Watch and read this 90 minute Ask Me Anything with Gene Kim, author of The Unicorn Project, from January 29, 2020. Recorded for the IT Revolution Book Club.
1. Johan Tegler (jump to question in video) — A mix of questions: Technology. How common is that word these days compared to IT? In companies producing cars, planes etc technology can be a lot more. —A— Code vs databases. How much work in TUP is about developing code vs developing databases? It’s not so easy to have databases loosely coupled. Scott Ambler with Agile Data (and Disciplined Agile) has more info. —B— Different types of databases. Is there some idea of the mix of different databases in TUP, like relational, graph, doc, search etc? How is the mix? —C— Make or buy. How much of the IT systems in TUP are developed in house and how much is bought? Suppliers can give the impression that you can buy systems have the functions you need (ERP, MRP, MES, PLM etc). —D— Visualize the teams. Is there some diagrams of the teams and the org in TUP? To better see the team topologies. —E— “The vast architecture beneath”. A good phrase in the book. How can you improve the understanding of complex software, when you are not working as a developer? Many people often underestimate software
2. Johan Tegler (jump to question in video) — TUP and TPP have a big focus on DevOps with the IT department and the business. These books describe many things that are really important and new to many people. The Goal has a focus on manufacturing. —A— What novels (and non-fiction books) can be found today in a wider area (see areas below)? Does IT Rev plans for more novels here, or knows of others who have plans?
(Jump to question in video) —B— Area 1: From Parts Unlimited to Cars Unlimited. Or from DevOps to Industrial DevOps. Product development of cyber physical systems (CPS) is often complex and dependant on software. Feedback and testing can be a lot harder. —C— Area 2. From Parts Unlimited to Hospital/Government Unlimited. Hospitals and the public sector can be complex and dependant on software. —D— I think area 1 and 2 can learn a lot from TUP. In the book The Age of Agile, Steve Denning describe some of this but as a non-fiction.
3. Georgii Ivankin (jump to question in video) — Copying from the discussions: what amuses me is how little attention is paid to Maxine’s family (at chapter 9 now). And she has kids! It seems she is not seeing them at all with all the overwork and hangouts at the bar. I wonder if this is on purpose and will change when they fix things in Parts Unlimited. However, this work-life imbalance isn’t explicit, like it was in TPP (or was it The Goal?). Can you comment on this @Gene Kim?
4. Ariel Kirshbom (jump to question in video) — Was interesting to see the inclusion of project/program management into de rebellion, as often I have seen PMs be associated with the problem and not necessarily the solution. Would like to know more on PM participation on DevOps environment and as part of the rebellion
5. Caroline Lange (jump to question in video) — You mention project Unicorn in The Phoenix Project. Did you have the idea/structure (or log ideas) for The Unicorn Project book while you were writing the Phoenix?
6. Theo Studer (jump to question in video) — Question about the Third Ideal : How much should be invest in improvement over really doing the work ? Push to an extreme it would mean producing nothing and just theorising about the best way of doing things, which is not good, right ?
8. William Judd (jump to question in video) — Kirsten is clearly a force with which to be reckoned and her expertise is held in high regard. But “project management” in software development seems to be going away (there’s no project manager in scrum, for example). Are you seeing anything like a “renaissance” of project management in areas like value stream management?
9. Kristian (jump to question in video) — How do you prevent having two departments in the same organization from having two drastically different practices in operations? Similar to how Maxine seemed to have been in a much more ready quiped team as oppose to what she experienced in the Phoenix project.
10. Keith Klundt (jump to question in video) — Dedicated, cross-functional teams probably gets the best attention in Accelerate on pp181-3 in the section that describes squads, chapters, etc, at ING Netherlands. I wonder if the Good Product Team described by Marty Cagan at SVPG is a construct that can be identified (or proven not to be critical)? In my experience, such teams, which are optimized for speed, trump other team forms.
