align:middle line:84% people than there are [INAUDIBLE] customers. 01:54:08.845 --> 01:54:10.220 align:middle line:84% HAZHIR RAHMANDAD: So these things 01:54:10.220 --> 01:54:12.080 align:middle line:90% depend on that conversion rate. 01:54:12.080 --> 01:54:15.900 align:middle line:84% So you actually could have 10 different sales force people 01:54:15.900 --> 01:54:17.300 align:middle line:90% working on a single customer. 01:54:17.300 --> 01:54:19.160 align:middle line:84% If your single customer is the US government 01:54:19.160 --> 01:54:21.480 align:middle line:84% and you are, I don't know, Boeing 01:54:21.480 --> 01:54:27.380 align:middle line:84% trying to sell a single missile to the government, 01:54:27.380 --> 01:54:30.060 align:middle line:84% you may have a whole army trying to sell that single unit. 01:54:30.060 --> 01:54:33.780 align:middle line:84% So that is just a productivity factor. 01:54:33.780 --> 01:54:36.140 align:middle line:90% 01:54:36.140 --> 01:54:36.640 align:middle line:90% OK. 01:54:36.640 --> 01:54:39.220 align:middle line:84% So let's look at these sales force formulations. 01:54:39.220 --> 01:54:42.080 align:middle line:90% So attrition is fine. 01:54:42.080 --> 01:54:46.400 align:middle line:84% You have hiring that comes from the adjustment of sales force. 01:54:46.400 --> 01:54:49.140 align:middle line:84% That adjustment is based on the desired sales force. 01:54:49.140 --> 01:54:50.980 align:middle line:90% So what's missing here? 01:54:50.980 --> 01:54:51.980 align:middle line:90% AUDIENCE: The attrition. 01:54:51.980 --> 01:54:53.480 align:middle line:84% HAZHIR RAHMANDAD: So attrition needs 01:54:53.480 --> 01:54:55.700 align:middle line:84% to be accounted for for the hiring rate. 01:54:55.700 --> 01:54:57.440 align:middle line:84% And then you would have basically 01:54:57.440 --> 01:55:02.400 align:middle line:84% attrition plus adjustment max of 0 and that thing. 01:55:02.400 --> 01:55:04.800 align:middle line:84% Otherwise, you have a steady state error. 01:55:04.800 --> 01:55:07.160 align:middle line:84% Once you have your sales force, you 01:55:07.160 --> 01:55:12.300 align:middle line:84% have max customers visited based on the-- 01:55:12.300 --> 01:55:14.300 align:middle line:84% AUDIENCE: This is a little bit confusing for me. 01:55:14.300 --> 01:55:15.340 align:middle line:90% AUDIENCE: Yeah. 01:55:15.340 --> 01:55:19.200 align:middle line:84% I just started getting negative customers at one point. 01:55:19.200 --> 01:55:20.860 align:middle line:84% And I was trying to figure out why. 01:55:20.860 --> 01:55:23.980 align:middle line:84% I thought that maybe it was that I was somehow 01:55:23.980 --> 01:55:25.582 align:middle line:90% visiting more customers. 01:55:25.582 --> 01:55:26.540 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 01:55:26.540 --> 01:55:28.400 align:middle line:84% So basically, you need the minimum. 01:55:28.400 --> 01:55:29.500 align:middle line:90% This is fine. 01:55:29.500 --> 01:55:32.360 align:middle line:84% But then the max customers visited 01:55:32.360 --> 01:55:34.900 align:middle line:84% and actual customers visited-- the terminology 01:55:34.900 --> 01:55:36.320 align:middle line:90% is a bit confusing. 01:55:36.320 --> 01:55:39.140 align:middle line:84% So this is kind of potential customers visited 01:55:39.140 --> 01:55:41.740 align:middle line:90% based on your sales force size. 01:55:41.740 --> 01:55:43.660 align:middle line:84% And this is the max customers visited 01:55:43.660 --> 01:55:46.040 align:middle line:84% based on the potential interested customers. 01:55:46.040 --> 01:55:47.620 align:middle line:84% And the equation for max customers 01:55:47.620 --> 01:55:50.500 align:middle line:84% visited should simply be potential interested customers 01:55:50.500 --> 01:55:52.740 align:middle line:84% divided by minimum time to visit customers 01:55:52.740 --> 01:55:53.920 align:middle line:90% or something like that. 01:55:53.920 --> 01:55:55.520 align:middle line:90% So that's fairly standard. 01:55:55.520 --> 01:55:57.900 align:middle line:90% Then you can use the-- 01:55:57.900 --> 01:56:01.660 align:middle line:84% and back to the same point we had with Alban 01:56:01.660 --> 01:56:04.100 align:middle line:84% and a couple other, this doesn't give you 01:56:04.100 --> 01:56:10.860 align:middle line:84% flexibility to have a non-linear shift in how 01:56:10.860 --> 01:56:15.100 align:middle line:90% effective your salespeople are. 01:56:15.100 --> 01:56:16.360 align:middle line:90% OK. 01:56:16.360 --> 01:56:16.900 align:middle line:90% Good. 01:56:16.900 --> 01:56:20.360 align:middle line:84% I think the rest are fairly straightforward. 01:56:20.360 --> 01:56:26.040 align:middle line:84% And we have sales force wage and total fraction of the revenue. 01:56:26.040 --> 01:56:29.680 align:middle line:84% AUDIENCE: So it seems basically everyone else 01:56:29.680 --> 01:56:33.000 align:middle line:84% made the same kind of idea in terms of the sales force, 01:56:33.000 --> 01:56:35.680 align:middle line:84% that you have some budget coming from the current money you're 01:56:35.680 --> 01:56:36.400 align:middle line:90% making. 01:56:36.400 --> 01:56:38.840 align:middle line:84% But then you would adjust your sales force with the delay. 01:56:38.840 --> 01:56:41.423 align:middle line:84% And I just felt kind of weird about making that simplification 01:56:41.423 --> 01:56:44.320 align:middle line:84% because it's possible that suddenly revenue drops, 01:56:44.320 --> 01:56:47.240 align:middle line:84% but your sales force doesn't immediately 01:56:47.240 --> 01:56:48.680 align:middle line:90% change because there's a delay. 01:56:48.680 --> 01:56:50.240 align:middle line:84% And so there's a little time period 01:56:50.240 --> 01:56:52.400 align:middle line:84% in which you're paying your sales force with money 01:56:52.400 --> 01:56:53.560 align:middle line:90% you don't have. 01:56:53.560 --> 01:56:56.020 align:middle line:84% And I guess you could borrow, right? 01:56:56.020 --> 01:56:57.937 align:middle line:84% But then that has other financial implications 01:56:57.937 --> 01:56:59.645 align:middle line:84% with interest rates and blah, blah, blah. 01:56:59.645 --> 01:57:00.320 align:middle line:90% So I don't know. 01:57:00.320 --> 01:57:02.600 align:middle line:84% It's just the sort of thing where [INAUDIBLE] 01:57:02.600 --> 01:57:03.850 align:middle line:90% HAZHIR RAHMANDAD: That's fair. 01:57:03.850 --> 01:57:06.400 align:middle line:84% So basically, the assumption there 01:57:06.400 --> 01:57:09.160 align:middle line:84% is that you're borrowing to pay your sales force. 01:57:09.160 --> 01:57:12.820 align:middle line:84% And if you're doing your cost-benefit correctly, 01:57:12.820 --> 01:57:17.620 align:middle line:84% that borrowing will show up in negative cash flow and reduction 01:57:17.620 --> 01:57:21.720 align:middle line:84% of your cumulative profit or whatever you call that. 01:57:21.720 --> 01:57:22.980 align:middle line:90% So it's not wrong. 