1 00:00:06,894 --> 00:00:10,762 - Learning Objective 7.3 Establish Portfolio Flow. 2 00:00:10,762 --> 00:00:13,981 We talked just a few minutes ago about establishing 3 00:00:13,981 --> 00:00:15,397 the funding for value streams. 4 00:00:15,397 --> 00:00:17,569 We kind of pretended that argument that 5 00:00:17,569 --> 00:00:18,945 you're done, you're good to go. 6 00:00:18,945 --> 00:00:21,615 That's the only significant role that the portfolio 7 00:00:21,615 --> 00:00:23,290 plays in funding projects. 8 00:00:23,290 --> 00:00:25,194 That's an oversimplification of the case. 9 00:00:25,194 --> 00:00:27,829 The reality is that no matter how my portfolio works, 10 00:00:27,829 --> 00:00:29,446 and how many value streams I have, 11 00:00:29,446 --> 00:00:32,195 they're gonna be some substantial initiatives 12 00:00:32,195 --> 00:00:33,660 that are gonna affect the portfolio 13 00:00:33,660 --> 00:00:37,207 that need to be driven from a centralized, 14 00:00:37,207 --> 00:00:39,774 strategic theme, or portfolio strategy perspective. 15 00:00:39,774 --> 00:00:41,837 Let me give you an example of one. 16 00:00:41,837 --> 00:00:43,707 Company has a whole series of solutions 17 00:00:43,707 --> 00:00:45,771 and software applications, and they want 18 00:00:45,771 --> 00:00:47,826 to migrate their web server applications. 19 00:00:47,826 --> 00:00:49,444 They want to move from where they are, I won't say 20 00:00:49,444 --> 00:00:50,958 where they are, to JBoss. 21 00:00:50,958 --> 00:00:53,793 They've decided that that is the right economics for them. 22 00:00:53,793 --> 00:00:57,081 So they say, you know, we need to move everyone to Jboss. 23 00:00:57,081 --> 00:00:59,129 That's a really big deal, okay? 24 00:00:59,129 --> 00:01:01,232 That's a lot of money, frankly, because there's 25 00:01:01,232 --> 00:01:03,901 multiple products, and some of those products are 26 00:01:03,901 --> 00:01:04,850 on the wane. 27 00:01:04,850 --> 00:01:07,592 Does it really make sense to go in and make that conversion 28 00:01:07,592 --> 00:01:10,188 in that case, or should be just re-up our old license 29 00:01:10,188 --> 00:01:13,569 for our old app server, instead of converting to JBoss? 30 00:01:13,569 --> 00:01:15,181 Not really sure, but we definitely need 31 00:01:15,181 --> 00:01:16,470 to reason about that. 32 00:01:16,470 --> 00:01:19,445 So we need a way to communicate these large initiatives. 33 00:01:19,445 --> 00:01:21,849 Single sign-on across a suite of products; 34 00:01:21,849 --> 00:01:25,120 things like gosh, we need to create a mobile app 35 00:01:25,120 --> 00:01:25,953 for the first time. 36 00:01:25,953 --> 00:01:28,127 That might come through as a mobile app and 37 00:01:28,127 --> 00:01:30,172 the technology to support that. 38 00:01:30,172 --> 00:01:33,115 So there are portfolio-level concerns virtually always 39 00:01:33,115 --> 00:01:35,545 that influence the entire line of business. 40 00:01:35,545 --> 00:01:37,510 And we need a word for that, and the word that we picked 41 00:01:37,510 --> 00:01:39,341 in SAFe is simply called Epic. 42 00:01:39,341 --> 00:01:41,921 Now in some versions of Agile and Scrum, 43 00:01:41,921 --> 00:01:44,287 an Epic is a thing bigger than a story. 44 00:01:44,287 --> 00:01:47,204 Well it's so much bigger that we didn't use that paradigm. 45 00:01:47,204 --> 00:01:50,011 We used the Story, the Feature, and the Epic. 46 00:01:50,011 --> 00:01:51,579 And I really like the word Epic for that 47 00:01:51,579 --> 00:01:53,366 because it's epic. 48 00:01:53,366 --> 00:01:56,389 And people respect the word because of its very 49 00:01:56,389 --> 00:01:58,520 connotation, that it's a pretty big deal. 50 00:01:58,520 --> 00:02:00,022 And it is a pretty big deal. 51 00:02:00,022 --> 00:02:03,145 So we define these Epics, and they can happen at any level, 52 00:02:03,145 --> 00:02:05,597 but for now we'll just talk about portfolio Epics, 53 00:02:05,597 --> 00:02:08,215 they typically cut across value streams. 54 00:02:08,215 --> 00:02:10,758 Most times the ones I mentioned here, let's 55 00:02:10,758 --> 00:02:14,761 just say migrate to JBoss, 64-byte back-office servers, 56 00:02:14,761 --> 00:02:16,344 build a mobile device. 57 00:02:16,344 --> 00:02:17,550 These are cross-cutting concerns. 58 00:02:17,550 --> 00:02:20,807 There's multiple value streams that get affected by that. 59 00:02:20,807 --> 00:02:22,279 So most typically that's one of the reasons 60 00:02:22,279 --> 00:02:23,531 they're impactful, 61 00:02:23,531 --> 00:02:25,885 is we have to communicate a lot of different things. 