11. JT Wall (jump to question in video) — One issue that I’ve seen organizations that have not gone through a transformation is that multiple people are effectively doing the same role. This manifests most clearly in the product management function. When Business and Technology still are considered separate from each other, then you effectively get two (or more) people providing pieces of the product management function, sometimes with conflicting overlaps. What ways have you seen this problem overcome?
12. Jerreck (jump to question in video) — Have you considered opening your own Dockside Bar?
13. Kristian Brito (jump to question in video) — You always seem to reference books in your responses, do you have a recommended book list somewhere and if so, an order in which to read those books?
14. Proctor (jump to question in video) — For clarification on the “how do we find the Erik for air cover” question, that you wanted to think about some: — In TPP: Bill quit because Steve wasn’t listening to his recommendations, and it was Erik that pushed Steve into getting Bill back and doing the Feature Freeze — In TUP: Erik was in the Redshirts meet Executive Meeting that Maggie was presenting at, pushing for Horizon 2 & 3 projects — In the budget cut meeting, Steve cites “Erik was right” (not the others) about the Horizon 3 projects — Also in the Budget Cut meeting Erik guided them on how to do the budget cut deeper and more correctly to meet the goal in regards to Core vs Contex. — Steve didn’t really seem to fully listen/accept what the others were telling him until Erik backed them up; which I am presuming was because his candidacy for being on the board of directors. — How does one find those that the executives will listen to until the rebellion has proven itself?
15. Philippe Guenet (jump to question in video) — I ran a large scale systemic workshop this week on Diversity – I am starting an effort on #Diversity4Digital based on the idea that the world of Digital is very unbalanced at the moment towards White/Male/Educated, yet Digital needs innovation and innovation needs Diversity. I am promoting conversations about diversity for people to embrace diversity rather than be forced into it (very unskilful approach from Corporates). The workshop has received fantastic echos which indicates that something has to be done on this subject and done better too. — I was wondering if there was some intentions about making the case for Diversity in TUP (apart from the obvious of having Maxine as the centre character) or if it was more about aligning to the typical standard in IT/Digital/Corporate leadership (which in reality and at Parts Unlimited seem to lack diversity).
16. Roman Pickl (jump to question in video) — How do the 5 ideals relate to the 3 ways?
17. Rainer Hansen (jump to question in video) — I have seen in many organizations a rush towards creating a solution before really trying to understand the problem. What are your recommendations for better understanding the problem and to cure such behavior.
18. Roman Pickl (jump to question in video) —What happened to Patty? She did so much of the heavy lifting in TPP.
19. Saman (jump to question in video) — Can you think of any instances where a Mainframe infrastructure team are practising in a DevOps organisation?
Welcome to the final book club AMA. This has been an incredible three weeks here that we’ve been doing this.
This has been so much fun, Alex.
Yeah, so thank you everybody for joining in. I hope the audio is great, maybe just give us a reaction on the AMA channel if everything is good, video, audio.
Okay. So this past week we had special guests, different IT revolution authors. We had authors of Agile conversations, Jeffrey and Squirrel. We had Dominica DeGrandis, we had Mik Kersten and we had Manuel Pais. Everybody coming in to give us their consulting advice to Parts Unlimited and those are all on YouTube, so go back and watch those. And today we have, as I said earlier, the final AMA. We’ve got Gene here with us live. Gene how’s it going?
I love it, that was so clever getting all the consultants to come in and make recommendations to the Parts Unlimited team. It was just great. I enjoyed the team typologies one, and I just got back from a week vacation. So I haven’t caught up on the other ones, but I’m going to do that literally right after this starting with Mik then Dominica, and Agile Conversations team. So that’s going to be great. Good job Alex. All right, so we’re coming up on the last 90 minutes and we have what? 22 questions, and so-
Right, so we’re going to try to go with the four minute pace and how are we doing on the transcripts being posted?
Those are coming. They’ll be live soon.
Okay, cool. I’m going to be looking forward to that, and then Alex, you and I have the action to follow up on a couple of the questions like for physical devices and so forth. So I’m going to be working on that over the next a week or two.