01:57:22.980 --> 01:57:24.280 align:middle line:90% It works fine. 01:57:24.280 --> 01:57:27.180 align:middle line:84% If you want to capture that more explicitly, 01:57:27.180 --> 01:57:30.300 align:middle line:84% you would have a stock of cash or something 01:57:30.300 --> 01:57:32.980 align:middle line:84% that is accounting for money in your bank account 01:57:32.980 --> 01:57:34.260 align:middle line:90% that you can pay. 01:57:34.260 --> 01:57:36.100 align:middle line:84% And if you don't have money to pay, 01:57:36.100 --> 01:57:38.180 align:middle line:84% you're not paying your sales force. 01:57:38.180 --> 01:57:40.840 align:middle line:84% That basically should automatically get them out. 01:57:40.840 --> 01:57:43.635 align:middle line:84% Or maybe if they are US government, federal employees, 01:57:43.635 --> 01:57:45.260 align:middle line:84% maybe they stick around and try to work 01:57:45.260 --> 01:57:46.320 align:middle line:90% without out or something. 01:57:46.320 --> 01:57:51.580 align:middle line:84% So we had this executive student who 01:57:51.580 --> 01:57:55.940 align:middle line:84% was telling us this story about payment system in Canada, 01:57:55.940 --> 01:57:57.093 align:middle line:90% that they had the project. 01:57:57.093 --> 01:57:59.260 align:middle line:84% And this was a project that was to be, I don't know, 01:57:59.260 --> 01:58:01.540 align:middle line:90% three years, $300 million. 01:58:01.540 --> 01:58:02.820 align:middle line:90% And now they are-- 01:58:02.820 --> 01:58:06.740 align:middle line:84% for re-vitalizing their payment system 01:58:06.740 --> 01:58:11.080 align:middle line:84% for the Canadian federal employees. 01:58:11.080 --> 01:58:13.160 align:middle line:84% Everybody who gets paid basically 01:58:13.160 --> 01:58:14.800 align:middle line:90% goes through the system. 01:58:14.800 --> 01:58:18.880 align:middle line:84% And then she said, now we are 10 years later and $3 billion 01:58:18.880 --> 01:58:22.020 align:middle line:84% of cost, and this project is not yet finished. 01:58:22.020 --> 01:58:25.360 align:middle line:84% And not only that, it has been operational. 01:58:25.360 --> 01:58:27.080 align:middle line:84% So they have been paying, but they 01:58:27.080 --> 01:58:29.980 align:middle line:84% have been paying wrong salaries and not paying some people. 01:58:29.980 --> 01:58:33.380 align:middle line:84% And it's not like just 1,000 people were affected. 01:58:33.380 --> 01:58:37.120 align:middle line:84% 50-60% of people were getting wrong salaries for months 01:58:37.120 --> 01:58:39.260 align:middle line:84% and not getting salaries for a few more months. 01:58:39.260 --> 01:58:43.660 align:middle line:84% It's like-- anyway, it was a really terrible story, 01:58:43.660 --> 01:58:47.520 align:middle line:84% but a good anecdote for project dynamics class. 01:58:47.520 --> 01:58:52.040 align:middle line:90% So that was just a digression. 01:58:52.040 --> 01:58:52.580 align:middle line:90% OK. 01:58:52.580 --> 01:58:57.240 align:middle line:84% So I think this is-- yeah, we covered much here. 01:58:57.240 --> 01:58:58.240 align:middle line:90% Other volunteers? 01:58:58.240 --> 01:59:01.320 align:middle line:84% Actually, one question-- yeah, let's do yours. 01:59:01.320 --> 01:59:02.040 align:middle line:90% And then we can-- 01:59:02.040 --> 01:59:04.473 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:59:04.473 --> 01:59:05.640 align:middle line:90% HAZHIR RAHMANDAD: That's OK. 01:59:05.640 --> 01:59:06.180 align:middle line:90% That's OK. 01:59:06.180 --> 01:59:08.347 align:middle line:84% AUDIENCE: And I think based on what you said so far, 01:59:08.347 --> 01:59:12.220 align:middle line:84% I think [INAUDIBLE] correctly, so. 01:59:12.220 --> 01:59:15.680 align:middle line:84% HAZHIR RAHMANDAD: That's all fine. 01:59:15.680 --> 01:59:16.180 align:middle line:90% OK. 01:59:16.180 --> 01:59:20.020 align:middle line:90% So let's take a look here. 01:59:20.020 --> 01:59:24.060 align:middle line:84% So actually, let's go with the factories. 01:59:24.060 --> 01:59:27.720 align:middle line:84% So factory depreciating-- we talked about it, not important. 01:59:27.720 --> 01:59:31.460 align:middle line:84% First order draining here-- we talked about that. 01:59:31.460 --> 01:59:33.800 align:middle line:84% Construction time, first order-- that's fine. 01:59:33.800 --> 01:59:35.280 align:middle line:90% But keep that in mind. 01:59:35.280 --> 01:59:37.340 align:middle line:84% And then you have target delivery delay 01:59:37.340 --> 01:59:41.680 align:middle line:84% versus actual delivery delay with capacity expansion rate-- 01:59:41.680 --> 01:59:46.300 align:middle line:84% so this formulation has the problem that we just discussed 01:59:46.300 --> 01:59:49.440 align:middle line:90% about [? Auriana's ?] model. 01:59:49.440 --> 01:59:53.720 align:middle line:90% 01:59:53.720 --> 01:59:54.220 align:middle line:90% OK. 01:59:54.220 --> 02:00:01.440 align:middle line:84% So this is very kind of discrete way of thinking about it. 02:00:01.440 --> 02:00:04.400 align:middle line:84% So you're saying, if you're above, let's expand. 02:00:04.400 --> 02:00:06.940 align:middle line:84% If you're below, let's just stop expanding. 02:00:06.940 --> 02:00:12.400 align:middle line:84% So that's kind of too discrete for our typical formulation. 02:00:12.400 --> 02:00:16.680 align:middle line:84% So you want to look at maybe the ratio of actual delivery delay 02:00:16.680 --> 02:00:20.840 align:middle line:84% to target delivery delay and use that as the pressure. 02:00:20.840 --> 02:00:23.800 align:middle line:84% The second thing is that it has the problem that it is not 02:00:23.800 --> 02:00:29.620 align:middle line:84% anchored to any of the kind of scale of your operations-- 02:00:29.620 --> 02:00:33.380 align:middle line:84% how many factories, what capacity you currently have. 02:00:33.380 --> 02:00:37.920 align:middle line:84% So that was kind of very similar to [? Auriana's ?] challenge. 02:00:37.920 --> 02:00:41.640 align:middle line:84% So these are the two main issues here. 02:00:41.640 --> 02:00:43.480 align:middle line:90% Then we have the delivery delay. 02:00:43.480 --> 02:00:45.220 align:middle line:90% That is current. 02:00:45.220 --> 02:00:47.180 align:middle line:90% That's standard. 02:00:47.180 --> 02:00:51.600 align:middle line:84% Your production capacity is impacting delivery. 02:00:51.600 --> 02:00:54.200 align:middle line:84% But this is another issue that you have. 02:00:54.200 --> 02:00:56.640 align:middle line:84% The minimum of backlog divided by target delivery-- 02:00:56.640 --> 02:00:57.140 align:middle line:90% OK. 02:00:57.140 --> 02:01:04.980 align:middle line:84% So you have the maximum delivery rate 02:01:04.980 --> 02:01:08.440 align:middle line:84% taken care of within the formulation, which is OK. 02:01:08.440 --> 02:01:11.300 align:middle line:84% It's a little bit kind of cleaner 02:01:11.300 --> 02:01:14.940 align:middle line:84% to take out the maximum delivery rate 02:01:14.940 --> 02:01:19.300 align:middle line:84% as the ratio of backlog divided by target delivery delay. 