62 00:02:25,885 --> 00:02:28,367 At the ART level, we can also have a program Epic. 63 00:02:28,367 --> 00:02:31,649 We can have a big thing, indeed migrate to JBoss 64 00:02:31,649 --> 00:02:34,177 for me might just be move to the JBoss app server 65 00:02:34,177 --> 00:02:35,306 in my application. 66 00:02:35,306 --> 00:02:36,931 That might be fine. 67 00:02:36,931 --> 00:02:38,435 And that's all I really need to know. 68 00:02:38,435 --> 00:02:41,659 But for this conversation we'll stick to the larger ones. 69 00:02:41,659 --> 00:02:43,616 Well, do we program an Epic? 70 00:02:43,616 --> 00:02:46,695 No, we don't even program features, do we? 71 00:02:46,695 --> 00:02:48,643 We program stories. 72 00:02:48,643 --> 00:02:50,709 But now we have a hierarchy. 73 00:02:50,709 --> 00:02:52,804 It's a form of a work breakdown structure. 74 00:02:52,804 --> 00:02:54,455 It's okay to say that word. 75 00:02:54,455 --> 00:02:58,064 It says that this work becomes this work becomes this work. 76 00:02:58,064 --> 00:02:59,942 That becomes a bunch of code and tests 77 00:02:59,942 --> 00:03:01,404 and that's how we achieve that. 78 00:03:01,404 --> 00:03:03,267 So now we have a way to discuss 79 00:03:03,267 --> 00:03:04,904 really, really big things, 80 00:03:04,904 --> 00:03:07,106 really, really medium-sized things, 81 00:03:07,106 --> 00:03:09,806 and small things, in a consistent way. 82 00:03:09,806 --> 00:03:12,137 That taxonomy is really valuable. 83 00:03:12,137 --> 00:03:14,135 As you grow, as your company grows, 84 00:03:14,135 --> 00:03:15,624 as you become more distributed, 85 00:03:15,624 --> 00:03:19,136 as you acquire, as you become acquired, 86 00:03:19,136 --> 00:03:21,797 a common way of talking about the work you need to do 87 00:03:21,797 --> 00:03:23,133 is incredibly valuable. 88 00:03:23,133 --> 00:03:25,631 And that's one of the benefits of SAFe is 89 00:03:25,631 --> 00:03:29,598 it certainly has a taxonomy associated with it. 90 00:03:29,598 --> 00:03:31,542 Most all the words are used in multiple context 91 00:03:31,542 --> 00:03:32,736 and other locations. 92 00:03:32,736 --> 00:03:34,956 That's fine, we just have to create a definition 93 00:03:34,956 --> 00:03:36,691 that works for us, so that we can talk 94 00:03:36,691 --> 00:03:38,615 consistently about that. 95 00:03:38,615 --> 00:03:40,029 Now we're gonna look at where they go, 96 00:03:40,029 --> 00:03:42,766 and we're gonna talk about Kanban system 97 00:03:42,766 --> 00:03:45,830 that we're gonna use to visualize that work. 98 00:03:45,830 --> 00:03:48,446 Before we do that, however, we're also gonna talk about 99 00:03:48,446 --> 00:03:51,173 one of the most critical aspects of 100 00:03:51,173 --> 00:03:53,368 initiating large-scale programs. 101 00:03:53,368 --> 00:03:55,195 To do this I'm going to share a little vignette 102 00:03:55,195 --> 00:03:56,893 that's from my recent history, 103 00:03:56,893 --> 00:04:00,182 that was about the time we were releasing 4.5. 104 00:04:00,182 --> 00:04:03,304 I was talking to a CIO of one of the world's 105 00:04:03,304 --> 00:04:07,566 largest cable companies about frameworks and SAFe. 106 00:04:07,566 --> 00:04:09,778 And he said, you know, I don't really care. 107 00:04:09,778 --> 00:04:11,062 That's an IT thing. 108 00:04:11,062 --> 00:04:13,502 How they build systems is up to them. 109 00:04:13,502 --> 00:04:16,862 He said, all I need to know is what I'm gonna get, 110 00:04:16,862 --> 00:04:19,181 when I'm gonna get it, exactly what it's gonna cost, 111 00:04:19,181 --> 00:04:20,873 what the business benefits are. 112 00:04:20,873 --> 00:04:22,523 Well, we've heard that story before. 113 00:04:22,523 --> 00:04:24,612 In other words, I'm gonna define large initiatives, 114 00:04:24,612 --> 00:04:27,114 I want a guarantee from IT that they're gonna 115 00:04:27,114 --> 00:04:29,417 deliver on time, and the result's gonna be there. 116 00:04:29,417 --> 00:04:30,986 We know the damage of that. 117 00:04:30,986 --> 00:04:32,777 That's the iron triangle big-time, the 118 00:04:32,777 --> 00:04:34,784 biggest iron triangle we can be in. 119 00:04:34,784 --> 00:04:37,835 Scope is fixed, resources are fixed, timeline is fixed. 120 00:04:37,835 --> 00:04:38,872 Do your best. 121 00:04:38,872 --> 00:04:40,006 Okay, we don't want to be there. 