Let’s go to the first question from Johann, “A mix of questions: Technology. How common is that word these days compared to IT? In companies producing cars, planes etc technology can be a lot more. —A— Code vs databases. How much work in TUP is about developing code vs developing databases? It’s not so easy to have databases loosely coupled. Scott Ambler with Agile Data (and Disciplined Agile) has more info. —B— Different types of databases. Is there some idea of the mix of different databases in TUP, like relational, graph, doc, search etc? How is the mix? —C— Make or buy. How much of the IT systems in TUP are developed in house and how much is bought? Suppliers can give the impression that you can buy systems have the functions you need (ERP, MRP, MES, PLM etc). —D— Visualize the teams. Is there some diagrams of the teams and the org in TUP? To better see the team topologies. —E— “The wast architecture beneath”. A good phrase in the book. How can you improve the understanding of complex software, when you are not working as a developer? Many people often underestimate software.”
Okay. This is going to be a big one.
In fact, Alex, can you post that question to the channel so that everyone can see it and I’m going to do my best to briefly address all four parts of that questions. So one code versus databases. I think code versus data, I guess there is a duality of everything that can be in code can be encoded as data and vice versa. It’s actually one of the tenants of functional programming. But yeah, I think databases are at the core of almost every data center and in core to almost every application. So I think databases are not going away in probably our lifetime and but yeah, it’s a notion about databases being loosely coupled. They’ve got to have bounded contexts. I think that makes database works so important is front and center because we have to tear these big databases apart. So that’s absolutely true.
But I think the focus is very much on code. Different types of databases. I recently did a project with some type with my good friend Dr. Steven McGill, a CEO of MuseDev. And so we got to take a look at this part of the state of the software supply chain report. And so this was so fun because as you know, my favorite programming language is a Clojure that actually runs on the JVM. And so we actually got access to almost two decades of the Maven central data. So just the usage patterns, we got to look at the dependency graphs of I think the quarter million components that are in there, and the millions of versions within those components. And Alex, let’s post a video to that and the slides to that GitHub universe presentation that we did.
That was actually the first kind of project where the data were that big. I think for the five years of data that we looked at, there was like 180 terabytes of data. And in that presentation, actually at the very end, I show the scores of pages of SQL Postgres statements that Steven McGill wrote, which was terrible. And I think that was indicative that to traverse these relational… to diverse these dependency graphs, it was actually very ill suited for SQL. So this next time around we’re actually going to be using a graph database. So I think we’re going to be evaluating about four of them, including Neo4j, Datomic and a couple of other ones, so that’s going to be really exciting. And so Steve and I will be doing a podcast on that.
So that was actually my first use of a graph database. And so in the Unicorn Project, no SQL databases were featured, things kind of… I’m not sure if we’ve mentioned by name, but it was actually based on the use of Cassandra by the target team in 2014. But you could substitute almost any no SQL database in there. So make or buy, how much does IT systems in Tripwire developed inhouse and how much is bought? Well, I think…. that’s an interesting one. I bet you it’s probably like 50/50 so probably high on the commercial side, but heading even higher, right, as the core versus contexts are probably going to be focusing their application development on the truly bespoke platforms like the power… the eCommerce platform probably. But certainly things coming out of the innovation work.
So all emissions sensor work and the data ingests, I think those are going to be far more customized. I think all the places where they can’t get what they need from a vendor, there’ll be building. So the whole outsourcing, moving to a software as a service, is trying to make more engineering capacity to do that. And then D, visualize a team. Is there some diagram of the team in Unicorn Project? You know what? I was so dazzled by the Team Topologies, work on Parts Unlimited that I’m going to post those with some slight corrections in terms of where it might have varied a little bit differently from my mental model of the teams. But I thought that was a fantastic basis for that. So Alex, let’s add that to our to do list. The vast architecture beneath. How can you improve the understanding of complex software when you are not working as a developer, people under estimate software?
