1 00:00:00,000 --> 00:00:02,430 The following content is provided under a Creative 2 00:00:02,430 --> 00:00:03,730 Commons license. 3 00:00:03,730 --> 00:00:06,060 Your support will help MIT OpenCourseWare 4 00:00:06,060 --> 00:00:10,090 continue to offer high quality educational resources for free. 5 00:00:10,090 --> 00:00:12,690 To make a donation or to view additional materials 6 00:00:12,690 --> 00:00:16,560 from hundreds of MIT courses, visit MIT OpenCourseWare 7 00:00:16,560 --> 00:00:17,744 at ocw.mit.edu. 8 00:00:28,040 --> 00:00:31,450 PROFESSOR: OK, so we're going to talk about Six Sigma. 9 00:00:31,450 --> 00:00:33,490 Now, if you've seen, the course is heavily 10 00:00:33,490 --> 00:00:37,690 weighted towards Lean with a little bit of Six Sigma. 11 00:00:37,690 --> 00:00:41,770 And we think that's the appropriate place to start. 12 00:00:41,770 --> 00:00:45,130 But we need to get you started on Six Sigma. 13 00:00:45,130 --> 00:00:48,640 And so learning objectives here are 14 00:00:48,640 --> 00:00:52,420 that Six Sigma is a valuable approach for improving process 15 00:00:52,420 --> 00:00:53,730 quality. 16 00:00:53,730 --> 00:00:54,460 OK? 17 00:00:54,460 --> 00:00:57,067 And we'll try to explain a little bit more about that. 18 00:00:57,067 --> 00:00:58,900 We're going to have you be able to interpret 19 00:00:58,900 --> 00:01:01,850 a basic statistical process control chart. 20 00:01:01,850 --> 00:01:05,170 That's what the active learning experiment's about. 21 00:01:05,170 --> 00:01:07,510 We're going to talk about the difference between process 22 00:01:07,510 --> 00:01:09,820 control limits and specified control limits. 23 00:01:09,820 --> 00:01:12,670 And you'll be able to describe what a capable process is. 24 00:01:12,670 --> 00:01:16,900 OK, so Six Sigma is a strategy to improve process quality 25 00:01:16,900 --> 00:01:19,720 by identifying and eliminating defects, 26 00:01:19,720 --> 00:01:22,750 and defects in the broadest possible sense, not just 27 00:01:22,750 --> 00:01:25,660 defects in the size of a hole being drilled, 28 00:01:25,660 --> 00:01:30,630 but defect in something that you want to deliver your customers, 29 00:01:30,630 --> 00:01:31,440 not what they want. 30 00:01:31,440 --> 00:01:32,170 That's a defect. 31 00:01:32,170 --> 00:01:33,480 OK? 32 00:01:33,480 --> 00:01:35,860 It's a very data-driven approach. 33 00:01:35,860 --> 00:01:37,240 So if you've seen Lean-- in fact, 34 00:01:37,240 --> 00:01:39,448 one of the comments we got back on the feedback sheet 35 00:01:39,448 --> 00:01:41,490 is it's kind of a lot of touchy-feely stuff 36 00:01:41,490 --> 00:01:42,660 you're covering. 37 00:01:42,660 --> 00:01:45,600 And Lean is, in some ways, a lot of touchy-feely stuff, 38 00:01:45,600 --> 00:01:47,700 because of the importance of people. 39 00:01:47,700 --> 00:01:49,920 Six Sigma, if you read the literature 40 00:01:49,920 --> 00:01:52,110 on Six Sigma, at least the literature I've read, 41 00:01:52,110 --> 00:01:55,050 people are never mentioned. 42 00:01:55,050 --> 00:01:58,290 It's about a problem-solving process. 43 00:01:58,290 --> 00:02:02,410 And that's important, but it has its limitations. 44 00:02:02,410 --> 00:02:04,530 And the important thing is to know when to use it 45 00:02:04,530 --> 00:02:08,440 and how it fits in with a more holistic Lean approach. 46 00:02:08,440 --> 00:02:10,930 And it's a very structured implementation approach 47 00:02:10,930 --> 00:02:12,160 with certified experts. 48 00:02:12,160 --> 00:02:14,452 You probably have heard of black belts and green belts, 49 00:02:14,452 --> 00:02:15,580 and things like that, OK. 50 00:02:15,580 --> 00:02:18,490 So Six Sigma has this hierarchy of experts. 51 00:02:18,490 --> 00:02:21,880 And that's really not the mindset of Lean. 52 00:02:21,880 --> 00:02:25,840 The mindset of Lean is the experts, the coach, not 53 00:02:25,840 --> 00:02:27,460 the person who knows the most. 54 00:02:27,460 --> 00:02:29,877 It's the person who gets the most out of the people there. 55 00:02:29,877 --> 00:02:31,600 So there's some big cultural differences 56 00:02:31,600 --> 00:02:33,477 between Six Sigma and Lean. 57 00:02:33,477 --> 00:02:35,560 But they've come together, and the important thing 58 00:02:35,560 --> 00:02:38,950 is to try to understand that. 59 00:02:38,950 --> 00:02:40,883 Now, where does the word sigma come from? 60 00:02:40,883 --> 00:02:42,550 Well, this is a pretty analytical group, 61 00:02:42,550 --> 00:02:43,840 so you probably all know this. 62 00:02:43,840 --> 00:02:47,008 But a normal distribution, something like what you just 63 00:02:47,008 --> 00:02:49,300 saw with the M&M's-- although that wasn't quite normal, 64 00:02:49,300 --> 00:02:51,520 but it was somewhat close to that-- 65 00:02:51,520 --> 00:02:55,930 is defined by the standard deviation, which is sigma. 66 00:02:55,930 --> 00:02:59,590 And in this chart-- so this is one standard deviation 67 00:02:59,590 --> 00:03:02,410 from the mean, two standard deviations, and three. 68 00:03:02,410 --> 00:03:05,395 And Six Sigma would be six standard deviations. 69 00:03:05,395 --> 00:03:06,520 And just keep this in mind. 70 00:03:06,520 --> 00:03:08,500 We're going to deal with some Three Sigma in a little bit, 71 00:03:08,500 --> 00:03:09,125 coming up here. 72 00:03:09,125 --> 00:03:12,620 Three Sigma is 99.73% of the area 73 00:03:12,620 --> 00:03:16,250 under a normal distribution, would be within a Three Sigma 74 00:03:16,250 --> 00:03:16,750 band. 75 00:03:19,310 --> 00:03:19,810 OK. 76 00:03:19,810 --> 00:03:22,150 An important part of Six Sigma is defects. 77 00:03:22,150 --> 00:03:26,500 Because the goal of Six Sigma is to reduce the number of defects 78 00:03:26,500 --> 00:03:28,620 per million opportunities. 79 00:03:28,620 --> 00:03:31,980 And a defect is defined as any process output that does not 80 00:03:31,980 --> 00:03:33,390 meet the customer specifications, 81 00:03:33,390 --> 00:03:36,560 as I mentioned earlier, in a very broad sense. 82 00:03:36,560 --> 00:03:37,343 OK? 83 00:03:37,343 --> 00:03:38,760 And so you've got two things here. 84 00:03:38,760 --> 00:03:41,100 You've got opportunities and defects. 85 00:03:41,100 --> 00:03:44,080 So you want to reduce the defects, 86 00:03:44,080 --> 00:03:46,360 but you also, in many ways, you want 87 00:03:46,360 --> 00:03:49,130 to reduce the number of opportunities to have defects. 88 00:03:49,130 --> 00:03:51,940 OK, so reducing the number of hand-offs-- 89 00:03:51,940 --> 00:03:52,990 Sue mentioned the baton. 90 00:03:55,890 --> 00:03:59,910 Hand-offs between people are always 91 00:03:59,910 --> 00:04:04,590 a potential source of mistakes or errors or misunderstandings. 92 00:04:04,590 --> 00:04:07,020 Just think about how many miscommunications have you 93 00:04:07,020 --> 00:04:09,660 had with your boyfriend or girlfriend or parents 94 00:04:09,660 --> 00:04:11,400 or best friends. 95 00:04:11,400 --> 00:04:14,340 So hand-offs of any kind are something 96 00:04:14,340 --> 00:04:17,718 that you want to minimize. 97 00:04:17,718 --> 00:04:19,260 I just read some literature I'm going 98 00:04:19,260 --> 00:04:21,870 to share this with you, since we're in the aviation field-- 99 00:04:21,870 --> 00:04:24,270 I'm in the aviation field. 100 00:04:24,270 --> 00:04:28,370 The data from last year's commercial air transport travel 101 00:04:28,370 --> 00:04:35,600 in the US was there was 0.02 fatalities for every million 102 00:04:35,600 --> 00:04:40,010 boardings, 0.02. 