02:01:19.300 --> 02:01:21.400 align:middle line:90% But that's OK. 02:01:21.400 --> 02:01:25.740 align:middle line:90% 02:01:25.740 --> 02:01:28.420 align:middle line:84% And the target delivery delay here 02:01:28.420 --> 02:01:34.600 align:middle line:84% being the driver makes things a little bit more problematic. 02:01:34.600 --> 02:01:43.520 align:middle line:84% So right now, this delivery rate will never be below backlog-- 02:01:43.520 --> 02:01:44.940 align:middle line:90% well, not never-- 02:01:44.940 --> 02:01:48.780 align:middle line:84% will not be above backlog divided by target delivery 02:01:48.780 --> 02:01:50.740 align:middle line:90% rate, delivery delay. 02:01:50.740 --> 02:01:55.060 align:middle line:84% And therefore, the delivery delay, the actual, 02:01:55.060 --> 02:02:01.380 align:middle line:84% will not go below target delivery delay under any case. 02:02:01.380 --> 02:02:03.040 align:middle line:90% Does that make sense? 02:02:03.040 --> 02:02:04.400 align:middle line:90% AUDIENCE: [INAUDIBLE] 02:02:04.400 --> 02:02:08.760 align:middle line:84% HAZHIR RAHMANDAD: So that because you're using the target 02:02:08.760 --> 02:02:13.440 align:middle line:84% delivery delay in the role of what we typically call minimum 02:02:13.440 --> 02:02:22.480 align:middle line:84% delivery delay, delivery rate will always be small enough that 02:02:22.480 --> 02:02:28.480 align:middle line:84% you cannot have a situation where actual delivery delay goes 02:02:28.480 --> 02:02:31.540 align:middle line:90% below target delivery delay. 02:02:31.540 --> 02:02:32.040 align:middle line:90% OK? 02:02:32.040 --> 02:02:36.280 align:middle line:84% And if it never goes below target delivery delay, 02:02:36.280 --> 02:02:39.440 align:middle line:90% this [? if and else ?] thingy-- 02:02:39.440 --> 02:02:44.520 align:middle line:84% so if actual delivery rate is bigger than target delivery 02:02:44.520 --> 02:02:46.600 align:middle line:90% delay-- 02:02:46.600 --> 02:02:49.120 align:middle line:84% so there are situations where the two will be equal. 02:02:49.120 --> 02:02:53.120 align:middle line:84% And therefore, this thing will go 0. 02:02:53.120 --> 02:02:56.440 align:middle line:84% But most of the time, this capacity expansion rate 02:02:56.440 --> 02:02:57.880 align:middle line:90% will be active. 02:02:57.880 --> 02:03:02.900 align:middle line:84% So it's not exactly what you want it to do, I think. 02:03:02.900 --> 02:03:03.400 align:middle line:90% OK. 02:03:03.400 --> 02:03:05.260 align:middle line:84% Then we have the product lifetime. 02:03:05.260 --> 02:03:06.920 align:middle line:90% People become potential. 02:03:06.920 --> 02:03:10.780 align:middle line:84% Once they became potential, they are now 02:03:10.780 --> 02:03:15.660 align:middle line:84% pitched to pitching capacity is from sales force 02:03:15.660 --> 02:03:20.900 align:middle line:84% and limited by a minimum pitch time-- 02:03:20.900 --> 02:03:24.280 align:middle line:90% so pitch time horizon. 02:03:24.280 --> 02:03:24.780 align:middle line:90% Yeah. 02:03:24.780 --> 02:03:26.322 align:middle line:84% So this is a very similar formulation 02:03:26.322 --> 02:03:29.140 align:middle line:90% as we had with other folks. 02:03:29.140 --> 02:03:32.360 align:middle line:84% It works with similar limitations. 02:03:32.360 --> 02:03:35.400 align:middle line:84% But the assumption that you have, which is different here, 02:03:35.400 --> 02:03:39.720 align:middle line:84% is that once you have been sold, the next time you come, again 02:03:39.720 --> 02:03:42.000 align:middle line:84% it takes the same amount of time to be sold to. 02:03:42.000 --> 02:03:43.580 align:middle line:90% And that's also OK. 02:03:43.580 --> 02:03:46.780 align:middle line:84% But we haven't yet seen any formulation that 02:03:46.780 --> 02:03:48.480 align:middle line:84% has kind of a balance between the two 02:03:48.480 --> 02:03:50.260 align:middle line:84% or allows different productivities 02:03:50.260 --> 02:03:52.820 align:middle line:84% for those different two purposes. 02:03:52.820 --> 02:03:56.740 align:middle line:84% And then you have the effect of delivery delay on conversion. 02:03:56.740 --> 02:04:00.240 align:middle line:84% Again, there is kind of the delay here for perception 02:04:00.240 --> 02:04:01.340 align:middle line:90% of the delivery delay. 02:04:01.340 --> 02:04:04.440 align:middle line:90% That is important, I think. 02:04:04.440 --> 02:04:07.140 align:middle line:84% Then attrition of sales force is fine, 02:04:07.140 --> 02:04:11.920 align:middle line:84% but you want to have replacement of sales force in the hiring 02:04:11.920 --> 02:04:17.000 align:middle line:84% process so that you have desired sales force 02:04:17.000 --> 02:04:20.780 align:middle line:84% minus sales force for the adjustment of sales force 02:04:20.780 --> 02:04:25.360 align:middle line:84% plus the attrition rate as the total adjustment of the-- 02:04:25.360 --> 02:04:27.360 align:middle line:90% or hiring rate. 02:04:27.360 --> 02:04:30.080 align:middle line:84% So one thing I kind of see here in multiple places 02:04:30.080 --> 02:04:32.920 align:middle line:84% is that you kind of combine multiple things 02:04:32.920 --> 02:04:36.180 align:middle line:90% inside the formulation. 02:04:36.180 --> 02:04:38.360 align:middle line:84% So you have fraction of revenue to sales force 02:04:38.360 --> 02:04:43.020 align:middle line:84% multiplied by revenue divided by cost per salesperson. 02:04:43.020 --> 02:04:47.520 align:middle line:84% All of this is your desired sales force. 02:04:47.520 --> 02:04:48.480 align:middle line:90% OK? 02:04:48.480 --> 02:04:50.900 align:middle line:84% So rather than this being the flow, 02:04:50.900 --> 02:04:53.120 align:middle line:84% it should be the desired sales force. 02:04:53.120 --> 02:04:55.400 align:middle line:84% And then you look at the current sales force 02:04:55.400 --> 02:04:58.540 align:middle line:84% compared to the desired sales force divided by some time 02:04:58.540 --> 02:04:59.580 align:middle line:90% to adjust-- 02:04:59.580 --> 02:05:02.660 align:middle line:90% close the gap for adjusting. 02:05:02.660 --> 02:05:05.240 align:middle line:84% So this one wouldn't actually do the job that you want. 02:05:05.240 --> 02:05:08.900 align:middle line:84% In fact, it may have even unit problems 02:05:08.900 --> 02:05:14.280 align:middle line:84% unless you have changed some of the units to be-- 02:05:14.280 --> 02:05:18.980 align:middle line:90% 02:05:18.980 --> 02:05:20.060 align:middle line:90% AUDIENCE: [INAUDIBLE] 02:05:20.060 --> 02:05:21.940 align:middle line:84% HAZHIR RAHMANDAD: Ah, this should be dollar 02:05:21.940 --> 02:05:23.680 align:middle line:90% per salesperson per month. 02:05:23.680 --> 02:05:26.500 align:middle line:84% That's why you don't get unit problem. 02:05:26.500 --> 02:05:31.240 align:middle line:84% So cost per sales force is not dollar per salesperson. 02:05:31.240 --> 02:05:32.440 align:middle line:90% This is not a slavery. 02:05:32.440 --> 02:05:37.100 align:middle line:84% You just are paying the monthly fees. 02:05:37.100 --> 02:05:43.660 align:middle line:84% So the units are not balancing here. 02:05:43.660 --> 02:05:46.580 align:middle line:90% All right. 