So I suspect Mik Kersten talked about that in his episode in Parts Unlimited and we certainly talked about it in the first episode of his podcast that just went live like last week, the Mik plus one podcast, which I thought was great. So I was the first guest, it’s coming in two parts. And if all goes according to plan his interview with Dr Carlotta Perez is going to be next. So it’s just a phenomenal piece of work. Anyway, so we actually do talk about how Mik kind of made the observation that in the 50% mark of the Unicorn Project, it was actually almost the worst case architecture possible where you couldn’t devise a more arduous and a difficult way to actually do what needed to get done.
So I think it does hint at what the real definition of architecture is, is to what extent is it made possible for developers do what needs to get done, done. So I think that’s a very different definition of architecture that certainly I grew up with. But I think that’s our degree of toil, the absence of locality and simplicity, the absence of ability to do things independently is a mark of poor architecture. And the degree to which enables it is a mark of good architecture. So anyway, sorry for the long rambling answer for the five point question. That was probably about four and a half minutes.
All right, question number two. Thanks Johann, and another question from Johann
Question two, Johann, “Unicorn Project and Phoenix Project have a big focus on DevOps with the IT department and the business. These books describe many things that are really important and new to many people. The Goal has a focus on manufacturing. —A— What novels (and non-fiction books) can be found today in a wider area (see areas below)? Does IT Rev plans for more novels here, or knows of others who have plans?”
Nothing planned at this point the other kind of the classic business novels kind of in business, I think I would take a look at all the Goldratt ones. ‘The Goal’ is his most famous, ‘Necessarily But Not Sufficient’ kind of takes the view from an ERP and an integrator perspective, which was I think, great. There’s some many, there’s ‘Critical Chain,’ all of his books really take place as novel. There was one called ‘The Promise.’ Yeah. So in my mind, my two favorites were his first two ‘The Goal’l and ‘Necessarily But Not Sufficient’.
There’re other books written by other authors. A one’s called ‘The Gold Mine ‘about lean, but man that is big, I’m guessing, that’s like 200,000 plus words. There’s one called… is actually about software development by, I think his name is Clark Chinn. There’s one about sales. Yeah, Alex, why don’t we take that on? And I’ve wanted to do that for a long time, just collect kind of the really great examples of mostly in theory constraints. There’s one on the healthcare system called… Oh gosh, I met the author. Anyway, we’ll come up with a list and I’ll post those.
“Part B— Area 1: From Parts Unlimited to Cars Unlimited. Or from DevOps to Industrial DevOps. Product development of cyber physical systems (CPS) is often complex and dependent on software. Feedback and testing can be a lot harder.”
We’ll certainly post the link that the cyber physical team did a forum, which was awesome as all of us sort of like, how do you sort of validate the software that goes into the autonomous driving systems in a car? Really, really wonderful work. I mean they basically said there’s kind of a couple of life cycles, right? You have the actual hardware that might have a three year shelf life. You might have a firmware that might need to be physically upgraded, and then there’s over the air software updates, all that operate at different cycles and have different assurance things. But all of them… well, in fact, I think that was the point, is that all of them jeopardize the assurance goals. So how do you build a software system and a software development system that makes sure that nothing can go wrong and cause the car to kill passengers or people around them.
That was just a phenomenal piece of work by Dean Leffingwell, Suzette Johnson and Robin Yeman and many others. From Parts Unlimited to Hospital Government Unlimited hospitals. Yeah, Pride and Joy, was the name of the theory constraints book about theory constraints applied to healthcare. Yeah, so Johann I owe you a more extensive catalog of books that I think do a great job in sort of educating a broad audience about all these different business processes and objectives that come with these different business processes. Great question. Okay. Three minutes, 30 seconds.