103 00:04:40,010 --> 00:04:42,330 I think that's a nine sigma. 104 00:04:42,330 --> 00:04:44,060 Yeah, OK. 105 00:04:44,060 --> 00:04:46,220 So that gives you kind of a benchmark of something 106 00:04:46,220 --> 00:04:46,970 being really good. 107 00:04:46,970 --> 00:04:49,800 But how good is good? 108 00:04:49,800 --> 00:04:52,095 Some people say, you know, 99% is good. 109 00:04:52,095 --> 00:04:54,785 I mean, If my car starts 99% of the time, 110 00:04:54,785 --> 00:04:56,160 and it's an old car, that's good. 111 00:04:56,160 --> 00:04:57,930 But that's three days out of the year, 112 00:04:57,930 --> 00:04:59,190 I'm not going to get out of my garage, 113 00:04:59,190 --> 00:05:01,190 and I'm going to miss getting to my appointment. 114 00:05:01,190 --> 00:05:03,000 OK, that's not so good. 115 00:05:03,000 --> 00:05:05,130 So 99% you might think is good. 116 00:05:05,130 --> 00:05:09,090 But that's 20,000 lost article of mail per hour; 117 00:05:09,090 --> 00:05:11,160 unsafe drinking water for 15 minutes a day; 118 00:05:11,160 --> 00:05:13,980 in this audience, 5,000 incorrect surgical operations 119 00:05:13,980 --> 00:05:17,820 per week; two short or long landings 120 00:05:17,820 --> 00:05:21,690 at a major airport each day; 200,000 wrong drug 121 00:05:21,690 --> 00:05:23,160 prescriptions a year. 122 00:05:23,160 --> 00:05:26,640 That's not really very good, 99%. 123 00:05:26,640 --> 00:05:28,200 This is much better, Six Sigma. 124 00:05:28,200 --> 00:05:30,180 This is what you're looking for. 125 00:05:30,180 --> 00:05:31,710 OK? 126 00:05:31,710 --> 00:05:35,010 1.7 incorrect operations per week-- even that's too much. 127 00:05:35,010 --> 00:05:38,830 But boy, that's a lot better than 5,000. 128 00:05:38,830 --> 00:05:41,860 If you're one of those 1.7 people, 129 00:05:41,860 --> 00:05:43,540 you don't care what sigma it is. 130 00:05:43,540 --> 00:05:44,380 You just got-- 131 00:05:44,380 --> 00:05:46,930 [LAUGHTER] 132 00:05:46,930 --> 00:05:47,793 Same here. 133 00:05:47,793 --> 00:05:49,210 And as I mentioned, airline travel 134 00:05:49,210 --> 00:05:52,720 is nine sigma for safety. 135 00:05:52,720 --> 00:05:56,560 OK, so a basic tool in Six Sigma is 136 00:05:56,560 --> 00:05:58,940 called the statistical process control chart. 137 00:05:58,940 --> 00:06:00,580 So a chart shows-- so it's a sequence. 138 00:06:00,580 --> 00:06:02,190 This is a time chart. 139 00:06:02,190 --> 00:06:03,940 And we're going to build one just shortly. 140 00:06:03,940 --> 00:06:05,607 It's time, and this is some measurement. 141 00:06:05,607 --> 00:06:08,800 And this shows how stable it is, how much it's 142 00:06:08,800 --> 00:06:10,780 tending to be around the center of the chart. 143 00:06:10,780 --> 00:06:13,260 It's a very common chart. 144 00:06:13,260 --> 00:06:16,360 And in both fields, if you go into a production facility, 145 00:06:16,360 --> 00:06:20,170 an engineering, production facility, now the operator 146 00:06:20,170 --> 00:06:23,830 will be showing you their control chart. 147 00:06:23,830 --> 00:06:27,190 I couldn't believe, when I first went in and saw the machinist 148 00:06:27,190 --> 00:06:32,270 proudly or the machine operators showing me their control chart. 149 00:06:32,270 --> 00:06:35,150 I see control charts all through health care. 150 00:06:35,150 --> 00:06:37,885 So let's make a control chart. 151 00:06:37,885 --> 00:06:39,760 OK, so we're going to do a little experiment, 152 00:06:39,760 --> 00:06:42,412 a little thing here. 153 00:06:42,412 --> 00:06:44,620 We're going to put it in the framework of a pharmacy. 154 00:06:44,620 --> 00:06:47,177 For those of you who are not in the framework of health care, 155 00:06:47,177 --> 00:06:48,760 you could think of another application 156 00:06:48,760 --> 00:06:50,230 where you're getting something. 157 00:06:50,230 --> 00:06:54,190 But our pharmacy here dispenses a medicine called a White Bean 158 00:06:54,190 --> 00:06:56,380 Medicine. 159 00:06:56,380 --> 00:07:00,350 And we get it in bulk from two different suppliers. 160 00:07:00,350 --> 00:07:04,900 This is our brand name supplier, Goya. 161 00:07:04,900 --> 00:07:06,690 And it appears to come from Spain. 162 00:07:06,690 --> 00:07:09,190 It's a little hard to tell what the original source is here. 163 00:07:09,190 --> 00:07:11,245 But at the end of it, there's some Spanish, 164 00:07:11,245 --> 00:07:13,450 and so I presume it comes from Spain. 165 00:07:13,450 --> 00:07:15,460 And this is our generic supplier, 166 00:07:15,460 --> 00:07:19,893 [? Shaw's. ?] So we get these two suppliers. 167 00:07:19,893 --> 00:07:22,060 And if you read the back of the [? Shaw's-- ?] well, 168 00:07:22,060 --> 00:07:24,477 you can read the back of the label, if you need to, later. 169 00:07:24,477 --> 00:07:27,520 So our pharmacy gets it and puts it in these bins. 170 00:07:27,520 --> 00:07:31,760 And then they've decided to measure dosages by volume. 171 00:07:31,760 --> 00:07:34,050 Now, usually a dosage is by weight. 172 00:07:34,050 --> 00:07:35,200 OK? 173 00:07:35,200 --> 00:07:39,437 But it's much easier just to scoop this up and fill it 174 00:07:39,437 --> 00:07:41,770 than to scoop it up and fill it and weigh it, and so on. 175 00:07:41,770 --> 00:07:43,600 So we're doing it by volume. 176 00:07:43,600 --> 00:07:47,830 And what we want to do is to track our pharmacy dispensary, 177 00:07:47,830 --> 00:07:53,270 to be sure that we're getting consistent weight. 178 00:07:53,270 --> 00:07:55,490 And so our pharmacy-- 179 00:07:55,490 --> 00:07:59,720 this is not a critical medicine, but it's widely 180 00:07:59,720 --> 00:08:00,970 used in our facility. 181 00:08:00,970 --> 00:08:03,360 So we dispense a lot of it each day. 182 00:08:03,360 --> 00:08:04,970 And so what we've done in our pharmacy 183 00:08:04,970 --> 00:08:08,990 is our pharmacy is going to take three samples of this medicine 184 00:08:08,990 --> 00:08:11,238 each day and weigh it. 185 00:08:11,238 --> 00:08:13,280 And the next day, it will take three more samples 186 00:08:13,280 --> 00:08:13,820 and weigh it. 187 00:08:13,820 --> 00:08:16,190 And we're going to build this chart to see what it's like. 188 00:08:16,190 --> 00:08:17,990 And the pharmacy has been very cooperative. 189 00:08:17,990 --> 00:08:20,970 They've given us 20 samples, that are already on your table. 190 00:08:20,970 --> 00:08:23,000 OK? 191 00:08:23,000 --> 00:08:24,750 And then I'm going to tell you what to do. 192 00:08:24,750 --> 00:08:26,107 We're going to enter this data. 193 00:08:26,107 --> 00:08:27,440 So we've got a three-cup sample. 194 00:08:27,440 --> 00:08:30,380 We're going to enter data into control charts-- one 195 00:08:30,380 --> 00:08:33,350 for the average weight and one for the range of the weight 196 00:08:33,350 --> 00:08:35,270 of the three cups. 197 00:08:35,270 --> 00:08:36,860 And we're going to use it for 20 days 198 00:08:36,860 --> 00:08:39,326 to establish our process capability. 199 00:08:39,326 --> 00:08:40,909 So we're going to measure for 20 days, 200 00:08:40,909 --> 00:08:42,284 and then that's going to tell us, 201 00:08:42,284 --> 00:08:43,490 do we have a stable process? 