122 00:04:40,006 --> 00:04:42,559 We want to be at a much more interactive mode. 123 00:04:42,559 --> 00:04:44,041 So I had a discussion with him because we'd just 124 00:04:44,041 --> 00:04:46,236 incorporated the Lean Startup cycle, 125 00:04:46,236 --> 00:04:49,024 which you can see here in the SAFe. 126 00:04:49,024 --> 00:04:51,391 And I said, give me an example of such an initiative. 127 00:04:51,391 --> 00:04:54,042 And he said, well we're going to accomplish foo 128 00:04:54,042 --> 00:04:55,809 in the next year or so. 129 00:04:55,809 --> 00:04:59,242 And I said, what's the hypothesis behind foo? 130 00:04:59,242 --> 00:05:01,619 And he looked at me funny for a second, and he said, 131 00:05:01,619 --> 00:05:02,452 well, what do you mean? 132 00:05:02,452 --> 00:05:04,483 I said, well what will the outcomes be? 133 00:05:04,483 --> 00:05:06,491 Well, we'll be able to do this, that and the other thing. 134 00:05:06,491 --> 00:05:08,162 We'll have this cool capability, we'll be able to 135 00:05:08,162 --> 00:05:11,080 compete more effectively with another enterprise, 136 00:05:11,080 --> 00:05:12,413 et cetera et cetera. 137 00:05:12,413 --> 00:05:13,839 I said, okay. 138 00:05:13,839 --> 00:05:18,006 What would you have to spend to just prove that hypothesis? 139 00:05:18,972 --> 00:05:19,805 What do you mean? 140 00:05:19,805 --> 00:05:22,478 I said, well would you fund the whole $30 million 141 00:05:22,478 --> 00:05:24,262 under the assumption that you're correct 142 00:05:24,262 --> 00:05:25,660 about the hypothesis? 143 00:05:25,660 --> 00:05:26,796 Well no, we wouldn't have to do that. 144 00:05:26,796 --> 00:05:29,534 We could just spend a portion of that amount of money. 145 00:05:29,534 --> 00:05:31,718 And if we just did this and these cable boxes 146 00:05:31,718 --> 00:05:32,853 and we modified them, and we put them 147 00:05:32,853 --> 00:05:34,315 in a single market place, 148 00:05:34,315 --> 00:05:36,277 we could test this hypothesis pretty cheaply. 149 00:05:36,277 --> 00:05:39,889 I said, okay, so you wouldn't really want to allocate 150 00:05:39,889 --> 00:05:42,392 the entire budget to that Epic, and you wouldn't really 151 00:05:42,392 --> 00:05:44,850 want to take the entire Epic and blow it out 152 00:05:44,850 --> 00:05:47,056 into the big fat business case, 153 00:05:47,056 --> 00:05:49,171 with all these requirements and all these markets served 154 00:05:49,171 --> 00:05:50,510 and all these assumptions. 155 00:05:50,510 --> 00:05:52,994 You'd rather build the MVP. 156 00:05:52,994 --> 00:05:53,827 What's that? 157 00:05:53,827 --> 00:05:56,123 Well, that's the minimum viable thing you can build, 158 00:05:56,123 --> 00:05:57,701 the least amount of money you can spend, 159 00:05:57,701 --> 00:05:58,878 that was his language, 160 00:05:58,878 --> 00:06:00,685 to prove the hypothesis. 161 00:06:00,685 --> 00:06:03,067 And then we'll evaluate that MVP. 162 00:06:03,067 --> 00:06:06,214 We'll look at that MVP and say, did we do it or not? 163 00:06:06,214 --> 00:06:09,578 And if the answer is no, let's not spend the rest 164 00:06:09,578 --> 00:06:10,427 of the $30 million. 165 00:06:10,427 --> 00:06:12,341 Oh no, of course not, we wouldn't do that. 166 00:06:12,341 --> 00:06:13,976 Great, let's stop work. 167 00:06:13,976 --> 00:06:17,353 Great experiment, we tried to innovate, had a good idea, 168 00:06:17,353 --> 00:06:20,928 had a great learning outcome, decided not to do it, great. 169 00:06:20,928 --> 00:06:22,601 You just saved $30 million. 170 00:06:22,601 --> 00:06:24,316 Did you waste 5 or save 30? 171 00:06:24,316 --> 00:06:26,004 I think you've saved 30. 172 00:06:26,004 --> 00:06:27,181 Well what if it works? 173 00:06:27,181 --> 00:06:29,626 Okay, we'll just implement additional features then, 174 00:06:29,626 --> 00:06:30,547 and we'll keep going. 175 00:06:30,547 --> 00:06:33,494 Now I discussed earlier, weighted shortest job first. 176 00:06:33,494 --> 00:06:35,435 That's really a powerful thing. 177 00:06:35,435 --> 00:06:37,195 It's kind of like a magic trick. 178 00:06:37,195 --> 00:06:40,762 If we create the MVP and we launch and we evaluate that, 179 00:06:40,762 --> 00:06:42,627 we then come to the next decision which says, 180 00:06:42,627 --> 00:06:45,342 this is good stuff, let's continue to invest 181 00:06:45,342 --> 00:06:47,190 down this direction. 