Question number three, Georgii, “Copying from the discussions: what amuses me is how little attention is paid to Maxine’s family (at chapter 9 now). And she has kids! It seems she is not seeing them at all with all the overwork and hangouts at the bar. I wonder if this is on purpose and will change when they fix things in Parts Unlimited. However, this work-life imbalance isn’t explicit, like it was in TPP (or was it The Goal?). Can you comment on this?”
Yeah, it’s interesting because that was actually a comment made during the novel development process and was actually something I worked pretty hard to address to make sure that people knew that Maxine had a life. But I don’t think the fact that Maxine had this rich family life came across because of word count pressure. So the final word count in the book is about 110,000 words and we decided to leave the functional programming parts in. And so it meant that in order to hit the word count target, which was about say 110,000 words at the max and in comparison Phoenix Project was 90,000 words. I actually did have to cut out a bunch of the family scenes. So that was kind of a last minute editorial decision. It was really a question of do we take this very scarce word count target and give it to functional programming or family and I chose functional programming.
So I would hope that it doesn’t seem like Maxine is being treated inhumanely and that she does have boundaries and she does care a lot about work life balance and that is certainly something that she advocates to the team. Anyway, so yeah, it’s a great observation and I would say it was not on purpose and in fact it’s really kind of, despite some things that I really, really tried to do with the book. A great question, Georgii.
Question number four, Ariel, “Was interesting to see the inclusion of project/program management in the rebellion, as often I have seen PMs be associated with the problem and not necessarily the solution. Would like to know more on PM participation on DevOps environment and as part of the rebellion.”
Great question. So this was very much modeled on the presentation that John Smart did at DevOps Enterprise, I think two years ago with the director of project management at Barclay’s. And so John Smart is now at Deloitte, but back then he worked closely as the head of better ways of working with the project management office to really tackle the problems and lead time that’s caused by the fuzzy front end. And so kind of the wonderful quote from his presentation I remember so well is we went Agile, we put in CICD. Yay, right. But there was only represented five to 10% of the total lead time. The rest was in budgeting approvals, proof of concept, fiscal release. And so that can take years. Alex, if you could post the link to that presentation, it was this wonderful woman, her name was not Morgan. Anyway, she told us incredible story of how at first project managers, the positions was actually being eliminated because of Agile. And then they brought them back because they realized there was actually this much larger problem that needs to be addressed that is outside the domain of the product owner or so forth.
So the fact that Kersten shows up, he’s the director of project management and he’s assistant to the analog of John Smarts co-presenter at Barclay’s and it was really meant to show kind of in the ideal what is the role that project management has and highlight how many things does project manager do that really don’t help with the advancement of delivering software in a better, faster, safer, happier way and kind of reimagine what that role should look like. So it’s probably not typical, but very much was meant to paint kind of like what it should ideally look like. And thank you for posting that Alex, it’s so great to be co-piloting this with you and I am pulling this up just so I can pull up… Morag’s name. Yeah, Morag McCall director of project and portfolio management at Barclay. So not at the project level, but at the portfolio level, which is one level higher. So great question and thank you Alex.
Question number five, Caroline, “You mention Project Unicorn in The Phoenix Project. Did you have the idea/structure (or log ideas) for The Unicorn Project book while you were writing the Phoenix?”
No, absolutely not. Maybe in the very back of my head that might’ve been a possibility kind of wrapping up the project, but by no means was actually turn that into a 110,000 word treatment ever really on my mind. The Unicorn Projects was really a distillation of all the Hero DevOps journeys I had seen at the time. So in 2013, so the period of writing that book was 2010 to 2013 and so DevOps Enterprise summit, the first one was in 2014 so I hadn’t even been exposed to that yet. So it was really based on kind of hero technical journeys at… they’re more like Google, Amazon, Facebook, Pinterest, Twitter, LinkedIn, and kind of synthesizing those stories. But then Unicorn Project, the book mirrored not so much on the tech giants and those success stories, but really how large complex organizations are being transformed.