02:05:46.580 --> 02:05:47.440 align:middle line:90% What else? 02:05:47.440 --> 02:05:52.000 align:middle line:90% 02:05:52.000 --> 02:05:52.500 align:middle line:90% Good. 02:05:52.500 --> 02:05:57.440 align:middle line:84% I think kind of you get the basic ideas. 02:05:57.440 --> 02:06:01.320 align:middle line:84% Question-- anybody has tried to make the formulation 02:06:01.320 --> 02:06:06.520 align:middle line:84% for fraction of budget going to sales force dynamic, 02:06:06.520 --> 02:06:10.880 align:middle line:84% depending on the size of the operations 02:06:10.880 --> 02:06:14.140 align:middle line:84% or kind of the effectiveness of the sales force? 02:06:14.140 --> 02:06:16.620 align:middle line:90% 02:06:16.620 --> 02:06:17.120 align:middle line:90% OK. 02:06:17.120 --> 02:06:20.700 align:middle line:90% 02:06:20.700 --> 02:06:21.200 align:middle line:90% OK. 02:06:21.200 --> 02:06:24.480 align:middle line:90% If nobody has done, that's OK. 02:06:24.480 --> 02:06:26.260 align:middle line:90% So the idea there-- 02:06:26.260 --> 02:06:28.980 align:middle line:84% I mean, we don't have an example to go through it in detail. 02:06:28.980 --> 02:06:30.730 align:middle line:84% But the idea there is that, right now, you 02:06:30.730 --> 02:06:34.480 align:middle line:84% have a fraction of budget that is going to sales force. 02:06:34.480 --> 02:06:37.200 align:middle line:84% And that as a heuristic may be fine. 02:06:37.200 --> 02:06:40.100 align:middle line:84% But during different stages of operation, 02:06:40.100 --> 02:06:42.240 align:middle line:90% that fraction should change. 02:06:42.240 --> 02:06:45.440 align:middle line:84% In practice, early on, you may actually put a lot more emphasis 02:06:45.440 --> 02:06:46.840 align:middle line:90% on the sales force. 02:06:46.840 --> 02:06:50.860 align:middle line:84% Later on, as you have a lot of customers who already know you 02:06:50.860 --> 02:06:53.800 align:middle line:84% and therefore the cost for converting them is lower, 02:06:53.800 --> 02:06:56.540 align:middle line:84% or because you just have captured all of the market 02:06:56.540 --> 02:06:59.380 align:middle line:84% and it's only the replacement rate that needs people, 02:06:59.380 --> 02:07:02.520 align:middle line:84% you want to reduce that budget for the sales force. 02:07:02.520 --> 02:07:05.060 align:middle line:84% You don't need to keep that big sales force, pay them. 02:07:05.060 --> 02:07:08.580 align:middle line:84% And that is right now not captured here 02:07:08.580 --> 02:07:10.380 align:middle line:90% in any of the models. 02:07:10.380 --> 02:07:12.520 align:middle line:84% So how would you think about capturing that? 02:07:12.520 --> 02:07:18.020 align:middle line:84% So what would be high-level, conceptual way 02:07:18.020 --> 02:07:20.100 align:middle line:90% of capturing that? 02:07:20.100 --> 02:07:23.460 align:middle line:84% AUDIENCE: [INAUDIBLE] base fraction of revenue 02:07:23.460 --> 02:07:25.740 align:middle line:84% that is allocated to the sales project. 02:07:25.740 --> 02:07:29.880 align:middle line:84% So if the effectiveness of the sales is at the maximum level, 02:07:29.880 --> 02:07:32.000 align:middle line:84% meaning that each sales person, for example, 02:07:32.000 --> 02:07:35.580 align:middle line:84% can recruit or teach 10 people, there 02:07:35.580 --> 02:07:37.960 align:middle line:84% would be no changes in that base fraction. 02:07:37.960 --> 02:07:40.360 align:middle line:84% But as soon as the fraction drops 02:07:40.360 --> 02:07:42.580 align:middle line:84% because the market is getting saturated, 02:07:42.580 --> 02:07:46.780 align:middle line:84% that sales budget can also drop [INAUDIBLE] 02:07:46.780 --> 02:07:48.780 align:middle line:84% HAZHIR RAHMANDAD: So basically calculate 02:07:48.780 --> 02:07:52.320 align:middle line:84% the net productivity of sales force 02:07:52.320 --> 02:07:56.120 align:middle line:84% and then use that net productivity to adjust. 02:07:56.120 --> 02:08:00.020 align:middle line:84% If that net productivity goes down, bring down that fraction. 02:08:00.020 --> 02:08:01.880 align:middle line:90% So that's a good first step. 02:08:01.880 --> 02:08:04.120 align:middle line:84% Anybody wants to improve upon that? 02:08:04.120 --> 02:08:04.683 align:middle line:90% Alexander. 02:08:04.683 --> 02:08:06.100 align:middle line:84% AUDIENCE: Just a very minor thing, 02:08:06.100 --> 02:08:08.480 align:middle line:84% but you could convert the productivity 02:08:08.480 --> 02:08:11.780 align:middle line:84% into the marginal revenue of each salesperson, right? 02:08:11.780 --> 02:08:14.320 align:middle line:84% So you just multiply how many customers 02:08:14.320 --> 02:08:18.910 align:middle line:84% each salesperson is getting times the order quantity 02:08:18.910 --> 02:08:19.410 align:middle line:90% and price. 02:08:19.410 --> 02:08:19.910 align:middle line:90% Right? 02:08:19.910 --> 02:08:22.280 align:middle line:84% And then that way you can compare that 02:08:22.280 --> 02:08:24.500 align:middle line:90% to the wage of the salesperson. 02:08:24.500 --> 02:08:27.000 align:middle line:84% So that way, if it's above 1, then salespeople 02:08:27.000 --> 02:08:28.620 align:middle line:90% are net profitable. 02:08:28.620 --> 02:08:30.100 align:middle line:90% You want to hire more of them. 02:08:30.100 --> 02:08:32.965 align:middle line:84% If it's less than 1, then you want to decrease [INAUDIBLE]. 02:08:32.965 --> 02:08:33.840 align:middle line:90% HAZHIR RAHMANDAD: OK. 02:08:33.840 --> 02:08:34.340 align:middle line:90% Very good. 02:08:34.340 --> 02:08:38.200 align:middle line:84% So that's kind of a very clean, economically logical way 02:08:38.200 --> 02:08:39.260 align:middle line:90% of doing it. 02:08:39.260 --> 02:08:43.580 align:middle line:84% Basically, what is the marginal profits from each salesperson 02:08:43.580 --> 02:08:44.080 align:middle line:90% now? 02:08:44.080 --> 02:08:45.920 align:middle line:84% For each sales that they do, there 02:08:45.920 --> 02:08:48.300 align:middle line:84% is a marginal profit that is generated. 02:08:48.300 --> 02:08:53.007 align:middle line:84% So is it worth having more of these salespeople or not? 02:08:53.007 --> 02:08:53.840 align:middle line:90% That's the question. 02:08:53.840 --> 02:08:56.220 align:middle line:84% And in fact, if you do it that way, 02:08:56.220 --> 02:09:04.240 align:middle line:84% you can basically make the sales force exposed to a pressure. 02:09:04.240 --> 02:09:06.080 align:middle line:84% And that pressure comes from this factor. 02:09:06.080 --> 02:09:08.720 align:middle line:90% If it is above 1, you expand. 02:09:08.720 --> 02:09:10.640 align:middle line:90% If it is below 1, you shrink. 02:09:10.640 --> 02:09:14.420 align:middle line:84% And that 1 is the ratio of the marginal profit 02:09:14.420 --> 02:09:18.000 align:middle line:84% per salesperson divided by the salary of the salesperson. 02:09:18.000 --> 02:09:22.020 align:middle line:84% So that becomes a pretty nice endogenous way 02:09:22.020 --> 02:09:24.460 align:middle line:84% of managing the size of the sales force 02:09:24.460 --> 02:09:27.540 align:middle line:84% so that it is consistent with your profitability 02:09:27.540 --> 02:09:28.710 align:middle line:90% from those people. 