182 00:06:47,190 --> 00:06:48,520 I want to bring in my new features, 183 00:06:48,520 --> 00:06:50,866 so let me put my backlog over here. 184 00:06:50,866 --> 00:06:53,329 Here's my new feature entering the backlog, okay? 185 00:06:53,329 --> 00:06:55,181 This is the one from the MVP. 186 00:06:55,181 --> 00:06:58,004 This is MVP plus another feature. 187 00:06:58,004 --> 00:07:00,656 Well it's now competing with somebody else's feature. 188 00:07:00,656 --> 00:07:04,135 What if this other feature has more business value? 189 00:07:04,135 --> 00:07:05,405 I want to do that. 190 00:07:05,405 --> 00:07:08,396 So we can kind of go on auto-pilot, 191 00:07:08,396 --> 00:07:09,293 because the system 192 00:07:09,293 --> 00:07:10,925 is going to naturally tend to sort 193 00:07:10,925 --> 00:07:12,639 based upon economic value, 194 00:07:12,639 --> 00:07:14,440 the right features in the right way. 195 00:07:14,440 --> 00:07:17,462 So that MVP, that Epic, will stop 196 00:07:17,462 --> 00:07:18,296 at the time that it's 197 00:07:18,296 --> 00:07:21,016 economically judicious to do so, 198 00:07:21,016 --> 00:07:22,974 and it will stop somewhat automatically. 199 00:07:22,974 --> 00:07:25,338 Now there's smart people making all these decisions. 200 00:07:25,338 --> 00:07:27,604 It's not an algorithm that we plug in, 201 00:07:27,604 --> 00:07:29,682 where people decide that wow, this next thing 202 00:07:29,682 --> 00:07:31,765 on that MVP is really valuable. 203 00:07:31,765 --> 00:07:34,312 That is the next most important one, no problem. 204 00:07:34,312 --> 00:07:35,962 But we don't have to worry so much about 205 00:07:35,962 --> 00:07:39,015 where the edge of the Epic is anymore. 206 00:07:39,015 --> 00:07:41,743 We're just gonna execute until such time as 207 00:07:41,743 --> 00:07:45,421 that feature can't compete with the next one on the backlog. 208 00:07:45,421 --> 00:07:47,047 And that's a really cool thing. 209 00:07:47,047 --> 00:07:49,677 Because it gives us a way to start our major initiatives 210 00:07:49,677 --> 00:07:53,086 with a minimum viable product, established hypothesis, 211 00:07:53,086 --> 00:07:55,052 prove the hypothesis and move forward. 212 00:07:55,052 --> 00:07:58,181 I think hypothesis is a magic word. 213 00:07:58,181 --> 00:07:59,908 It's a magic word that starts to make you 214 00:07:59,908 --> 00:08:01,156 think about things differently. 215 00:08:01,156 --> 00:08:03,177 Even in our own shop. 216 00:08:03,177 --> 00:08:05,662 I had an interesting idea the other day. 217 00:08:05,662 --> 00:08:06,689 People were looking at that and saying, 218 00:08:06,689 --> 00:08:08,712 well you're a pretty big guy here, 219 00:08:08,712 --> 00:08:11,770 if you say it's interesting, we probably should do it. 220 00:08:11,770 --> 00:08:15,550 We discussed it and said, no it's a thought experiment. 221 00:08:15,550 --> 00:08:16,638 It's a hypothesis. 222 00:08:16,638 --> 00:08:18,490 Everybody said, oh okay. 223 00:08:18,490 --> 00:08:20,416 Let's go take the hypothesis and let's 224 00:08:20,416 --> 00:08:21,913 reason about it a little bit 225 00:08:21,913 --> 00:08:23,246 and come back and say, 226 00:08:23,246 --> 00:08:25,703 yeah okay, let's try that hypothesis or a different one. 227 00:08:25,703 --> 00:08:26,536 That's great. 228 00:08:26,536 --> 00:08:30,328 It gets you out of the boss said do foo, and everybody's 229 00:08:30,328 --> 00:08:32,424 launched on foo, and then to the fact when the 230 00:08:32,424 --> 00:08:34,648 cable guy says, hey folks. 231 00:08:34,648 --> 00:08:36,808 What's the minimum amount we can spend 232 00:08:36,808 --> 00:08:37,988 to test the hypothesis? 233 00:08:37,988 --> 00:08:40,721 Can you imagine the reaction of developers 234 00:08:40,721 --> 00:08:43,563 four or five levels deep when they hear the fact that 235 00:08:43,563 --> 00:08:46,290 hey, what do we need to do to prove the hypothesis 236 00:08:46,290 --> 00:08:48,398 of this next level spend initiative. 237 00:08:48,398 --> 00:08:50,779 They're gonna go, boy I'm glad I work here. 238 00:08:50,779 --> 00:08:52,359 Things are really starting to change. 239 00:08:52,359 --> 00:08:54,516 A very powerful notion. 240 00:08:54,516 --> 00:08:56,453 That's what the Lean Startup cycle does. 241 00:08:56,453 --> 00:08:59,565 That's why it was so critical for us to integrate it 242 00:08:59,565 --> 00:09:01,642 with the Version 4.5. 243 00:09:01,642 --> 00:09:03,714 I'll stretch the point one more. 244 00:09:03,714 --> 00:09:06,448 Back to what we talked about when we talked about DevOps. 