So Caroline, I would say decisively no, on that. And in fact the source material for inspiration of that actually came from different places, right. In the Phoenix Project era, all we had were the case studies of how technology companies were transforming and in the two to three years working in the Unicorn Project that was inspired really by the success stories coming out of the DevOps Enterprise summit. So it’s interesting that kind of the study areas for those were actually different.
Maybe one last observation, Caroline, I had mentioned last time, there’s a spreadsheet of all the timelines that served as the biggest constraints in terms of writing the Unicorn Project. That was very much the case writing the Unicorn Project was trying to fit in all the demos and the sprints, and what got built in time to fit the Phoenix Project timeline was actually really quite a challenge. In fact, there was a whole category of things that I actually threw out just because they couldn’t be reasonably have been built within a month. So anyway, there was so many sort of creative constraints that were formed just as a result of the Phoenix Project timeline. Anyway, I thought you might get a kick out of that.
Question number six, Theo, “Question about the Third Ideal : How much should be invest in improvement over really doing the work? Push to an extreme it would mean producing nothing and just theorizing about the best way of doing things, which is not good, correct?”
Yes. Theo, I would absolutely agree. I’ll take a run out of that from a couple of angles. I think the best practice, the guidance that I hold in high regards is from the gentleman who wrote ‘Inspired: How To Create Products Customers Love,’ Marty Cagan. And in his book he talks about how he has trained generations of product owners to reserve 20% of capacity to pay down technical debt.
I was able to confirm this in some interactions with Twitter that was really informed… he actually in the book talks about his experience at eBay where essentially he didn’t ship a major class of feature at eBay for two years because every engineer was basically dedicated, just trying to keep the site up. So this is in the late 90s early 2000s when that made front page news when eBay would go down for hours. And so he would say essentially to paraphrase him, every product owner must take 20% of all DevOps cycles off the table. They’re not for you to spend, they are for dev and ops, and every engineer to use however they best see fit to fix problematic areas of code, to re-architect, to re-platform, to automate, to fix security issues because if you don’t do that, you’re going to spend 100% of your time paying down technical debt. So it was like a 20% tax. So I think that guidance goes a long way, especially with product owners.
I think the other thing that I’ve referenced over the last couple of Ask Me Anythings is to what extent the tech giants are dedicating their best engineers to helping improve dev productivity. So Google famously spends 1,500 developers on dev productivity and some of their most senior engineers are often people with PhDs. And in fact a lot of the papers coming out of Google are from that team, like the papers about their Blaze or Basil build systems about their static code analysis systems. I mean it really is amazing to see how much is dedicated to that. So I would say even on top of like how much time, who should be focused on sort of dev productivity as a 100% of their effort. So absolutely right. Some people call them nonfunctional requirements. I just saw a quote, a tweet from Martin Fowler about his trying to use a term cross-functional requirements. Just saying that it’s not about just nonfunctional invisible things. It’s about things that span every domain, like QA and security.
I love that phrase, but if you spend all your time on quote non-function requirements, that means you’re not actually building features that people care about. So absolutely right on. But really these are investments to increase safety and also improved speed, the speed of delivery, the speed of the ability to create new functionality that customers actually care about. So yeah, I think your mental intuition is correct, but there’s a lot of evidence to suggest that 20% is a good number. And in fact, Marty Cagan even says if there’s any technical debt extant he wants to even see 30% or more. So that’s great advice. All right. That was question number six. Four minutes, Alex.
Let’s go to question from Quint.
Question number seven, Quint, “Do you have any relationship to Burr Sutter? Yes, from red hat, right? If yes, what are your thoughts and similarities on DevOps and Agile?”
Burr Sutter, yeah, I met him before and we were joking about how we actually do look physically somewhat similar but no relationship that I know of. So that’s the first question.