02:09:28.710 --> 02:09:30.460 align:middle line:84% AUDIENCE: And the sales force productivity 02:09:30.460 --> 02:09:32.860 align:middle line:84% would have to come from some relationship 02:09:32.860 --> 02:09:36.515 align:middle line:84% between the stock of potential customers and your [INAUDIBLE] 02:09:36.515 --> 02:09:38.640 align:middle line:84% HAZHIR RAHMANDAD: So the sales force productivity-- 02:09:38.640 --> 02:09:41.860 align:middle line:84% you don't need to worry about doing all of those calculations 02:09:41.860 --> 02:09:42.780 align:middle line:90% separately. 02:09:42.780 --> 02:09:45.160 align:middle line:84% You have a total sales that is happening. 02:09:45.160 --> 02:09:48.040 align:middle line:90% 02:09:48.040 --> 02:09:53.260 align:middle line:84% And you have total sales force you have. 02:09:53.260 --> 02:09:57.480 align:middle line:84% So you can get the average sales force productivity right away. 02:09:57.480 --> 02:09:58.860 align:middle line:90% Now, that is average. 02:09:58.860 --> 02:10:00.360 align:middle line:90% It is not marginal. 02:10:00.360 --> 02:10:04.240 align:middle line:84% So this is actually one level more complex, 02:10:04.240 --> 02:10:06.340 align:middle line:84% if you want to think about it this way. 02:10:06.340 --> 02:10:08.513 align:middle line:84% So calculating the average is very easy. 02:10:08.513 --> 02:10:10.180 align:middle line:84% You already have all of the information. 02:10:10.180 --> 02:10:13.800 align:middle line:84% You just divide the total sales to the total-- 02:10:13.800 --> 02:10:17.100 align:middle line:84% but this is not exactly the marginal productivity. 02:10:17.100 --> 02:10:20.560 align:middle line:84% Marginal productivity is, for each extra sales 02:10:20.560 --> 02:10:23.120 align:middle line:90% force, what will happen? 02:10:23.120 --> 02:10:28.000 align:middle line:84% And doing that requires actually a little bit more tracking. 02:10:28.000 --> 02:10:30.300 align:middle line:84% And the way you do it, it's not very standard. 02:10:30.300 --> 02:10:33.560 align:middle line:84% We don't have the infrastructure built into this structure 02:10:33.560 --> 02:10:34.520 align:middle line:90% right now. 02:10:34.520 --> 02:10:37.320 align:middle line:90% So you could do it analytically. 02:10:37.320 --> 02:10:40.960 align:middle line:84% That is, basically, if you have simpler formulations, 02:10:40.960 --> 02:10:47.540 align:middle line:84% you can calculate the marginal value analytically 02:10:47.540 --> 02:10:50.000 align:middle line:84% as a separate path that is saying, OK, 02:10:50.000 --> 02:10:54.620 align:middle line:84% if my formulation for sales force impact on sales 02:10:54.620 --> 02:11:01.420 align:middle line:84% is like this, what is the marginal productivity 02:11:01.420 --> 02:11:05.740 align:middle line:84% of the sales force, which is the derivative of profitability 02:11:05.740 --> 02:11:07.240 align:middle line:90% with respect to sales force? 02:11:07.240 --> 02:11:10.160 align:middle line:84% You can write that, and calculate that, and track that. 02:11:10.160 --> 02:11:12.940 align:middle line:84% So that would be one way to do it. 02:11:12.940 --> 02:11:16.700 align:middle line:84% For more complex models, it's even harder. 02:11:16.700 --> 02:11:20.780 align:middle line:84% But one way to do it is to actually always use a-- 02:11:20.780 --> 02:11:23.780 align:middle line:84% so this is starting to become more complex. 02:11:23.780 --> 02:11:32.580 align:middle line:84% So you add a subscript to track the value of sales 02:11:32.580 --> 02:11:36.140 align:middle line:84% if you had one more person, one more salesperson. 02:11:36.140 --> 02:11:40.300 align:middle line:84% So you're calculating the current sales force and what 02:11:40.300 --> 02:11:42.220 align:middle line:90% is the impact on the sales. 02:11:42.220 --> 02:11:43.840 align:middle line:90% There is one pathway. 02:11:43.840 --> 02:11:45.360 align:middle line:84% And then you would add a subscript 02:11:45.360 --> 02:11:50.440 align:middle line:84% along this pathway for extras or marginally more sales force 02:11:50.440 --> 02:11:56.360 align:middle line:84% and calculate the sales under that alternative formulation. 02:11:56.360 --> 02:11:58.920 align:middle line:84% And then you look at that delta and divide it 02:11:58.920 --> 02:12:03.720 align:middle line:84% by 1, which is in this case the total difference 02:12:03.720 --> 02:12:04.720 align:middle line:90% between the two. 02:12:04.720 --> 02:12:09.060 align:middle line:84% So I haven't seen many people use it, 02:12:09.060 --> 02:12:11.840 align:middle line:84% but that's how you would do it if the formulations are 02:12:11.840 --> 02:12:16.640 align:middle line:84% too hard to do analytical tracking of the marginal. 02:12:16.640 --> 02:12:19.200 align:middle line:84% Is the difference between marginal and average clear? 02:12:19.200 --> 02:12:21.260 align:middle line:84% Because that's the first starting point. 02:12:21.260 --> 02:12:23.520 align:middle line:84% So marginal productivity of a salesperson 02:12:23.520 --> 02:12:26.880 align:middle line:84% is, if I add one more salesperson, how many dollars 02:12:26.880 --> 02:12:28.440 align:middle line:90% would they make? 02:12:28.440 --> 02:12:32.160 align:middle line:84% Average productivity is saying, I have 100 salespeople 02:12:32.160 --> 02:12:36.320 align:middle line:84% and I'm making $1 million, so that's $10,000 per person 02:12:36.320 --> 02:12:37.720 align:middle line:90% on average. 02:12:37.720 --> 02:12:40.000 align:middle line:84% The question of, do I add one more person 02:12:40.000 --> 02:12:41.980 align:middle line:84% or do I reduce one more person, should 02:12:41.980 --> 02:12:44.660 align:middle line:84% be determined by the marginal, not the average. 02:12:44.660 --> 02:12:49.540 align:middle line:84% Average is not always but most of the time overestimating 02:12:49.540 --> 02:12:51.500 align:middle line:90% the benefit of the salespeople. 02:12:51.500 --> 02:12:55.100 align:middle line:84% And therefore, expanding or shrinking based on average 02:12:55.100 --> 02:12:56.920 align:middle line:90% will give you the wrong answer. 02:12:56.920 --> 02:12:58.940 align:middle line:84% That is, if you increase your sales force 02:12:58.940 --> 02:13:03.780 align:middle line:84% until you have the average be 0, then 02:13:03.780 --> 02:13:06.763 align:middle line:84% you already have a very bloated sales force. 02:13:06.763 --> 02:13:08.180 align:middle line:84% Many of those people you shouldn't 02:13:08.180 --> 02:13:09.430 align:middle line:90% have hired in the first place. 02:13:09.430 --> 02:13:12.680 align:middle line:84% So that is why using the marginal is important. 02:13:12.680 --> 02:13:13.180 align:middle line:90% Good. 02:13:13.180 --> 02:13:15.700 align:middle line:84% So I think that was kind of a conversation that 02:13:15.700 --> 02:13:21.340 align:middle line:84% wasn't obvious without having it, but I think-- 02:13:21.340 --> 02:13:22.220 align:middle line:90% OK. 02:13:22.220 --> 02:13:25.320 align:middle line:90% We could look at one more model. 