245 00:09:06,448 --> 00:09:10,031 That small thing, if I can't take that MVP 246 00:09:10,031 --> 00:09:12,073 and get it through the pipeline quickly 247 00:09:12,073 --> 00:09:14,500 because the transaction costs of deployment 248 00:09:14,500 --> 00:09:16,519 are too brutal, I can't do this. 249 00:09:16,519 --> 00:09:19,631 So how do you do innovation without DevOps? 250 00:09:19,631 --> 00:09:20,571 Beats me. 251 00:09:20,571 --> 00:09:23,666 Seems like we have to have DevOps as an optional 252 00:09:23,666 --> 00:09:25,475 continuous delivery pipeline as an optional. 253 00:09:25,475 --> 00:09:27,489 It's part of a larger picture, which is 254 00:09:27,489 --> 00:09:31,331 how I can take what was about to be a gigantic batch size, 255 00:09:31,331 --> 00:09:33,480 and deliver it in little pieces, 256 00:09:33,480 --> 00:09:35,742 and get feedback on each and every feature. 257 00:09:35,742 --> 00:09:37,504 Very powerful construct. 258 00:09:37,504 --> 00:09:39,683 Really happy to have achieved that in 4.5. 259 00:09:39,683 --> 00:09:42,656 I think you can we don't see Lean Startup and 260 00:09:42,656 --> 00:09:44,893 continuous delivery as different things. 261 00:09:44,893 --> 00:09:46,725 Yeah, they're different conferences. 262 00:09:46,725 --> 00:09:47,894 Ten thousand go to each. 263 00:09:47,894 --> 00:09:49,702 And they maybe don't communicate with each other. 264 00:09:49,702 --> 00:09:52,355 We see it as one unifying paradigm. 265 00:09:52,355 --> 00:09:55,600 Big initiatives, small MVP, quickly validated 266 00:09:55,600 --> 00:09:56,858 in the marketplace. 267 00:09:56,858 --> 00:09:59,048 That's the Lean Startup cycle. 268 00:09:59,048 --> 00:10:01,364 In order to support that, you can find on SAFe 269 00:10:01,364 --> 00:10:03,512 this Epic Hypothesis Statement. 270 00:10:03,512 --> 00:10:05,004 For those of you who might have been around 271 00:10:05,004 --> 00:10:06,224 in earlier times, 272 00:10:06,224 --> 00:10:08,128 we used to have the Epic Value Statement, 273 00:10:08,128 --> 00:10:09,516 but it was a little flat. 274 00:10:09,516 --> 00:10:12,129 It kind of said, here's the value, we're sure. 275 00:10:12,129 --> 00:10:13,990 Now we look at it slightly different. 276 00:10:13,990 --> 00:10:16,710 We continue with our more derived 277 00:10:16,710 --> 00:10:18,456 product position statement. 278 00:10:18,456 --> 00:10:20,219 But we look at it a little bit different and we say, 279 00:10:20,219 --> 00:10:23,165 okay, what is the outcome hypothesis? 280 00:10:23,165 --> 00:10:26,811 And then most importantly, or at least equally importantly, 281 00:10:26,811 --> 00:10:28,776 we talk about innovation accounting, 282 00:10:28,776 --> 00:10:31,590 which I mentioned early on in Lesson 3: 283 00:10:31,590 --> 00:10:33,359 The Lean Agile Principles. 284 00:10:33,359 --> 00:10:36,121 How will we know before we've spent all the money, 285 00:10:36,121 --> 00:10:38,485 that the hypothesis is tracking? 286 00:10:38,485 --> 00:10:42,681 A fundamental change in the paradigm of really taking 287 00:10:42,681 --> 00:10:46,785 the concepts of Agile to the full enterprise level 288 00:10:46,785 --> 00:10:50,464 and allow us to flex, invest a little, invest a little more, 289 00:10:50,464 --> 00:10:51,932 decide not to invest. 290 00:10:51,932 --> 00:10:55,099 As we remember, when we pivot there's no blame, 291 00:10:55,099 --> 00:10:58,479 there's no moral outrage, there's no guilt, 292 00:10:58,479 --> 00:11:00,078 it's just the right thing to do. 293 00:11:00,078 --> 00:11:02,179 It's a really big deal to start converting our enterprise 294 00:11:02,179 --> 00:11:04,618 to a truly Lean Agile enterprise, 295 00:11:04,618 --> 00:11:06,393 which has the ability to experiment, 296 00:11:06,393 --> 00:11:08,844 make small experiments, and get feedback really quickly. 297 00:11:08,844 --> 00:11:11,598 An Epic Hypothesis Statement is just a template. 298 00:11:11,598 --> 00:11:14,573 But it really does drive that kind of behavior. 299 00:11:14,573 --> 00:11:16,480 You know, let's take a minute. 300 00:11:16,480 --> 00:11:19,423 Epics are kind of abstract, and kind of hard to think about. 301 00:11:19,423 --> 00:11:20,789 They're just big initiatives. 302 00:11:20,789 --> 00:11:22,360 They're different from strategic themes. 303 00:11:22,360 --> 00:11:23,711 How are they different? 304 00:11:23,711 --> 00:11:25,199 Take a minute right now. 305 00:11:25,199 --> 00:11:28,335 Identify an Epic from one of your contexts. 