202 00:08:43,490 --> 00:08:46,078 And we can monitor after that. 203 00:08:46,078 --> 00:08:47,620 And so that's what we're going to do. 204 00:08:47,620 --> 00:08:52,130 So our first phase is now each of you are going to weigh this. 205 00:08:52,130 --> 00:08:54,080 You each have four days. 206 00:08:54,080 --> 00:08:57,560 In the center of your table, you have 12 cups. 207 00:08:57,560 --> 00:09:00,950 And the cups, there are three cups for each day. 208 00:09:00,950 --> 00:09:03,230 It says right on there, this is day 16. 209 00:09:03,230 --> 00:09:05,120 There are three day 16 cups. 210 00:09:05,120 --> 00:09:06,870 And you have a sheet. 211 00:09:06,870 --> 00:09:09,810 And your sheet says, here's day 16. 212 00:09:09,810 --> 00:09:13,320 So you measure cup A, B, C. You compute the average-- 213 00:09:13,320 --> 00:09:15,600 we can get some calculators, if you need it-- 214 00:09:15,600 --> 00:09:17,860 the max and the min and then the range, 215 00:09:17,860 --> 00:09:19,320 which is the max minus the min. 216 00:09:19,320 --> 00:09:22,290 And then you bring that up here to data central. 217 00:09:22,290 --> 00:09:24,720 And we'll record it. 218 00:09:24,720 --> 00:09:26,550 And you have a scale. 219 00:09:26,550 --> 00:09:27,900 And you want to use grams. 220 00:09:27,900 --> 00:09:31,780 This is in pounds and kilograms. 221 00:09:31,780 --> 00:09:34,840 You want to use the kilograms scale to get it in grams. 222 00:09:34,840 --> 00:09:36,280 And don't use this now. 223 00:09:36,280 --> 00:09:38,738 But when we're all finished, we'll dump the beans in there. 224 00:09:38,738 --> 00:09:40,720 But don't do it now, because we may 225 00:09:40,720 --> 00:09:43,160 need to go back and do some second measurements. 226 00:09:43,160 --> 00:09:47,380 OK, so please now start at your tables measuring these samples, 227 00:09:47,380 --> 00:09:50,480 filling out the chart, and bringing it up here. 228 00:09:50,480 --> 00:09:52,956 [INTERPOSING VOICES] 229 00:10:06,550 --> 00:10:07,420 STUDENT: 98. 230 00:10:07,420 --> 00:10:08,462 STUDENT: Oh, my goodness. 231 00:10:12,088 --> 00:10:12,880 STUDENT: Excellent. 232 00:10:12,880 --> 00:10:15,087 Very [? well done. ?] 233 00:10:15,087 --> 00:10:15,670 PROFESSOR: OK. 234 00:10:15,670 --> 00:10:18,400 So we have all the data. 235 00:10:18,400 --> 00:10:24,660 And my helper over here is going to enter it day 236 00:10:24,660 --> 00:10:27,690 by day, so it unfolds like you would see a normal control 237 00:10:27,690 --> 00:10:29,940 chart, if you're watching it day by day. 238 00:10:29,940 --> 00:10:32,400 So he's now on day five. 239 00:10:32,400 --> 00:10:37,080 Now, what we have-- this is the mean average. 240 00:10:37,080 --> 00:10:39,330 This is the average, and this is the range. 241 00:10:39,330 --> 00:10:40,890 And this is the mean of the average, 242 00:10:40,890 --> 00:10:42,640 and this is the mean of the range. 243 00:10:42,640 --> 00:10:43,530 OK? 244 00:10:43,530 --> 00:10:46,590 And then what we have here are these red lines 245 00:10:46,590 --> 00:10:49,060 are the Three Sigma lines. 246 00:10:49,060 --> 00:10:51,240 Remember, on a normal distribution, 247 00:10:51,240 --> 00:10:52,860 I said Three Sigma. 248 00:10:52,860 --> 00:10:57,600 So the way these control charts work is you get this data. 249 00:10:57,600 --> 00:11:01,080 And it's been established that the processes-- 250 00:11:01,080 --> 00:11:03,630 you look at a three sigma deviation. 251 00:11:03,630 --> 00:11:07,390 And as long as you're within that three sigma deviation, 252 00:11:07,390 --> 00:11:09,510 your process is under control. 253 00:11:09,510 --> 00:11:13,840 OK, that's just sort of what's evolved. 254 00:11:13,840 --> 00:11:16,390 And of course, these red lines will 255 00:11:16,390 --> 00:11:18,940 change as he enters the data, because he's 256 00:11:18,940 --> 00:11:21,210 getting more and more samples. 257 00:11:21,210 --> 00:11:23,840 So this is like watching the election returns 258 00:11:23,840 --> 00:11:24,680 on election eve. 259 00:11:24,680 --> 00:11:25,610 [LAUGHTER] 260 00:11:25,610 --> 00:11:27,470 STUDENT: We're all [? down here. ?] 261 00:11:27,470 --> 00:11:29,446 PROFESSOR: OK. 262 00:11:29,446 --> 00:11:32,750 STUDENT: [INAUDIBLE] 263 00:11:32,750 --> 00:11:34,372 PROFESSOR: So we're up to day 10. 264 00:11:34,372 --> 00:11:37,180 STUDENT: [INAUDIBLE] 265 00:11:37,180 --> 00:11:39,430 PROFESSOR: OK, and what's going on in the background-- 266 00:11:39,430 --> 00:11:40,960 one of our colleagues at Cal Poly 267 00:11:40,960 --> 00:11:42,260 put this spreadsheet together. 268 00:11:42,260 --> 00:11:46,660 So this shows the [? app, ?] the mean, what's called x bar. 269 00:11:46,660 --> 00:11:49,030 This is x bar, and this is r bar. 270 00:11:49,030 --> 00:11:54,130 X is the average, and r is the range. 271 00:11:54,130 --> 00:11:56,650 And then we've got the upper and lower control limits. 272 00:11:56,650 --> 00:11:59,050 These are called upper control limits and lower control 273 00:11:59,050 --> 00:11:59,695 limits. 274 00:11:59,695 --> 00:12:02,260 We've got the upper and lower control limits for the average 275 00:12:02,260 --> 00:12:03,040 and for the range. 276 00:12:05,920 --> 00:12:10,240 And usually, the way this field works is 20 samples 277 00:12:10,240 --> 00:12:12,940 is considered a good enough baseline 278 00:12:12,940 --> 00:12:14,950 to establish the process capability. 279 00:12:14,950 --> 00:12:17,890 So oh. 280 00:12:17,890 --> 00:12:19,350 STUDENT: Oh-oh. 281 00:12:19,350 --> 00:12:20,495 PROFESSOR: [INAUDIBLE] 282 00:12:20,495 --> 00:12:22,248 STUDENT: I'm double-checking the data. 283 00:12:22,248 --> 00:12:23,790 PROFESSOR: Can you double-check that? 284 00:12:23,790 --> 00:12:24,730 STUDENT: Yeah, I'm double-checking. 285 00:12:24,730 --> 00:12:26,210 [INAUDIBLE] what's on the sheet? 286 00:12:26,210 --> 00:12:27,460 We may have to go [INAUDIBLE]. 287 00:12:27,460 --> 00:12:29,168 PROFESSOR: OK, well, let's finish it off. 288 00:12:29,168 --> 00:12:30,030 Let's keep going. 289 00:12:30,030 --> 00:12:33,580 STUDENT: OK, so that's it for 20 data points. 290 00:12:33,580 --> 00:12:36,640 PROFESSOR: OK, so something's obviously funny with day 15. 291 00:12:36,640 --> 00:12:38,796 So when you have some funny data, what do you do? 292 00:12:38,796 --> 00:12:40,740 [INTERPOSING VOICES] 293 00:12:40,740 --> 00:12:43,950 PROFESSOR: Go to the [? gemba. ?] So who has 15? 294 00:12:43,950 --> 00:12:45,755 OK, so let's go take a look at 15. 295 00:12:45,755 --> 00:12:46,630 What's going on here? 296 00:12:46,630 --> 00:12:49,300 STUDENT: [INAUDIBLE] Which sample is that one? 297 00:12:49,300 --> 00:12:51,396 PROFESSOR: Can you weigh it again? 298 00:12:51,396 --> 00:12:53,550 STUDENT: It's cup C. 299 00:12:53,550 --> 00:12:56,368 PROFESSOR: It doesn't look quite full, does it? 300 00:12:56,368 --> 00:12:58,160 Why don't you put some more beans in there. 301 00:12:58,160 --> 00:13:01,611 [LAUGHTER] 302 00:13:04,080 --> 00:13:07,398 Just fill it up kind of like the rest of these. 303 00:13:07,398 --> 00:13:11,206 [INTERPOSING VOICES] 304 00:13:12,160 --> 00:13:13,330 PROFESSOR: What's now? 305 00:13:13,330 --> 00:13:14,800 STUDENT: 73. 306 00:13:14,800 --> 00:13:19,240 PROFESSOR: OK, so now, on day 15, 307 00:13:19,240 --> 00:13:21,190 something happened in the pharmacy. 308 00:13:21,190 --> 00:13:26,070 And all three cops are shy by 1/4 inch, 1/8 inch, 309 00:13:26,070 --> 00:13:27,370 something like that. 310 00:13:27,370 --> 00:13:30,703 So what we would probably do is go back to the pharmacy records 311 00:13:30,703 --> 00:13:31,620 and see what happened. 