02:13:25.320 --> 02:13:27.665 align:middle line:84% Anybody wants their model to be looked at? 02:13:27.665 --> 02:13:30.040 align:middle line:84% AUDIENCE: I just have a question about what we just said. 02:13:30.040 --> 02:13:30.998 align:middle line:90% HAZHIR RAHMANDAD: Sure. 02:13:30.998 --> 02:13:32.340 align:middle line:90% Go ahead, Jake. 02:13:32.340 --> 02:13:36.380 align:middle line:84% AUDIENCE: So if we're looking at a marginal profit 02:13:36.380 --> 02:13:41.800 align:middle line:84% of another sales person, when you get into actually making 02:13:41.800 --> 02:13:46.955 align:middle line:84% that, would you also have to consider the fact that if you 02:13:46.955 --> 02:13:49.080 align:middle line:84% add too many too quickly, [INAUDIBLE] might go down 02:13:49.080 --> 02:13:51.080 align:middle line:84% because you reputation goes down because you haven't built 02:13:51.080 --> 02:13:52.040 align:middle line:90% that factory yet. 02:13:52.040 --> 02:13:53.980 align:middle line:90% And you've introduced a delay. 02:13:53.980 --> 02:13:56.340 align:middle line:84% Like, do you need to get that granular with it? 02:13:56.340 --> 02:13:59.760 align:middle line:84% Or are you thinking long run [INAUDIBLE] 02:13:59.760 --> 02:14:02.300 align:middle line:84% gets to equilibrium state with a new salesperson? 02:14:02.300 --> 02:14:03.258 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 02:14:03.258 --> 02:14:08.140 align:middle line:84% So in the first cut at it, I would just do this. 02:14:08.140 --> 02:14:11.320 align:middle line:84% That is, I would say, OK, sales force 02:14:11.320 --> 02:14:13.660 align:middle line:90% is creating pitching capacity. 02:14:13.660 --> 02:14:15.200 align:middle line:84% So I'll make the pitching capacity 02:14:15.200 --> 02:14:18.980 align:middle line:84% in this model subscripted with two versions. 02:14:18.980 --> 02:14:20.780 align:middle line:84% One is with the current sales force, 02:14:20.780 --> 02:14:23.520 align:middle line:84% another with sales force plus one person. 02:14:23.520 --> 02:14:26.400 align:middle line:84% And then wherever it goes-- so it is changing the customer 02:14:26.400 --> 02:14:27.560 align:middle line:90% conversion. 02:14:27.560 --> 02:14:29.980 align:middle line:84% I'll have two versions of customer conversion. 02:14:29.980 --> 02:14:31.820 align:middle line:84% One of them is operating on the stocks. 02:14:31.820 --> 02:14:34.280 align:middle line:84% The other version I'm just-- for keeping track 02:14:34.280 --> 02:14:36.400 align:middle line:90% of how things are happening. 02:14:36.400 --> 02:14:38.540 align:middle line:84% And then that customer conversion somewhere 02:14:38.540 --> 02:14:40.220 align:middle line:90% shows up in my profits. 02:14:40.220 --> 02:14:43.580 align:middle line:90% 02:14:43.580 --> 02:14:44.160 align:middle line:90% Yeah. 02:14:44.160 --> 02:14:45.860 align:middle line:84% In the accounting profits, I'm also 02:14:45.860 --> 02:14:48.900 align:middle line:84% tracking that sales force and so on. 02:14:48.900 --> 02:14:55.280 align:middle line:84% So that version is fairly short term-focused. 02:14:55.280 --> 02:14:59.140 align:middle line:84% It is not really capturing the longer-term consequences 02:14:59.140 --> 02:15:03.340 align:middle line:84% of bringing this new salesperson. 02:15:03.340 --> 02:15:06.940 align:middle line:84% Doing those longer-term consequences is much harder. 02:15:06.940 --> 02:15:08.720 align:middle line:90% It is actually more precise. 02:15:08.720 --> 02:15:13.280 align:middle line:84% That is a good calculation of what the right balance is, 02:15:13.280 --> 02:15:15.540 align:middle line:84% should account for second and third order 02:15:15.540 --> 02:15:19.820 align:middle line:84% effects that would require adjustments of other stocks. 02:15:19.820 --> 02:15:22.380 align:middle line:84% But then you basically would be needing 02:15:22.380 --> 02:15:25.380 align:middle line:84% to-- for every time step, you need 02:15:25.380 --> 02:15:31.140 align:middle line:84% to simulate the model with that alternative formulation 02:15:31.140 --> 02:15:33.900 align:middle line:84% or that alternative number of sales force, 02:15:33.900 --> 02:15:36.400 align:middle line:84% do a calculation of the marginal, 02:15:36.400 --> 02:15:41.600 align:middle line:84% then come back and decide how many people you want to have. 02:15:41.600 --> 02:15:45.380 align:middle line:84% So in a way, it is what organizations are doing. 02:15:45.380 --> 02:15:47.960 align:middle line:84% Organizations have their simulation models 02:15:47.960 --> 02:15:51.960 align:middle line:84% or their spreadsheet models that try 02:15:51.960 --> 02:15:55.680 align:middle line:84% to forecast what would be the impact of each additional sales 02:15:55.680 --> 02:15:58.760 align:middle line:84% force over the time and therefore bring it back 02:15:58.760 --> 02:16:01.340 align:middle line:84% into that decision every period of time. 02:16:01.340 --> 02:16:04.900 align:middle line:84% Now, they don't do it every time step in the simulation sense. 02:16:04.900 --> 02:16:06.600 align:middle line:84% They would do this maybe periodically, 02:16:06.600 --> 02:16:08.640 align:middle line:90% every quarter or something. 02:16:08.640 --> 02:16:13.760 align:middle line:84% So conceptually, that is not too far from what the organizations 02:16:13.760 --> 02:16:18.740 align:middle line:84% also do, but doing it in these models becomes exceedingly hard. 02:16:18.740 --> 02:16:21.480 align:middle line:90% 02:16:21.480 --> 02:16:27.400 align:middle line:84% So that would require a separate simulation of some simpler 02:16:27.400 --> 02:16:30.460 align:middle line:84% version of the model-- not all of the model necessarily, 02:16:30.460 --> 02:16:33.940 align:middle line:84% but some simpler version of the model every time period 02:16:33.940 --> 02:16:36.400 align:middle line:84% to inform just one of the decisions in the model. 02:16:36.400 --> 02:16:39.879 align:middle line:84% And that is logistically hard to do. 02:16:39.879 --> 02:16:42.340 align:middle line:84% In Vensim, you basically need to define 02:16:42.340 --> 02:16:45.180 align:middle line:84% some kind of external function that goes and calculates 02:16:45.180 --> 02:16:48.059 align:middle line:84% that, maybe calling Vensim separately, doing 02:16:48.059 --> 02:16:50.219 align:middle line:84% that simulation, then coming back, and plugging 02:16:50.219 --> 02:16:52.379 align:middle line:84% in that number into the next time step. 02:16:52.379 --> 02:16:56.780 align:middle line:84% And it's just a headache I'm not going to worry too much about. 02:16:56.780 --> 02:17:00.760 align:middle line:84% But sometimes it becomes important and you need to do it. 02:17:00.760 --> 02:17:03.700 align:middle line:84% So in those cases, yes, think about it. 02:17:03.700 --> 02:17:09.500 align:middle line:84% So Aria-- for her general exam, one part of the question 02:17:09.500 --> 02:17:14.020 align:middle line:84% was about the marginal value of investing 02:17:14.020 --> 02:17:17.100 align:middle line:84% in new shows for streaming services. 