306 00:11:28,335 --> 00:11:29,168 What is it? 307 00:11:29,168 --> 00:11:31,328 Maybe it's 64-byte back-office servers. 308 00:11:31,328 --> 00:11:34,220 It could be technical Epic, could be a business Epic. 309 00:11:34,220 --> 00:11:35,991 Write it down in this format. 310 00:11:35,991 --> 00:11:37,229 For, who, the, is. 311 00:11:37,229 --> 00:11:39,438 That's a standard format that we use a lot 312 00:11:39,438 --> 00:11:42,499 for defining initiatives and future value. 313 00:11:42,499 --> 00:11:46,282 Now, what's the outcome's hypothesis for that thing? 314 00:11:46,282 --> 00:11:49,519 How will you know that you've achieved the business result 315 00:11:49,519 --> 00:11:52,002 long before the ROI comes in? 316 00:11:52,002 --> 00:11:55,672 What could an MVP be that could validate that? 317 00:11:55,672 --> 00:11:58,457 So what's the minimum amount you would have to do 318 00:11:58,457 --> 00:12:00,752 to prove that your hypothesis is correct? 319 00:12:00,752 --> 00:12:02,435 So take a couple of minutes and write one down. 320 00:12:02,435 --> 00:12:03,738 Doesn't take very long. 321 00:12:03,738 --> 00:12:05,482 I'll check back in in about 3-5 minutes, 322 00:12:05,482 --> 00:12:08,947 and we'll see what you came up with. 323 00:12:08,947 --> 00:12:11,561 Okay, I suspect that you wrote one out. 324 00:12:11,561 --> 00:12:13,328 And you probably had to think a little deeper thoughts 325 00:12:13,328 --> 00:12:15,674 about the MVP and the outcomes hypothesis and 326 00:12:15,674 --> 00:12:17,005 the innovation accounting measures. 327 00:12:17,005 --> 00:12:17,930 Great. 328 00:12:17,930 --> 00:12:19,267 It's a moral victory for all of us. 329 00:12:19,267 --> 00:12:21,394 We're now thinking differently about the problem. 330 00:12:21,394 --> 00:12:24,220 In the same way that we want our executives and leaders 331 00:12:24,220 --> 00:12:26,032 to think differently about the problem. 332 00:12:26,032 --> 00:12:28,379 We don't want to think about big bang, waterfall, 333 00:12:28,379 --> 00:12:30,912 proven business case, guaranteed ROI. 334 00:12:30,912 --> 00:12:32,433 We want to think about the ability to innovate 335 00:12:32,433 --> 00:12:33,681 at the portfolio level. 336 00:12:33,681 --> 00:12:36,165 And SAFe 4.5 delivers that capability. 337 00:12:36,165 --> 00:12:39,199 Now, okay so let's say, how are we gonna approve this work? 338 00:12:39,199 --> 00:12:42,726 Well, okay, we've already allocated the funding 339 00:12:42,726 --> 00:12:44,730 to the value stream, but now we're going to have to 340 00:12:44,730 --> 00:12:47,880 go into that value stream and say, okay, but some of that 341 00:12:47,880 --> 00:12:49,714 has to be allocated to this. 342 00:12:49,714 --> 00:12:51,837 So we want to take that Epic and invest 343 00:12:51,837 --> 00:12:52,743 a little bit more in it. 344 00:12:52,743 --> 00:12:55,641 It's not a simple matter of weighted shortest job first. 345 00:12:55,641 --> 00:12:57,690 That's one measure you can establish with an Epic, 346 00:12:57,690 --> 00:12:59,633 but prioritization requires other things. 347 00:12:59,633 --> 00:13:01,225 Like, where are you gonna get the return, 348 00:13:01,225 --> 00:13:02,934 and what makes you think you're gonna do that. 349 00:13:02,934 --> 00:13:06,793 So if you go to the site and download this free document. 350 00:13:06,793 --> 00:13:09,650 It's a template for the Lean Business Case. 351 00:13:09,650 --> 00:13:12,411 And the Lean Business Case is derived from Lean thinking 352 00:13:12,411 --> 00:13:15,108 and also traditional business case thinking, 353 00:13:15,108 --> 00:13:16,715 but it's two pages maximum. 354 00:13:16,715 --> 00:13:18,893 It will force you to fill in some things like 355 00:13:18,893 --> 00:13:20,723 what's the outcome's hypothesis, 356 00:13:20,723 --> 00:13:22,642 what's in scope, what isn't. 357 00:13:22,642 --> 00:13:24,100 64-byte back-office servers, 358 00:13:24,100 --> 00:13:25,961 well what about for the CRM system. 359 00:13:25,961 --> 00:13:27,087 No, we don't care about that. 360 00:13:27,087 --> 00:13:28,911 Well we care about it we just don't need to upgrade it. 361 00:13:28,911 --> 00:13:29,744 No problem. 362 00:13:29,744 --> 00:13:30,908 Okay, that's out of scope. 363 00:13:30,908 --> 00:13:33,324 It will cause you to think about the parameters of that. 364 00:13:33,324 --> 00:13:35,227 Now you have a thing that's defined that says 365 00:13:35,227 --> 00:13:36,766 now I can consider it. 366 00:13:36,766 --> 00:13:39,171 I want to consider this with respect to what? 367 00:13:39,171 --> 00:13:40,149 By itself? 