312 00:13:31,620 --> 00:13:33,078 And we might find out, for example, 313 00:13:33,078 --> 00:13:35,280 that some employee had been ill that day, 314 00:13:35,280 --> 00:13:36,720 and they'd hired a temp. 315 00:13:36,720 --> 00:13:39,360 Or a fill-in wasn't cross-trained properly, 316 00:13:39,360 --> 00:13:42,150 and they thought this was a full cup, a full dosage. 317 00:13:42,150 --> 00:13:43,530 So we can eliminate. 318 00:13:43,530 --> 00:13:45,000 Let's delete that point. 319 00:13:45,000 --> 00:13:48,540 We know that that point is no good, because for certain-- 320 00:13:48,540 --> 00:13:49,170 we can see it. 321 00:13:49,170 --> 00:13:50,520 We can verify it's no good. 322 00:13:50,520 --> 00:13:51,520 So-- 323 00:13:51,520 --> 00:13:53,270 STUDENT: What happens if I [? blank it? ?] 324 00:13:53,270 --> 00:13:54,230 PROFESSOR: OK, that's fine. 325 00:13:54,230 --> 00:13:54,980 STUDENT: Going to zero? 326 00:13:54,980 --> 00:13:55,235 PROFESSOR: OK. 327 00:13:55,235 --> 00:13:56,320 STUDENT: Is it going to go to zero? 328 00:13:56,320 --> 00:13:56,870 PROFESSOR: No, that's fine. 329 00:13:56,870 --> 00:13:57,050 STUDENT: [INAUDIBLE] 330 00:13:57,050 --> 00:13:57,830 PROFESSOR: No, no, that's fine. 331 00:13:57,830 --> 00:13:58,770 It's fine. 332 00:13:58,770 --> 00:13:59,540 Yeah. 333 00:13:59,540 --> 00:14:00,200 Thank you, sir. 334 00:14:00,200 --> 00:14:00,830 STUDENT: All right. 335 00:14:00,830 --> 00:14:01,550 PROFESSOR: Yep. 336 00:14:01,550 --> 00:14:05,780 OK, so now we've got our process capability established. 337 00:14:05,780 --> 00:14:09,500 We've got 20 days minus the 1 bogus day. 338 00:14:09,500 --> 00:14:13,340 And our upper and lower control limits are-- 339 00:14:13,340 --> 00:14:17,540 we know now, for our process, it's 75 and 69. 340 00:14:17,540 --> 00:14:19,630 And we got them for the range. 341 00:14:19,630 --> 00:14:21,110 OK? 342 00:14:21,110 --> 00:14:21,660 So good. 343 00:14:21,660 --> 00:14:22,160 Thank you. 344 00:14:22,160 --> 00:14:25,307 So that's how we would build a control chart. 345 00:14:25,307 --> 00:14:27,390 So now we're going to do a little bit more slides. 346 00:14:27,390 --> 00:14:28,220 And then we're going to come back 347 00:14:28,220 --> 00:14:29,762 to see how we use this control chart. 348 00:14:29,762 --> 00:14:34,310 So I'm going to change now from the basic control-charting 349 00:14:34,310 --> 00:14:38,600 to a little bit about the basic Six Sigma method, called DMAIC. 350 00:14:38,600 --> 00:14:40,790 And DMAIC is another Deming cycle. 351 00:14:40,790 --> 00:14:45,000 It's like plan, do, study, act, except it's slightly different. 352 00:14:45,000 --> 00:14:48,980 It's called Define, Measure, Analyze, Improve, and Control. 353 00:14:48,980 --> 00:14:52,600 And so DMAIC is the easy thing to remember. 354 00:14:52,600 --> 00:14:57,600 So if you're doing a Six Sigma intervention, 355 00:14:57,600 --> 00:15:00,643 like in our pharmacy, we would have to define, 356 00:15:00,643 --> 00:15:01,560 who are the customers? 357 00:15:01,560 --> 00:15:02,893 And what are their requirements? 358 00:15:02,893 --> 00:15:05,460 So in the case of the pharmacy, the customers 359 00:15:05,460 --> 00:15:08,740 would be whoever has ordered the medicine, 360 00:15:08,740 --> 00:15:10,705 and they have some requirements. 361 00:15:10,705 --> 00:15:13,080 And then we want to know what the key characteristics are 362 00:15:13,080 --> 00:15:14,163 important to the customer. 363 00:15:14,163 --> 00:15:15,540 Well, the customer for medication 364 00:15:15,540 --> 00:15:17,520 is really interested in weight. 365 00:15:17,520 --> 00:15:20,670 They don't care too much what it looks like. 366 00:15:20,670 --> 00:15:22,470 They really don't care how it's packed 367 00:15:22,470 --> 00:15:24,610 in the cup, things like that. 368 00:15:24,610 --> 00:15:26,340 But they're really interested in weight. 369 00:15:26,340 --> 00:15:28,620 And of course, we're using volume, 370 00:15:28,620 --> 00:15:30,570 because it's a little quicker, but they're 371 00:15:30,570 --> 00:15:31,445 interested in weight. 372 00:15:31,445 --> 00:15:33,580 So OK, then you measure. 373 00:15:33,580 --> 00:15:37,260 So once we know the key characteristics, then 374 00:15:37,260 --> 00:15:40,290 what are the key input and output characteristics? 375 00:15:40,290 --> 00:15:42,490 And you want to verify the measurement system. 376 00:15:42,490 --> 00:15:44,460 So here are output characteristics, 377 00:15:44,460 --> 00:15:45,930 is the weight of the cup. 378 00:15:45,930 --> 00:15:47,490 That's what we're measuring. 379 00:15:47,490 --> 00:15:49,245 Later on, we might get more sophisticated 380 00:15:49,245 --> 00:15:51,120 and measure the input characteristics. 381 00:15:51,120 --> 00:15:55,352 Like, we might weigh these two samples, 382 00:15:55,352 --> 00:15:56,310 or something like that. 383 00:15:56,310 --> 00:15:57,685 But right now, we're just focused 384 00:15:57,685 --> 00:15:59,250 on the output characteristics. 385 00:15:59,250 --> 00:16:00,990 And we collect the data and establish 386 00:16:00,990 --> 00:16:02,130 the baseline performance. 387 00:16:02,130 --> 00:16:04,540 That's what we've done. 388 00:16:04,540 --> 00:16:07,730 OK, and then we've done that so far. 389 00:16:07,730 --> 00:16:10,120 Then we're going to look at the raw data. 390 00:16:10,120 --> 00:16:11,770 We convert the raw data information 391 00:16:11,770 --> 00:16:13,570 and get insights into the process. 392 00:16:13,570 --> 00:16:15,820 Well, we already found out we had something 393 00:16:15,820 --> 00:16:18,130 wrong with poor Molly. 394 00:16:18,130 --> 00:16:19,780 On day 15, she got the flu. 395 00:16:19,780 --> 00:16:23,410 And we had a good temp, but we didn't train the temp. 396 00:16:23,410 --> 00:16:26,630 So we've learned now that training is important 397 00:16:26,630 --> 00:16:29,570 so we don't dispense the wrong medication. 398 00:16:29,570 --> 00:16:32,450 OK, so we've improved. 399 00:16:32,450 --> 00:16:34,360 We've developed a solution. 400 00:16:34,360 --> 00:16:38,040 And now we want to put our process under process control. 401 00:16:38,040 --> 00:16:39,580 I [? mean, ?] under process control, 402 00:16:39,580 --> 00:16:43,060 we're going to now monitor day by day 403 00:16:43,060 --> 00:16:47,650 the medication dispensing and see if it stays within control. 404 00:16:47,650 --> 00:16:49,000 That's DMAIC. 405 00:16:49,000 --> 00:16:51,050 So here's a simple example. 406 00:16:51,050 --> 00:16:52,400 You have a process. 407 00:16:52,400 --> 00:16:57,390 So you have a defined process with an input and an output. 408 00:16:57,390 --> 00:17:03,000 You measure the weight with a scale. 409 00:17:03,000 --> 00:17:05,310 You analyze the data with the control chart. 410 00:17:08,368 --> 00:17:09,910 Then you want to improve the process. 411 00:17:09,910 --> 00:17:13,869 We just talked about training is one improvement. 412 00:17:13,869 --> 00:17:17,720 And then you put it under process control. 413 00:17:17,720 --> 00:17:21,160 So that's the basic Six Sigma cycle. 414 00:17:21,160 --> 00:17:23,643 This is one area where I said, in Lean, when you came out 415 00:17:23,643 --> 00:17:25,060 of this course, you could probably 416 00:17:25,060 --> 00:17:27,465 pick up any book in Lean, and you could go for it. 417 00:17:27,465 --> 00:17:29,590 If you pick up Six Sigma books, they do a deep dive 418 00:17:29,590 --> 00:17:31,030 into a lot of math. 419 00:17:31,030 --> 00:17:34,130 So it's a lot of statistical mathematics. 