Second question is like what is the relationship of agile and DevOps? Yeah, I think at the highest level can certainly say they are compatible, and I love the Agile principles. I mean I think even the form of them, right? We want to favor outcomes. What does it like? We want favor people over process, we want to… just a lovely, beautifully written set of principles. But I do think that Agile was primarily focused on software developers. So when you look at the words that were written almost 20 years ago, they do focus primarily on developers. So the functional requirements, those cross-functional requirements are kind of absent or never explicitly addressed. So I think DevOps at least explicitly says there’s something outside of just developers. And then I think a lot of the DevOps practices, do very much speak to the infrastructure and operations side in a way that the agile principles do not.
However, I think what people may not often realize is that… Give me one second to sort of frame this, and Alex, if you could sort of mark the time on this for a specific notion about Agile versus DevOps. I’d like to publish it out separately. One of the things that I think was lost in the Agile industrial complex is a focus on technical practices. So one of the things that was definitely part of Agile in around 2004 was things like pair programming, continuous integration. And that almost disappeared from the mainstream Agile literature. Certainly by the time you get to the late two thousands early 2010s and I don’t know why. But I think it’s maybe the way the certifications worked as they focused more on sort of the Agile rituals, less about the technical practices. Kent Beck, he was a phenomenal and a strong proponent of technical practices and one of the things I really like about the SAFe, scaled agile framework, is that, boy, it really elevates technical practices front and center.
And the illusion that you can be working in an Agile like way without technical practices is pretty much demolished by the SAFe framework. So that’s something I really did like about that. So yeah, whether by design, technical practices kind of disappeared in the mainstream Agile thing. But if you look back at the Kent Beck school, right? That was really there from the beginning. Anyway, this is my long way of saying they are absolutely compatible. DevOps certainly brought the ops and the cross-functional concerns to the forefront, but it’s not like they were absolutely actually missing. It was there in the beginning, but somehow got deemphasized for reasons that maybe another book will explore. Not from me, but from people who went through that whole Agile historical experience. In fact, a great person to ask for this Alex, might be Squirrel as one of the co-authors of them Cruise Control, which we use at Tripwire around 2005, right. He was there from the beginning.
I think he would actually have some incredible views on what happened to technical practices in the Agile movement, and how did people feel about that? What a cool community we have. All right. Okay. How am I doing here, Alex? Can we prompt for emojis? How are we doing? And while we wait for those responses, I’m going to keep going. In fact, I’ll ask for the emoji reactions too. If everyone by a sign of emoji respond to Alex’s question on how are we doing, that’d be great. Thank you. Okay.
Question number eight, William, “Kirsten is clearly a force to be reckoned and her expertise is held in high regard. But “project management” in software development seems to be going away (there’s no project manager in scrum, for example). Are you seeing anything like a “renaissance” of project management in areas like value stream management?”
That’s a great question. In fact, one of the things I really want to do analysis, the reason why I asked you to bookmark the time is this I got to spend, I think this is about two years ago, I had the privilege of being able to attend the SAFe training by Dean Leffingwell in Boulder, Colorado. So Dean Leffingwell is an amazing person he was a one of the forces behind the rational unified process that was almost the early 2000s, and so that was just kind of very ambitious efforts to unify the schools of thoughts. Grady Booch and the whole… I think that was a UML diagram phase of the era. And he was the brains behind the scaled Agile framework. And I’ll just say this, I have so much respect for Dean Leffingwell’s experiences and his expertise. I don’t think there’s anyone who would sit through four days of training and not learn something. I mean, I thought it was absolutely dazzling and one of the things I realized was so many of the rituals that I admired when I studied organizations like Rally software also based in Boulder, Colorado actually originated in conjunction with Dean Leffingwell.
So Ryan Martin is a co-founder of a Rally software it has a decade’s long history with Dean. So the big room planning, they had this thing called a quarterly steering where once a quarter they would assemble the top leaders of the company. So I think it was senior managers and up, or maybe it was managers and up, they would have basically like a QBR, quarterly business review. But it was like no QBR I’ve ever been a part of, always were kind of stilted. Everything was kind of pre-ordained, it was a waste of time, this was the exact opposite. It was this incredible ritual that was done in front of customers about the state of the business. They had report outs from every kind of functional area of the business, sales, marketing, operations, finance that was the first part.