368 00:13:40,149 --> 00:13:43,181 No, with respect to other opportunities. 369 00:13:43,181 --> 00:13:46,297 In order to do that, we're gonna make that work visible. 370 00:13:46,297 --> 00:13:48,563 And we're gonna exploit the incredible power 371 00:13:48,563 --> 00:13:51,651 and visibility of Kanban, and where necessary, 372 00:13:51,651 --> 00:13:52,853 work-in-progress limits 373 00:13:52,853 --> 00:13:54,830 to make sure that we see the work coming, 374 00:13:54,830 --> 00:13:56,297 we know what the work looks like, 375 00:13:56,297 --> 00:13:57,719 we have a rough idea of the scope, 376 00:13:57,719 --> 00:13:59,701 and we can start to anticipate. 377 00:13:59,701 --> 00:14:02,460 Now maybe I'm down in the trenches as a developer, 378 00:14:02,460 --> 00:14:04,904 and I'm seeing single sign-on in that portfolio. 379 00:14:04,904 --> 00:14:06,484 It's coming my way. 380 00:14:06,484 --> 00:14:09,575 I make decisions every day that could make it easier 381 00:14:09,575 --> 00:14:11,894 or harder to implement single sign-on. 382 00:14:11,894 --> 00:14:13,701 If I know it's probably coming, 383 00:14:13,701 --> 00:14:14,905 I'll make the right decision. 384 00:14:14,905 --> 00:14:17,508 If I have no idea, oh they're meeting independently, 385 00:14:17,508 --> 00:14:18,762 they've got some new ideas up there 386 00:14:18,762 --> 00:14:20,052 but we don't know what they are, 387 00:14:20,052 --> 00:14:22,001 I'm gonna tend to make the wrong decision. 388 00:14:22,001 --> 00:14:23,900 So we build a Portfolio Kanban for this 389 00:14:23,900 --> 00:14:26,509 and then we use that to manage the flow of Epics. 390 00:14:26,509 --> 00:14:27,970 You can see the benefits here. 391 00:14:27,970 --> 00:14:29,245 It makes them visible. 392 00:14:29,245 --> 00:14:30,903 It brings some structure to it. 393 00:14:30,903 --> 00:14:32,898 There's more than one opportunity folks, 394 00:14:32,898 --> 00:14:35,889 so this MVP here versus this MVP here, 395 00:14:35,889 --> 00:14:37,471 which do we think is better. 396 00:14:37,471 --> 00:14:40,822 Also, the fact that it's in a Kanban system says, 397 00:14:40,822 --> 00:14:42,968 you're not working on it yet. 398 00:14:42,968 --> 00:14:45,029 We're doing some analysis work maybe, 399 00:14:45,029 --> 00:14:46,692 we're doing some prototyping work. 400 00:14:46,692 --> 00:14:50,442 But it's not yet being implemented at the ART level. 401 00:14:50,442 --> 00:14:53,455 And that means that we don't have unrealistic expectations. 402 00:14:53,455 --> 00:14:55,979 This particular Kanban was driven by a vignette 403 00:14:55,979 --> 00:14:57,722 where I was working in the field with an 404 00:14:57,722 --> 00:15:00,663 enterprise that was very sharp and pretty Agile. 405 00:15:00,663 --> 00:15:02,832 But because the thing had been defined, 406 00:15:02,832 --> 00:15:04,904 they assumed it was getting executed. 407 00:15:04,904 --> 00:15:07,460 And we went through a planning process to determine 408 00:15:07,460 --> 00:15:09,885 how much they could actually implement, 409 00:15:09,885 --> 00:15:11,497 they were worlds and worlds apart. 410 00:15:11,497 --> 00:15:13,334 They had scoped and committed 411 00:15:13,334 --> 00:15:15,205 probably three or four times the amount of work 412 00:15:15,205 --> 00:15:17,259 that could actually be done by the teams, 413 00:15:17,259 --> 00:15:19,700 because they didn't understand the team's local context, 414 00:15:19,700 --> 00:15:22,548 and they hadn't done the right and proper job of analysis. 415 00:15:22,548 --> 00:15:24,203 Now Kanban systems are configurable. 416 00:15:24,203 --> 00:15:26,017 You can make them be whatever you want. 417 00:15:26,017 --> 00:15:29,154 Our out-of-box one, if you will, has a few simple states. 418 00:15:29,154 --> 00:15:31,487 It is a state called the funnel which is ... 419 00:15:31,487 --> 00:15:32,553 Here's an idea. 420 00:15:32,553 --> 00:15:34,140 If your boss comes along and says, 421 00:15:34,140 --> 00:15:36,595 we should think about this big thing, say 422 00:15:36,595 --> 00:15:39,053 agree with that, I think we should think about it too, 423 00:15:39,053 --> 00:15:40,322 and put it in the funnel. 424 00:15:40,322 --> 00:15:42,205 That's not the same as doing it. 425 00:15:42,205 --> 00:15:43,640 Go through a reviewing process. 426 00:15:43,640 --> 00:15:45,637 We look at the Epic Hypothesis Statement. 