420 00:17:34,130 --> 00:17:37,510 So you may need more than what we've taught you, 421 00:17:37,510 --> 00:17:41,080 because we're not focusing on math. 422 00:17:41,080 --> 00:17:41,920 OK. 423 00:17:41,920 --> 00:17:44,780 It's easy to see in a process control application like this. 424 00:17:44,780 --> 00:17:48,250 It may get more complicated if you're looking at something 425 00:17:48,250 --> 00:17:50,260 that's not quite so visual. 426 00:17:50,260 --> 00:17:52,780 But anyway, that's the way it is. 427 00:17:52,780 --> 00:17:56,590 OK, now there are certain kind of variation in the process. 428 00:17:56,590 --> 00:18:00,220 And actually, our control chart was a good example of that. 429 00:18:00,220 --> 00:18:02,710 There's what's called Common Cause Variation. 430 00:18:02,710 --> 00:18:05,260 And this is just the sort of randomness 431 00:18:05,260 --> 00:18:11,380 in the system, the things that affect all the samples. 432 00:18:11,380 --> 00:18:13,792 And then you have Special Cause Variation. 433 00:18:13,792 --> 00:18:15,250 And we just had an example of that. 434 00:18:15,250 --> 00:18:19,320 Common Cause was the different variations of 19 435 00:18:19,320 --> 00:18:20,050 of the 20 days. 436 00:18:20,050 --> 00:18:22,780 And then we had one case which was a Special Cause Variation. 437 00:18:22,780 --> 00:18:24,430 We sort of went out of bounds. 438 00:18:24,430 --> 00:18:27,670 In that, we could go in, and we could intervene fairly quickly 439 00:18:27,670 --> 00:18:28,600 and correct. 440 00:18:28,600 --> 00:18:30,795 To get our Comma Cause Variation down, 441 00:18:30,795 --> 00:18:32,920 we're going to have to do something different here. 442 00:18:32,920 --> 00:18:35,650 We're going to have to be much more careful about how we fill 443 00:18:35,650 --> 00:18:37,330 the cups, and stuff like that. 444 00:18:37,330 --> 00:18:39,430 We might have to have more careful procedures. 445 00:18:39,430 --> 00:18:42,250 But the Special Cause Variation, we 446 00:18:42,250 --> 00:18:43,720 want to drive out of the system. 447 00:18:43,720 --> 00:18:45,640 And the control charts give you a way 448 00:18:45,640 --> 00:18:48,630 to quickly see the differences between these two. 449 00:18:48,630 --> 00:18:50,210 So let's look at an example. 450 00:18:50,210 --> 00:18:52,990 This is actual data for patient falls. 451 00:18:52,990 --> 00:18:55,290 I've forgotten the facility. 452 00:18:55,290 --> 00:18:57,660 Anyway, the reference is just off the back of the bottom 453 00:18:57,660 --> 00:18:58,950 of the screen here. 454 00:18:58,950 --> 00:19:00,990 But this is patient falls. 455 00:19:00,990 --> 00:19:06,340 And we've heard sue mention this morning how patient falls are-- 456 00:19:06,340 --> 00:19:07,270 you had it on a slide. 457 00:19:07,270 --> 00:19:09,610 It was something about there being a source of waste, 458 00:19:09,610 --> 00:19:11,545 or something. 459 00:19:11,545 --> 00:19:14,170 STUDENT: [? Patient ?] falls are considered [? to have ?] never 460 00:19:14,170 --> 00:19:15,070 have been something [INAUDIBLE]. 461 00:19:15,070 --> 00:19:15,610 PROFESSOR: [? They've ?] never [INAUDIBLE].. 462 00:19:15,610 --> 00:19:17,590 Great-- never [? met, ?] OK. 463 00:19:17,590 --> 00:19:22,450 And here we've got, in this facility, this is by month. 464 00:19:22,450 --> 00:19:27,510 And we got an average of about six patient falls a month. 465 00:19:27,510 --> 00:19:29,590 And they've been charting it. 466 00:19:29,590 --> 00:19:32,020 And we have an upper control limit of about eight 467 00:19:32,020 --> 00:19:34,380 and a lower control limit of about four. 468 00:19:34,380 --> 00:19:36,860 Now, how many patient falls would you want? 469 00:19:36,860 --> 00:19:37,840 STUDENT: [INAUDIBLE] 470 00:19:37,840 --> 00:19:38,507 PROFESSOR: None. 471 00:19:38,507 --> 00:19:42,220 OK, that's a customer-specified limit. 472 00:19:42,220 --> 00:19:44,980 There's a difference between your process limit. 473 00:19:44,980 --> 00:19:46,840 This is your as is process. 474 00:19:46,840 --> 00:19:48,250 In this case, your process is not 475 00:19:48,250 --> 00:19:51,163 very capable in the eyes of the customer. 476 00:19:51,163 --> 00:19:53,330 So we're going to get into this towards the back end 477 00:19:53,330 --> 00:19:55,580 of the module about the difference between the process 478 00:19:55,580 --> 00:19:58,610 capability, the upper and lower control limits, 479 00:19:58,610 --> 00:20:00,720 and the upper and lower specified limits. 480 00:20:00,720 --> 00:20:02,840 So I just introduce it here as a teaser 481 00:20:02,840 --> 00:20:04,910 so you don't fall asleep. 482 00:20:04,910 --> 00:20:10,240 OK, so now, in this facility, they tracked it for 20 days. 483 00:20:10,240 --> 00:20:11,550 And then they kept tracking it. 484 00:20:11,550 --> 00:20:13,530 And then something starts going kind of bad 485 00:20:13,530 --> 00:20:17,255 here towards the end of the around 28th month. 486 00:20:17,255 --> 00:20:19,130 We've got something that's kind of getting up 487 00:20:19,130 --> 00:20:19,910 to the control limits. 488 00:20:19,910 --> 00:20:21,368 And then we've got something that's 489 00:20:21,368 --> 00:20:23,130 going outside the control limits. 490 00:20:23,130 --> 00:20:27,080 And this is when the red flags start saying, OK, we're 491 00:20:27,080 --> 00:20:30,380 outside the normal bounds we consider acceptable. 492 00:20:30,380 --> 00:20:32,720 We better go investigate what's happening here. 493 00:20:32,720 --> 00:20:37,500 And that's what we did with our pharmacy thing. 494 00:20:37,500 --> 00:20:41,560 So now what we're going to do is continue our exercise. 495 00:20:41,560 --> 00:20:43,950 And the first thing I'd like you to do is take the cups. 496 00:20:43,950 --> 00:20:45,840 And there's a white tray in here. 497 00:20:45,840 --> 00:20:49,020 And dump all the beans that you have into there. 498 00:20:49,020 --> 00:20:52,710 And just nest the cups back together. 499 00:20:52,710 --> 00:20:56,685 And then the pharmacy's going to bring you a new batch of cups. 500 00:20:56,685 --> 00:21:00,600 So let's get these out of the way first. 501 00:21:00,600 --> 00:21:03,870 OK, so while we're getting some help here, 502 00:21:03,870 --> 00:21:07,050 what's going to happen is, each of the stations on their easel 503 00:21:07,050 --> 00:21:11,380 has a chart for the average and the range. 504 00:21:11,380 --> 00:21:17,970 And what I want you to do is to come put a scale on this 505 00:21:17,970 --> 00:21:19,440 where you're going to put the-- 506 00:21:19,440 --> 00:21:21,090 OK, the upper control limits, we don't 507 00:21:21,090 --> 00:21:24,840 have to be accurate to the second decimal place here. 508 00:21:24,840 --> 00:21:27,900 It's about 75.4 and 69.5. 509 00:21:27,900 --> 00:21:30,670 So we want to do something like this. 510 00:21:30,670 --> 00:21:32,580 First of all, let's get the mean. 511 00:21:32,580 --> 00:21:35,790 I'll just do one, and I'll let the other tables do theirs. 512 00:21:35,790 --> 00:21:40,480 So this is 72.5. 513 00:21:40,480 --> 00:21:42,070 That's the mean. 514 00:21:42,070 --> 00:21:48,370 And then the upper control limit is about 75.4. 515 00:21:48,370 --> 00:21:49,720 So we'll just put that here. 516 00:21:55,000 --> 00:22:05,380 And the lower control limit is 69.5. 