02:17:17.100 --> 02:17:20.299 align:middle line:84% And these investments have long term impacts. 02:17:20.299 --> 02:17:23.920 align:middle line:84% Because it is not just, next period, how many I sell. 02:17:23.920 --> 02:17:28.860 align:middle line:84% But it is over the life of this show how many would I sell, 02:17:28.860 --> 02:17:35.920 align:middle line:84% and how many it kind of reduces the value for my other shows-- 02:17:35.920 --> 02:17:39.020 align:middle line:84% it cannibalizes my other shows-- and then how many new customers 02:17:39.020 --> 02:17:40.080 align:middle line:90% it brings to me. 02:17:40.080 --> 02:17:43.860 align:middle line:84% So it was a very complex decision. 02:17:43.860 --> 02:17:45.879 align:middle line:84% So coming up with a marginal value 02:17:45.879 --> 02:17:48.400 align:middle line:90% was a very non-trivial task. 02:17:48.400 --> 02:17:50.280 align:middle line:84% And ultimately, it was kind of, how can you 02:17:50.280 --> 02:17:53.740 align:middle line:84% approximate that using relatively simple, 02:17:53.740 --> 02:17:57.920 align:middle line:84% manageable analytical formulations? 02:17:57.920 --> 02:18:02.760 align:middle line:84% Anyway, that's kind of a long answer to a question that 02:18:02.760 --> 02:18:05.400 align:middle line:90% is not easy, ultimately. 02:18:05.400 --> 02:18:06.559 align:middle line:90% OK. 02:18:06.559 --> 02:18:08.879 align:middle line:84% But now probably it's a good time 02:18:08.879 --> 02:18:13.240 align:middle line:84% to just review the problem for next week, 02:18:13.240 --> 02:18:15.900 align:middle line:90% given the limited time we have. 02:18:15.900 --> 02:18:27.040 align:middle line:84% So next week, we are having the endogenous diffusion model. 02:18:27.040 --> 02:18:32.500 align:middle line:84% So this is kind of motivated by Jack Homer's research 02:18:32.500 --> 02:18:35.620 align:middle line:90% on pacemakers. 02:18:35.620 --> 02:18:39.280 align:middle line:90% So this he did for his thesis. 02:18:39.280 --> 02:18:41.040 align:middle line:84% So this is an old, old piece of work. 02:18:41.040 --> 02:18:44.400 align:middle line:84% But he continued working on this space. 02:18:44.400 --> 02:18:47.084 align:middle line:84% It's a very generic way of thinking about diffusion. 02:18:47.084 --> 02:18:48.459 align:middle line:84% So typically in diffusion models, 02:18:48.459 --> 02:18:53.480 align:middle line:84% we think about a single concrete product and how it gets adopted. 02:18:53.480 --> 02:18:56.139 align:middle line:84% But if you think about a product that is evolving over 02:18:56.139 --> 02:18:59.219 align:middle line:84% time-- it is getting better, features and functionalities 02:18:59.219 --> 02:19:00.400 align:middle line:90% are added to it. 02:19:00.400 --> 02:19:01.980 align:middle line:84% So pacemakers are a good example. 02:19:01.980 --> 02:19:03.879 align:middle line:84% Because when they were introduced-- 02:19:03.879 --> 02:19:07.059 align:middle line:84% I don't know, in the 1970s or early '80s-- no, '70s, I think. 02:19:07.059 --> 02:19:11.340 align:middle line:84% They were very big, bulky, only useful 02:19:11.340 --> 02:19:13.980 align:middle line:84% for the most critical patients who would die without them 02:19:13.980 --> 02:19:15.860 align:middle line:90% or something like that. 02:19:15.860 --> 02:19:18.559 align:middle line:90% It was a very invasive surgery. 02:19:18.559 --> 02:19:21.940 align:middle line:84% But over time, they became more efficient, smaller. 02:19:21.940 --> 02:19:25.100 align:middle line:84% Their signaling became more sophisticated. 02:19:25.100 --> 02:19:27.160 align:middle line:90% Their battery life extended. 02:19:27.160 --> 02:19:29.360 align:middle line:84% All of those increased the use cases 02:19:29.360 --> 02:19:31.520 align:middle line:84% that pacemakers could be used for so 02:19:31.520 --> 02:19:35.320 align:middle line:84% that these days many people who otherwise can survive very well 02:19:35.320 --> 02:19:38.799 align:middle line:84% and have a decently normal life would still 02:19:38.799 --> 02:19:39.959 align:middle line:90% benefit from pacemaker. 02:19:39.959 --> 02:19:43.500 align:middle line:84% Because on the margin, it adds value for them as well. 02:19:43.500 --> 02:19:46.920 align:middle line:84% So that is expanding the size of the market. 02:19:46.920 --> 02:19:49.000 align:middle line:84% Now there are additional things that 02:19:49.000 --> 02:19:52.120 align:middle line:84% come into play-- that you add these new features 02:19:52.120 --> 02:19:53.260 align:middle line:90% and functionalities. 02:19:53.260 --> 02:19:55.680 align:middle line:84% You make the product potentially more applicable 02:19:55.680 --> 02:20:00.740 align:middle line:84% to a larger customer set, but you're also introducing bugs. 02:20:00.740 --> 02:20:04.400 align:middle line:84% You're introducing problems that are not known at the beginning. 02:20:04.400 --> 02:20:07.620 align:middle line:84% So discovery of those problems, fixing of those problems, 02:20:07.620 --> 02:20:10.600 align:middle line:84% and the reputational impact that those problems have 02:20:10.600 --> 02:20:14.700 align:middle line:84% on the product are also important in the adoption. 02:20:14.700 --> 02:20:17.140 align:middle line:84% So it can even create oscillatory modes. 02:20:17.140 --> 02:20:19.520 align:middle line:84% So he showed that, for example, typically 02:20:19.520 --> 02:20:23.120 align:middle line:84% in the case of pacemakers, there is delays in research 02:20:23.120 --> 02:20:28.880 align:middle line:84% that identifies and captures those side effects and problems. 02:20:28.880 --> 02:20:32.580 align:middle line:84% And that delay leads to some oscillatory mode 02:20:32.580 --> 02:20:35.700 align:middle line:84% in adoption of pacemakers-- that kind of they 02:20:35.700 --> 02:20:37.720 align:middle line:90% come out with new versions. 02:20:37.720 --> 02:20:41.020 align:middle line:90% They create lots of excitement. 02:20:41.020 --> 02:20:43.580 align:middle line:84% But then early research typically 02:20:43.580 --> 02:20:46.420 align:middle line:84% kind shows some disappointing results 02:20:46.420 --> 02:20:49.880 align:middle line:84% that then needs to be fixed, those problems. 02:20:49.880 --> 02:20:52.460 align:middle line:84% And that brings down adoption for a while. 02:20:52.460 --> 02:20:55.180 align:middle line:90% And then it starts to recover. 02:20:55.180 --> 02:21:01.580 align:middle line:84% So that's what your task is to formulate, 02:21:01.580 --> 02:21:06.460 align:middle line:84% a model of adoption that is not your typical Bass diffusion 02:21:06.460 --> 02:21:07.000 align:middle line:90% model. 02:21:07.000 --> 02:21:09.560 align:middle line:84% In fact, the way Jack formulated it-- 02:21:09.560 --> 02:21:12.900 align:middle line:84% and I suggested that you can formulate it with that idea-- 02:21:12.900 --> 02:21:17.840 align:middle line:84% doesn't use the Bass infrastructure at all. 02:21:17.840 --> 02:21:23.400 align:middle line:84% It kind of says there is a indicated 02:21:23.400 --> 02:21:26.040 align:middle line:90% level of potential users. 