427 00:15:45,637 --> 00:15:47,912 What is the Epic and what do we think about that? 428 00:15:47,912 --> 00:15:49,857 Absolutely use WSJF. 429 00:15:49,857 --> 00:15:54,243 WSJF is too light a tool for making $5 and $10 million, 430 00:15:54,243 --> 00:15:55,670 or $30 million decisions. 431 00:15:55,670 --> 00:15:57,916 It's the right tool for making feature decisions, 432 00:15:57,916 --> 00:16:01,019 but it's one parameter of a prioritization scheme 433 00:16:01,019 --> 00:16:02,567 for the larger scale Epics. 434 00:16:02,567 --> 00:16:04,196 We'll move those through. 435 00:16:04,196 --> 00:16:05,350 We'll analyze them, we'll figure out 436 00:16:05,350 --> 00:16:06,672 what the alternatives are. 437 00:16:06,672 --> 00:16:08,114 Different WSJF. 438 00:16:08,114 --> 00:16:09,966 Identify the MVP. 439 00:16:09,966 --> 00:16:12,324 Get them in the backlog and then we'll decide 440 00:16:12,324 --> 00:16:14,042 the go-no go right here. 441 00:16:14,042 --> 00:16:16,270 Are we gonna do this or not? 442 00:16:16,270 --> 00:16:19,658 We'll decide based upon the capacity to actually implement. 443 00:16:19,658 --> 00:16:21,259 And you can have a definition of done. 444 00:16:21,259 --> 00:16:23,569 It might just be that you run the algorithm 445 00:16:23,569 --> 00:16:25,533 until such time as the features aren't there. 446 00:16:25,533 --> 00:16:26,977 Or achieve the business outcomes. 447 00:16:26,977 --> 00:16:28,204 That's totally up to you. 448 00:16:28,204 --> 00:16:31,577 But this Kanban system causes a collaboration 449 00:16:31,577 --> 00:16:33,035 amongst the right people. 450 00:16:33,035 --> 00:16:35,136 We haven't talked about Epic owners yet. 451 00:16:35,136 --> 00:16:36,855 They're just people in the enterprise. 452 00:16:36,855 --> 00:16:37,831 Portfolio people. 453 00:16:37,831 --> 00:16:40,773 Could be program manager, a system architect, 454 00:16:40,773 --> 00:16:43,613 a chief engineer, anybody that has some technical 455 00:16:43,613 --> 00:16:45,174 background and knowledge, and knows how to 456 00:16:45,174 --> 00:16:47,298 get things done around here, big things. 457 00:16:47,298 --> 00:16:49,312 They might take that Epic and babysit it. 458 00:16:49,312 --> 00:16:52,623 Because 64-byte back-office servers are not going to 459 00:16:52,623 --> 00:16:54,766 just magically appear in production, okay? 460 00:16:54,766 --> 00:16:56,415 It's gonna take time to do that. 461 00:16:56,415 --> 00:16:58,296 We're gonna need somebody to shepherd it through. 462 00:16:58,296 --> 00:16:59,826 They're going to work with the teams, 463 00:16:59,826 --> 00:17:01,838 the solution and systems architects, 464 00:17:01,838 --> 00:17:03,878 the business owners and product management, 465 00:17:03,878 --> 00:17:06,737 to determine what the right set of decision is 466 00:17:06,737 --> 00:17:09,189 with respect to getting that Epic done. 467 00:17:09,189 --> 00:17:11,549 Visualizing the highest level work is one of the most 468 00:17:11,549 --> 00:17:13,575 important things you can do in the large enterprise. 469 00:17:13,575 --> 00:17:16,514 Because it gives you a chance to see the big stuff coming. 470 00:17:16,514 --> 00:17:19,548 For most of us, when we take all the ideas that we've heard 471 00:17:19,548 --> 00:17:21,118 back to Jeffrey Moore's statement, 472 00:17:21,118 --> 00:17:24,008 we don't have a really great way to reason about strategy. 473 00:17:24,008 --> 00:17:26,333 So in that office that guy thinks 474 00:17:26,333 --> 00:17:29,167 this is a great initiative and I probably agree. 475 00:17:29,167 --> 00:17:31,982 In that office this person thinks it a great initiative 476 00:17:31,982 --> 00:17:33,644 and he probably agrees. 477 00:17:33,644 --> 00:17:35,692 And the other office there's another initiative. 478 00:17:35,692 --> 00:17:36,906 Nobody argues with that. 479 00:17:36,906 --> 00:17:38,906 But when you add all those initiatives up, 480 00:17:38,906 --> 00:17:41,049 and you pretend like that's your strategy, 481 00:17:41,049 --> 00:17:42,570 you may well be doomed. 482 00:17:42,570 --> 00:17:45,218 Because your appetite is bigger than your wallet, 483 00:17:45,218 --> 00:17:47,225 and you won't be able to fund those things. 484 00:17:47,225 --> 00:17:49,502 If we make that more visible, we're going to put some 485 00:17:49,502 --> 00:17:52,116 regulators on that, so you can see that work 486 00:17:52,116 --> 00:17:54,969 and make the right decisions economically.