517 00:22:05,380 --> 00:22:08,440 OK, so you're going to want to finish that. 518 00:22:08,440 --> 00:22:10,990 This is for the average, and this is for the range. 519 00:22:10,990 --> 00:22:13,880 And I'll let each team finish their own. 520 00:22:13,880 --> 00:22:16,640 And you can come up here and get the data if you need to. 521 00:22:16,640 --> 00:22:17,540 And then-- 522 00:22:17,540 --> 00:22:18,400 STUDENT: [INAUDIBLE] 523 00:22:18,400 --> 00:22:19,900 PROFESSOR: So we want to draw the control 524 00:22:19,900 --> 00:22:20,860 limits on our chart. 525 00:22:20,860 --> 00:22:22,360 That's now fixed. 526 00:22:22,360 --> 00:22:24,340 That's our process capability. 527 00:22:24,340 --> 00:22:26,990 And now we're going to start moderating it 528 00:22:26,990 --> 00:22:31,030 for days 21, 22, 23, 24, but you're each going to do it. 529 00:22:31,030 --> 00:22:33,250 You're going to do 21, 22, 23, 24. 530 00:22:33,250 --> 00:22:34,840 You're going to do 21, 22, 23, 24. 531 00:22:34,840 --> 00:22:36,630 OK? 532 00:22:36,630 --> 00:22:40,020 And you're going to measure three samples each day 533 00:22:40,020 --> 00:22:42,810 and plot the data on your control charts. 534 00:22:42,810 --> 00:22:46,980 So whatever the data is for the average on day 21, 535 00:22:46,980 --> 00:22:48,180 you put here in the range. 536 00:22:48,180 --> 00:22:49,920 On day 21, you put here. 537 00:22:49,920 --> 00:22:50,698 OK? 538 00:22:50,698 --> 00:22:51,990 And you're going to look at it. 539 00:22:51,990 --> 00:22:54,073 And you're going to see whether your process stays 540 00:22:54,073 --> 00:22:57,960 under control or whether it-- 541 00:22:57,960 --> 00:23:00,900 is it doing this, or is it doing that, or that, 542 00:23:00,900 --> 00:23:01,770 or something else? 543 00:23:01,770 --> 00:23:05,190 And if it starts going out of control, 544 00:23:05,190 --> 00:23:07,570 once you've decided it's out of control-- 545 00:23:07,570 --> 00:23:09,093 you may have to watch a little bit-- 546 00:23:09,093 --> 00:23:10,635 then it's time to stop to investigate 547 00:23:10,635 --> 00:23:11,858 in what the root cause is. 548 00:23:11,858 --> 00:23:13,650 Just like we did here, we went to the table 549 00:23:13,650 --> 00:23:16,350 and found out the root cause was it wasn't filled. 550 00:23:16,350 --> 00:23:19,870 But we want you to do this in a structured way. 551 00:23:19,870 --> 00:23:20,868 OK? 552 00:23:20,868 --> 00:23:22,410 The tendency you're all going to have 553 00:23:22,410 --> 00:23:25,110 is to say that I know what the root cause is. 554 00:23:25,110 --> 00:23:26,852 Somebody didn't fill it. 555 00:23:26,852 --> 00:23:27,810 But maybe or maybe not. 556 00:23:27,810 --> 00:23:30,660 So we're going to ask you to use a fishbone diagram that Analisa 557 00:23:30,660 --> 00:23:31,950 just told you about. 558 00:23:31,950 --> 00:23:33,720 And that's on the second page. 559 00:23:33,720 --> 00:23:36,390 And start thinking about all the reasons 560 00:23:36,390 --> 00:23:40,840 that the behavior you're observing might be happening. 561 00:23:40,840 --> 00:23:43,448 For instance, could it be that Aubrey, 562 00:23:43,448 --> 00:23:45,990 who was thinking about something else, didn't weigh it right? 563 00:23:45,990 --> 00:23:50,970 Maybe it's a measurement thing, so that would be a personnel. 564 00:23:50,970 --> 00:23:56,330 Or maybe you got a crummy weighing scale. 565 00:23:56,330 --> 00:23:57,280 That's a machine. 566 00:23:57,280 --> 00:23:59,470 So start filling out all the things you can. 567 00:23:59,470 --> 00:24:01,510 And then, when you get those filled out, 568 00:24:01,510 --> 00:24:03,730 then start looking at them systematically. 569 00:24:03,730 --> 00:24:05,420 Which one do you think is most likely? 570 00:24:05,420 --> 00:24:06,340 OK? 571 00:24:06,340 --> 00:24:08,090 So it's a detective operation. 572 00:24:08,090 --> 00:24:10,900 So now you're going to actually do a root cause analysis. 573 00:24:10,900 --> 00:24:11,900 OK? 574 00:24:11,900 --> 00:24:12,940 OK. 575 00:24:12,940 --> 00:24:15,160 Does anybody have any questions about-- yeah? 576 00:24:15,160 --> 00:24:17,600 STUDENT: So the upper and lower limit 577 00:24:17,600 --> 00:24:20,810 and the average that we're using are from our first sample? 578 00:24:20,810 --> 00:24:22,810 PROFESSOR: We've established our baseline, yeah. 579 00:24:22,810 --> 00:24:24,393 So that's not going to change [? anymore. ?] 580 00:24:24,393 --> 00:24:24,846 STUDENT: Do we all use the same one? 581 00:24:24,846 --> 00:24:26,450 PROFESSOR: You all use the same one. 582 00:24:26,450 --> 00:24:28,290 Would it help if I write down those numbers? 583 00:24:28,290 --> 00:24:30,040 STUDENT: Yes, especially there on the left 584 00:24:30,040 --> 00:24:30,460 [? in the screen. ?] 585 00:24:30,460 --> 00:24:30,710 PROFESSOR: Yeah. 586 00:24:30,710 --> 00:24:31,418 Can you see them? 587 00:24:31,418 --> 00:24:32,551 STUDENT: Yeah, there it is. 588 00:24:32,551 --> 00:24:35,710 STUDENT: 75.4 and 69.5. 589 00:24:35,710 --> 00:24:36,670 PROFESSOR: Yeah. 590 00:24:36,670 --> 00:24:37,720 Oh, you've got good eyes. 591 00:24:37,720 --> 00:24:39,760 OK, it's nice to be with young people. 592 00:24:39,760 --> 00:24:42,018 [INTERPOSING VOICES] 593 00:24:42,018 --> 00:24:43,810 And then now, you just operate on your own. 594 00:24:43,810 --> 00:24:45,250 You're all independent. 595 00:24:45,250 --> 00:24:48,610 Think of yourself as the independent measurement quality 596 00:24:48,610 --> 00:24:50,230 control group of a pharmacy. 597 00:24:50,230 --> 00:24:51,016 OK? 598 00:24:51,016 --> 00:24:52,141 STUDENT: [INAUDIBLE] for A. 599 00:24:52,141 --> 00:24:54,105 STUDENT: 0 to 7 and 1/2. 600 00:24:54,105 --> 00:24:55,578 STUDENT: 74. 601 00:24:55,578 --> 00:24:57,940 [INAUDIBLE] There's a rock in mine. 602 00:24:57,940 --> 00:24:59,150 [? Turn ?] [? it ?] around. 603 00:24:59,150 --> 00:25:01,400 STUDENT: Yeah, [INAUDIBLE] [? wrong ?] with it. 604 00:25:01,400 --> 00:25:02,045 [INAUDIBLE] 605 00:25:02,045 --> 00:25:03,920 STUDENT: Average is going to be [INAUDIBLE].. 606 00:25:03,920 --> 00:25:07,920 [INTERPOSING VOICES] 607 00:25:09,972 --> 00:25:12,310 STUDENT: So it will be [INAUDIBLE] 16. 608 00:25:12,310 --> 00:25:14,782 STUDENT: [? RG ?] is-- OK. 609 00:25:14,782 --> 00:25:16,240 STUDENT: [? And ?] are the machines 610 00:25:16,240 --> 00:25:18,400 that you use for calculating [INAUDIBLE] 611 00:25:18,400 --> 00:25:20,765 [INTERPOSING VOICES] 612 00:25:20,765 --> 00:25:23,810 STUDENT: So let's talk about how those might have been issues. 613 00:25:23,810 --> 00:25:25,110 STUDENT: OK. 614 00:25:25,110 --> 00:25:26,230 [INAUDIBLE] materials. 615 00:25:26,230 --> 00:25:27,230 STUDENT: The cup itself? 616 00:25:27,230 --> 00:25:28,080 STUDENT: Yes, sir. 617 00:25:28,080 --> 00:25:30,065 [INTERPOSING VOICES] 618 00:25:30,910 --> 00:25:33,317 STUDENT: You want to write those down? 619 00:25:33,317 --> 00:25:34,650 STUDENT: Quality of [? scale? ?] 620 00:25:34,650 --> 00:25:36,692 STUDENT: Yeah, quality of [? scale ?] [INAUDIBLE] 621 00:25:36,692 --> 00:25:38,144 machine [INAUDIBLE]. 622 00:25:38,144 --> 00:25:41,975 [INTERPOSING VOICES] 623 00:25:45,093 --> 00:25:48,640 STUDENT: It could be not leaving it long enough on the scale. 624 00:25:48,640 --> 00:25:51,450 STUDENT: Well, I just sampled three different ones. 625 00:25:51,450 --> 00:25:54,425 [LAUGHTER] 626 00:25:54,425 --> 00:25:55,700 STUDENT: [INAUDIBLE] 627 00:25:55,700 --> 00:25:57,920 STUDENT: So bean variations. 