02:21:26.040 --> 02:21:29.600 align:middle line:84% And the actual users move towards that indicated level. 02:21:29.600 --> 02:21:32.660 align:middle line:84% But the speed by which they move towards that indicated level, 02:21:32.660 --> 02:21:34.240 align:middle line:84% the time constant there, could be 02:21:34.240 --> 02:21:37.680 align:middle line:84% a function of how many other adopters are out there. 02:21:37.680 --> 02:21:41.280 align:middle line:84% So this is a different way of capturing the Bass diffusion 02:21:41.280 --> 02:21:42.620 align:middle line:90% model's reinforcing loop. 02:21:42.620 --> 02:21:45.160 align:middle line:84% So the Bass diffusion model says people 02:21:45.160 --> 02:21:47.660 align:middle line:84% who adopt it talk with people who don't have it. 02:21:47.660 --> 02:21:50.120 align:middle line:90% And that leads to adoption. 02:21:50.120 --> 02:21:53.520 align:middle line:84% This alternative says there is an indicated level of how 02:21:53.520 --> 02:21:55.320 align:middle line:90% many ultimately would adopt. 02:21:55.320 --> 02:21:56.760 align:middle line:90% There is a current level. 02:21:56.760 --> 02:21:57.900 align:middle line:90% And you look at the gap. 02:21:57.900 --> 02:22:00.100 align:middle line:84% You close that gap over a time constant. 02:22:00.100 --> 02:22:04.040 align:middle line:84% But that time constant can become smaller 02:22:04.040 --> 02:22:05.720 align:middle line:84% if you have a lot of people who already 02:22:05.720 --> 02:22:09.000 align:middle line:84% have adopted because then the information spreads faster. 02:22:09.000 --> 02:22:13.520 align:middle line:84% So that's a different version of a Bass core idea 02:22:13.520 --> 02:22:16.960 align:middle line:84% which accounts for the advertisement 02:22:16.960 --> 02:22:19.420 align:middle line:84% channel or the alternative channels as well. 02:22:19.420 --> 02:22:22.820 align:middle line:84% Because the time constant is never infinite. 02:22:22.820 --> 02:22:24.540 align:middle line:90% So there is a max time constant. 02:22:24.540 --> 02:22:31.060 align:middle line:84% If nobody else's have it, there is still a fractional rate 02:22:31.060 --> 02:22:33.875 align:middle line:90% of movement towards this level. 02:22:33.875 --> 02:22:35.500 align:middle line:84% So that's a kind of an alternative way. 02:22:35.500 --> 02:22:36.320 align:middle line:90% So think about it. 02:22:36.320 --> 02:22:38.580 align:middle line:90% Use it just to see how it works. 02:22:38.580 --> 02:22:41.300 align:middle line:90% 02:22:41.300 --> 02:22:47.260 align:middle line:84% So you'll see the problem definition in the folder. 02:22:47.260 --> 02:22:50.820 align:middle line:84% It is a little bit less well defined. 02:22:50.820 --> 02:22:53.300 align:middle line:84% So you will need to make more architectural choices. 02:22:53.300 --> 02:22:56.100 align:middle line:84% I think this week, the problem-- even 02:22:56.100 --> 02:22:57.940 align:middle line:84% though it was a broader model boundary 02:22:57.940 --> 02:23:01.560 align:middle line:84% compared to the previous weeks, it was fairly well defined. 02:23:01.560 --> 02:23:06.560 align:middle line:84% So we see a lot of similarity across different submissions. 02:23:06.560 --> 02:23:08.920 align:middle line:84% And overall, you folks had done well. 02:23:08.920 --> 02:23:14.020 align:middle line:84% I think many of the submissions were in the right space. 02:23:14.020 --> 02:23:19.220 align:middle line:84% Now, we could improve and change some of these things. 02:23:19.220 --> 02:23:21.960 align:middle line:84% So this one might be a little bit more challenging. 02:23:21.960 --> 02:23:23.720 align:middle line:84% That's why I put it a week later. 02:23:23.720 --> 02:23:27.520 align:middle line:90% But it shouldn't be too hard. 02:23:27.520 --> 02:23:28.020 align:middle line:90% OK. 02:23:28.020 --> 02:23:31.720 align:middle line:84% So that's the plan for next week. 02:23:31.720 --> 02:23:35.860 align:middle line:90% Any questions on this exercise? 02:23:35.860 --> 02:23:41.040 align:middle line:90% 02:23:41.040 --> 02:23:44.060 align:middle line:84% AUDIENCE: So there is a bit of word of mouth here, right? 02:23:44.060 --> 02:23:45.680 align:middle line:84% HAZHIR RAHMANDAD: Yes, there is a bit of word of mouth. 02:23:45.680 --> 02:23:47.263 align:middle line:84% AUDIENCE: Both positive and negative-- 02:23:47.263 --> 02:23:49.380 align:middle line:84% more doctors tell people about the increased scope 02:23:49.380 --> 02:23:51.842 align:middle line:84% but also the bugs as a result of the newly [INAUDIBLE]. 02:23:51.842 --> 02:23:52.800 align:middle line:90% HAZHIR RAHMANDAD: Yeah. 02:23:52.800 --> 02:23:55.760 align:middle line:84% So how you deal with that is kind of a little bit 02:23:55.760 --> 02:23:56.940 align:middle line:90% of an open question. 02:23:56.940 --> 02:24:02.760 align:middle line:84% So I left it there at this level, that bugs are there. 02:24:02.760 --> 02:24:04.460 align:middle line:84% And they get discovered and fixed. 02:24:04.460 --> 02:24:07.280 align:middle line:84% But when they get discovered, they reduce utility. 02:24:07.280 --> 02:24:10.240 align:middle line:84% In fact, some of the bugs, they don't 02:24:10.240 --> 02:24:13.420 align:middle line:84% need to be discovered broadly and perceived broadly. 02:24:13.420 --> 02:24:18.640 align:middle line:84% Just the fact that it is there, you as a user will feel it. 02:24:18.640 --> 02:24:20.520 align:middle line:84% It does have an impact on your utility. 02:24:20.520 --> 02:24:22.740 align:middle line:84% But the people who have not adopted it usually 02:24:22.740 --> 02:24:23.760 align:middle line:90% need to hear about it. 02:24:23.760 --> 02:24:26.000 align:middle line:84% So that discovery process is important. 02:24:26.000 --> 02:24:28.900 align:middle line:90% 02:24:28.900 --> 02:24:30.850 align:middle line:90% OK. 02:24:30.850 --> 02:24:34.460 align:middle line:90% So that's for the next week. 02:24:34.460 --> 02:24:38.580 align:middle line:84% Do please also revisit your proposals. 02:24:38.580 --> 02:24:43.220 align:middle line:84% Let's kind of finalize those so that next week everybody has 02:24:43.220 --> 02:24:48.500 align:middle line:84% a version that is kind of close to final that others-- 02:24:48.500 --> 02:24:49.940 align:middle line:84% when they go through the exercise 02:24:49.940 --> 02:24:52.080 align:middle line:84% we talked about for the week of 18, 02:24:52.080 --> 02:24:57.500 align:middle line:84% they can look at those problem definitions. 02:24:57.500 --> 02:24:58.040 align:middle line:90% Cool. 02:24:58.040 --> 02:24:59.040 align:middle line:90% Any questions? 02:24:59.040 --> 02:25:02.220 align:middle line:90% 02:25:02.220 --> 02:25:03.720 align:middle line:90% OK. 02:25:03.720 --> 02:25:06.860 align:middle line:90% So enjoy your weekend, folks. 02:25:06.860 --> 02:25:09.020 align:middle line:90% Good to see those online. 02:25:09.020 --> 02:25:11.360 align:middle line:90% And see you next week. 02:25:11.360 --> 02:25:12.910 align:middle line:90% Yeah. 02:25:12.910 --> 02:25:17.000 align:middle line:90%