628 00:25:57,920 --> 00:26:01,049 [INTERPOSING VOICES] 629 00:26:05,450 --> 00:26:08,390 STUDENT: I think it's pretty full. 630 00:26:08,390 --> 00:26:11,156 PROFESSOR: OK, well let's see. 631 00:26:11,156 --> 00:26:12,700 STUDENT: [? If ?] we dump it out? 632 00:26:12,700 --> 00:26:15,200 STUDENT: Well, I'll try dumping it out and see [INAUDIBLE].. 633 00:26:15,200 --> 00:26:17,442 [INTERPOSING VOICES] 634 00:26:17,442 --> 00:26:19,400 PROFESSOR: [INAUDIBLE] [? It's ?] [INAUDIBLE].. 635 00:26:19,400 --> 00:26:20,858 STUDENT: Why are they [INAUDIBLE].. 636 00:26:20,858 --> 00:26:22,130 STUDENT: Oh. 637 00:26:22,130 --> 00:26:23,360 Look at that. 638 00:26:23,360 --> 00:26:25,220 PROFESSOR: OK, so going around the room, 639 00:26:25,220 --> 00:26:31,040 we had a range of a root cause analysis things, ranging from-- 640 00:26:31,040 --> 00:26:33,808 this table just said, we think this is the cause. 641 00:26:33,808 --> 00:26:35,600 And they tested it, and that was the cause. 642 00:26:35,600 --> 00:26:37,440 So they just jumped right to a conclusion. 643 00:26:37,440 --> 00:26:39,950 We had a very structured process over here. 644 00:26:39,950 --> 00:26:41,853 But this is a way to make it visible. 645 00:26:41,853 --> 00:26:44,270 If you go into a facility that's under statistical process 646 00:26:44,270 --> 00:26:46,520 control, the employees are trained on this. 647 00:26:46,520 --> 00:26:49,700 And they're trained to recognize this kind of deviation. 648 00:26:49,700 --> 00:26:52,680 And then they do something about it. 649 00:26:52,680 --> 00:26:56,580 OK, here's a chart for resident falls in a long-term care 650 00:26:56,580 --> 00:26:59,370 facility, control chart. 651 00:26:59,370 --> 00:27:01,260 And here's their 20 days. 652 00:27:01,260 --> 00:27:04,890 And this was their control. 653 00:27:04,890 --> 00:27:07,530 Now, in this case, they were intervening 654 00:27:07,530 --> 00:27:09,160 to try to reduce patient falls. 655 00:27:09,160 --> 00:27:12,030 So this is not now just doing nothing. 656 00:27:12,030 --> 00:27:13,710 They're doing a bunch of interventions. 657 00:27:13,710 --> 00:27:15,600 And these are different interventions. 658 00:27:15,600 --> 00:27:18,240 Actually, what they did here was-- 659 00:27:18,240 --> 00:27:19,740 it's not on the chart, but what they 660 00:27:19,740 --> 00:27:21,480 did was they just simply identified 661 00:27:21,480 --> 00:27:23,550 the patients in this long-term care facility who 662 00:27:23,550 --> 00:27:25,320 were susceptible to falling. 663 00:27:25,320 --> 00:27:29,370 And they put stickers on their wheelchairs or their walkers 664 00:27:29,370 --> 00:27:31,410 in their room and on their charts. 665 00:27:31,410 --> 00:27:34,110 And they picked stickers that were so interesting, 666 00:27:34,110 --> 00:27:35,910 that the residents who weren't in danger 667 00:27:35,910 --> 00:27:38,520 wanted the stickers, too. 668 00:27:38,520 --> 00:27:41,080 But just being conscious of-- 669 00:27:41,080 --> 00:27:43,780 those were at-- the at-risk patients started to drop it. 670 00:27:43,780 --> 00:27:47,070 And you can see their interventions lowered the mean 671 00:27:47,070 --> 00:27:49,860 and lowered the control limits and lowered them again. 672 00:27:49,860 --> 00:27:51,990 And this was a process improvement study 673 00:27:51,990 --> 00:27:54,900 that I actually heard at the last IHI meeting. 674 00:27:54,900 --> 00:27:57,640 OK, so now our last topic-- 675 00:27:57,640 --> 00:27:59,790 this is Process Capability. 676 00:27:59,790 --> 00:28:02,220 Process Capability is defined as the ability of a process 677 00:28:02,220 --> 00:28:04,720 to meet the customer expectations. 678 00:28:04,720 --> 00:28:07,402 And what we have here is we know what the capability is, 679 00:28:07,402 --> 00:28:09,360 but we don't know whether it meets the customer 680 00:28:09,360 --> 00:28:10,170 expectations. 681 00:28:10,170 --> 00:28:13,060 And to get the customer expectations on there, 682 00:28:13,060 --> 00:28:14,850 we have to find out what they want. 683 00:28:14,850 --> 00:28:16,200 HUGH MCMANUS: The customer-determined limit, 684 00:28:16,200 --> 00:28:18,575 the spec limits, are essentially what the customer wants. 685 00:28:18,575 --> 00:28:20,310 The upper and lower values between which 686 00:28:20,310 --> 00:28:23,790 a process must be controlled-- that's what the customer wants. 687 00:28:23,790 --> 00:28:27,390 The bounds in which we choose to control the process, 688 00:28:27,390 --> 00:28:29,340 or which we try to control the process, 689 00:28:29,340 --> 00:28:33,430 are the upper and lower control limits of the process. 690 00:28:33,430 --> 00:28:36,720 So how do we measure, given those two definitions, 691 00:28:36,720 --> 00:28:39,010 the process capability? 692 00:28:39,010 --> 00:28:41,160 One way to do it is to think about the process. 693 00:28:41,160 --> 00:28:42,700 And again, this is an approximation. 694 00:28:42,700 --> 00:28:46,110 But it's a pretty good one, under a lot of circumstances. 695 00:28:46,110 --> 00:28:48,910 The process is having a normally distributed behavior. 696 00:28:48,910 --> 00:28:51,090 So if we collect some data, we can 697 00:28:51,090 --> 00:28:55,950 assume a normal distribution, do some statistics using 698 00:28:55,950 --> 00:28:58,740 our statistical tools, and decide 699 00:28:58,740 --> 00:29:01,650 how good or bad the process is. 700 00:29:01,650 --> 00:29:05,130 And if we have a normally distributed process, 701 00:29:05,130 --> 00:29:07,560 its behavior is going to be characterized 702 00:29:07,560 --> 00:29:10,120 by its standard deviation, sigma. 703 00:29:10,120 --> 00:29:12,792 And if we draw our spec limits around that normal 704 00:29:12,792 --> 00:29:14,625 distribution-- here is our lower spec limit, 705 00:29:14,625 --> 00:29:16,950 and here's our upper spec limit-- 706 00:29:16,950 --> 00:29:25,170 we can define a quantity CP, the process capability, 707 00:29:25,170 --> 00:29:29,250 which is essentially this distance here, the upper limit 708 00:29:29,250 --> 00:29:31,090 less this lower spec limit-- 709 00:29:31,090 --> 00:29:33,450 so that the range which is acceptable 710 00:29:33,450 --> 00:29:37,000 divided by the Six Sigma of the process. 711 00:29:37,000 --> 00:29:43,470 And if we have, say, a CP of 2, that's a very tight. 712 00:29:43,470 --> 00:29:48,630 Basically, the spec limits are twice Six Sigma of the process. 713 00:29:48,630 --> 00:29:51,060 Then it's a very tight distribution 714 00:29:51,060 --> 00:29:52,770 inside the spec limits. 715 00:29:52,770 --> 00:29:55,860 And the chance of something bad happening are very, very small, 716 00:29:55,860 --> 00:29:57,170 the chance of going out. 717 00:29:57,170 --> 00:30:00,395 So a CP of 2 is a very good process. 718 00:30:00,395 --> 00:30:01,020 That, in fact-- 719 00:30:01,020 --> 00:30:03,398 STUDENT: [INAUDIBLE] Six Sigma is [INAUDIBLE]---- 720 00:30:03,398 --> 00:30:04,440 HUGH MCMANUS: Six Sigma-- 721 00:30:04,440 --> 00:30:05,580 STUDENT: --standard deviations [INAUDIBLE].. 722 00:30:05,580 --> 00:30:07,747 HUGH MCMANUS: That's right-- six standard deviations 723 00:30:07,747 --> 00:30:09,990 on either side of the mean, right. 724 00:30:09,990 --> 00:30:13,260 So a CP of 2 would be a Six Sigma process. 725 00:30:13,260 --> 00:30:15,563 In that sense, it would be a very good one. 726 00:30:15,563 --> 00:30:16,980 And just graphically, you can just 727 00:30:16,980 --> 00:30:19,740 see there's basically no tail sticking out of the bad zone. 728 00:30:19,740 --> 00:30:22,560 CP of 1 means you've got three sigma on each side. 729 00:30:22,560 --> 00:30:28,050 Three sigma, you've got 97 point whatever percent of the process 730 00:30:28,050 --> 00:30:29,500 falling in that range. 731 00:30:29,500 --> 00:30:31,770 So the chance of something falling outside the range 732 00:30:31,770 --> 00:30:33,090 is quite small. 733 00:30:33,090 --> 00:30:34,020 But it's not 0. 734 00:30:34,020 --> 00:30:35,670 We can see a little bit of ink here 735 00:30:35,670 --> 00:30:38,500 outside of the acceptable range. 736 00:30:38,500 --> 00:30:40,860 And once we get below 1, it starts getting ugly. 737 00:30:40,860 --> 00:30:43,920 We really don't want to even talk about that-- 738 00:30:43,920 --> 00:30:48,280 so from a statistical process control point of view. 739 00:30:48,280 --> 00:30:51,420 Another issue which one can ignore 740 00:30:51,420 --> 00:30:54,690 is the fact that we may not be centered. 741 00:30:54,690 --> 00:30:59,760 And we can actually redefine our process capability 742 00:30:59,760 --> 00:31:03,600 using a metric called CPK, which takes into a fact 743 00:31:03,600 --> 00:31:06,270 that we may drift off mean. 744 00:31:06,270 --> 00:31:07,440 And it's defined this way. 745 00:31:07,440 --> 00:31:13,380 It's basically the distance to the closer boundary from where 746 00:31:13,380 --> 00:31:15,540 we are divided by three sigma. 747 00:31:15,540 --> 00:31:18,190 And it's whichever one of those is worse. 748 00:31:18,190 --> 00:31:22,980 And those same curves shifted over 1 and 1/2 sigma 749 00:31:22,980 --> 00:31:25,080 look a lot less pretty. 750 00:31:25,080 --> 00:31:30,120 Our former Six Sigma process now has a CPK at 1.5. 751 00:31:30,120 --> 00:31:31,873 But it's still OK. 752 00:31:31,873 --> 00:31:33,540 There's a tiny, tiny-- there's a couple, 753 00:31:33,540 --> 00:31:36,150 like two pixels worth of ink sticking out there. 754 00:31:36,150 --> 00:31:40,620 The chance of a defect is still very, very small. 755 00:31:40,620 --> 00:31:46,110 But our formally OK-looking CP of 1 process 756 00:31:46,110 --> 00:31:49,290 now has quite a bad-looking tail. 757 00:31:49,290 --> 00:31:53,090 And our process that was bad is now horrible, 758 00:31:53,090 --> 00:31:55,113 our relatively uncontrolled process. 759 00:31:55,113 --> 00:31:57,530 So that's a different way of measuring process capability, 760 00:31:57,530 --> 00:32:00,740 which takes into account the fact that the mean may not 761 00:32:00,740 --> 00:32:02,390 be on the center. 762 00:32:02,390 --> 00:32:08,450 What we do with these two measures is different. 763 00:32:08,450 --> 00:32:10,730 Here we have a archery example. 764 00:32:10,730 --> 00:32:17,090 This is a classic statistically characterize-able process. 765 00:32:17,090 --> 00:32:19,220 We have shots that are both widely 766 00:32:19,220 --> 00:32:21,650 dispersed and off-center. 767 00:32:21,650 --> 00:32:24,630 So how would we characterize that? 768 00:32:24,630 --> 00:32:26,640 It's basically low on both metrics, right? 769 00:32:26,640 --> 00:32:31,490 It's widely dispersed, and it's off-center. 770 00:32:31,490 --> 00:32:35,340 Here's another archer. 771 00:32:35,340 --> 00:32:39,407 It has a tight distribution, but it's way off-center. 772 00:32:39,407 --> 00:32:40,900 STUDENT: So CP is high. 773 00:32:40,900 --> 00:32:43,610 HUGH MCMANUS: So CP is high, and CPK is low. 774 00:32:43,610 --> 00:32:44,170 That's right. 775 00:32:44,170 --> 00:32:46,868 So that's the difference between those metrics. 776 00:32:46,868 --> 00:32:48,160 On this metric, it looks great. 777 00:32:48,160 --> 00:32:50,140 On that metric, it does not. 778 00:32:50,140 --> 00:32:53,300 And this, of course, is what we want. 779 00:32:53,300 --> 00:32:56,680 If you're an archer or any other person who 780 00:32:56,680 --> 00:33:01,790 needs to control their process, what do you do first? 781 00:33:01,790 --> 00:33:03,770 Yeah, I'm cheating a little bit on this one. 782 00:33:03,770 --> 00:33:06,440 I do archery sometimes, not very much. 783 00:33:06,440 --> 00:33:08,630 But you have to do this first, actually. 784 00:33:08,630 --> 00:33:10,820 You have to get your process repeatable, 785 00:33:10,820 --> 00:33:12,920 even if the meat is way off. 786 00:33:12,920 --> 00:33:15,290 Because if you're correcting every time, 787 00:33:15,290 --> 00:33:18,150 then it just goes all over the place. 788 00:33:18,150 --> 00:33:23,200 But if you just go, OK, same thing every time, OK. 789 00:33:23,200 --> 00:33:25,630 Five times in a row, I've hit up and to the right. 790 00:33:25,630 --> 00:33:26,740 Now I'm going to adjust. 791 00:33:29,290 --> 00:33:34,060 So often, you concentrate on knocking your CP down first, 792 00:33:34,060 --> 00:33:36,580 to understand your process, to see where the mean is. 793 00:33:36,580 --> 00:33:38,350 Because you can't really measure the mean 794 00:33:38,350 --> 00:33:41,200 in a situation like-- you can statistically, but it's hard. 795 00:33:41,200 --> 00:33:43,480 OK, enough of that. 796 00:33:46,890 --> 00:33:49,860 This actually-- remember we said we would-- 797 00:33:49,860 --> 00:33:52,530 I'd tell you where the definition of Six Sigma 798 00:33:52,530 --> 00:33:53,190 comes from. 799 00:33:53,190 --> 00:33:57,810 Interestingly, it comes from a process 800 00:33:57,810 --> 00:34:00,270 with a mean shift of 1 and 1/2, which 801 00:34:00,270 --> 00:34:04,660 would give you three defects per million opportunities. 802 00:34:04,660 --> 00:34:06,630 So if you have a very good process, 803 00:34:06,630 --> 00:34:11,790 but you still don't control your mean to more than 1 804 00:34:11,790 --> 00:34:17,880 and 1/2 sigma, you still have a very low defects per million. 805 00:34:17,880 --> 00:34:19,949 That's where that idea comes from. 806 00:34:19,949 --> 00:34:21,540 Those are some of the concepts that we 807 00:34:21,540 --> 00:34:25,260 use to get processes under control and keep them there. 808 00:34:25,260 --> 00:34:31,560 And Six Sigma, as an overall method, 809 00:34:31,560 --> 00:34:36,030 is very useful for taking variation down, especially 810 00:34:36,030 --> 00:34:40,800 in critical applications like large-scale manufacturing 811 00:34:40,800 --> 00:34:44,190 or health care, where you want your defects to be very low, 812 00:34:44,190 --> 00:34:45,832 your variation to be very low. 813 00:34:45,832 --> 00:34:47,790 We've talked to you about control charts, which 814 00:34:47,790 --> 00:34:49,610 is a great place to start. 815 00:34:49,610 --> 00:34:50,469 It's not the end. 816 00:34:50,469 --> 00:34:53,730 And we've glossed over all the statistical tools 817 00:34:53,730 --> 00:34:55,739 you need to really understand the numbers. 818 00:34:55,739 --> 00:34:59,550 But the visual is very powerful in and of itself. 819 00:34:59,550 --> 00:35:01,140 Even if you don't do the statistics, 820 00:35:01,140 --> 00:35:04,590 having the visual evidence of what your process can do 821 00:35:04,590 --> 00:35:06,435 and whether it's deviating is very powerful. 822 00:35:11,140 --> 00:35:13,830 And once we understand what our process can do, 823 00:35:13,830 --> 00:35:18,420 we can compare that to what the customer wants and understand 824 00:35:18,420 --> 00:35:21,030 the capabilities of our process and whether they're acceptable 825 00:35:21,030 --> 00:35:22,640 or not.