The second part was goals, objectives, SWAT analysis in terms of kind of what the company challenges are. And then there was a recommendation to top leadership and then cautions and then basically an open mic critique of the day of which customers participated. It was the most magnificent thing I’ve ever seen, and that’s something that’s very much embodied in SAFe, and that came from SAFe. It was co-created with SAFe and I thought that was just amazing to see kind of like some of the rituals I so much admired how much it was co-created with Dean Leffingwell. The other… Okay. That’s why I was dazzled too, was how much I learned from the four days, and I’m not going to say that every person delivering SAFe training is Dean Leffingwell. In fact, I think the numbers showed decisively that they’re not. But if you take the train from Dean, right, I mean it was utterly a world-class amazing experience. But there was something I did mention to Dean over drinks and I asked him like two questions…
Sorry before we go there, so what also amazed me about Safe was to what extent he’s voraciously taking the best from almost every body of knowledge, whether it’s lean, DevOps, Agile and so forth. I mean, retaining the terminology and it’s so clear that from that perspective he’s voraciously taking the best of every domain. So we were talking about project management, so I got a chance to hang out with a Dean for drinks and I asked him like a couple of questions that both a couple of I’m just embarrassed by because he had such a great response.
The first question I asked him was exactly to your question, are you afraid that project managers and ITIL people are going to flock to SAFe now that their God is dead? It was based on the observation that I think ITIL project managers having come from ITIL and PMBOK or PRINCE2 or whatever, right? SAFe looks very familiar. But my question was really like, are you afraid that’s going to get misused by these project managers and ITIL folks? And his answer almost made me feel ashamed. He said, afraid, no, they’re already here.
So making the observation that a good chunk say 15 to 20% of the people attending training with me came from a project management background. And so this is great, right? This is a way to help advance their expertise, to make sure that they’re relevant and they can do more than run PRDs and MRDs through Stage-Gate processes. I mean, it was I think a phenomenal service to those communities. And incidentally, the reason why I felt ashamed was this is exactly the answer that I laughed at when I first tried to get an enterprise track at the velocity summit. So I was trying to get a track created in the velocity conference and the answer was no, we don’t want them here. And my response was, but they’re already here whether you know it or not. Anyway, so Dean kind of corrected my claim using words that I had said six years before, just for completeness’s sake and your amusement. I also asked Dean, are you embarrassed or ashamed about your work in the rational unified process?
I asked out of a place of genuine curiosity. He goes, no, it was the best we had at the time. And again, I felt ashamed and embarrassed because he made me realize if someone asked me, am I ashamed of the work I did studying high performers and the visible ops book, which is based in the ITIL processes, and I would have answered exactly like Dean did, right? He’s like, well, no, ITIL is that the best now? But it’s certainly the best language we had back then. I love the candidness and forthrightness of these answers. So something else I was going to mention about it. Yeah, I’ve been talking about scaled Agile a lot or rather the Safe framework. And this is not an endorsement, but I certainly endorse Dean. Dean is a first class intellect and I will be the first to acknowledge that. I think that what made people’s complaints about Safe I think is the fact that it’s being delivered by people who aren’t Dean Leffingwell.
So in my experience, I mean that really is kind of a QA process at the Institute of internal auditors. I got exposed to the QAR process to make sure that anyone in the audit profession can clear a certain bar according to a set of professional practices. Anyway so, I mean, I think that might be my unsolicited recommendation to this SAFe community is maybe a need for something like that. Okay. So hopefully that answered your question. Oh wait. So let me read your question again even though I met eight minutes. But it was something I wanted to talk about for a long time because I’ve heard from a couple of people that Gene hates Safe, which is absolutely not true. So yeah, what is happening to project management? Oh, so I’m sorry, before we leave this, so there is a great link somewhere that, I think it’s in the DevOps handbook. It came from Charles bets about how target eliminated… this is like in 2014, hundreds of project managers exactly for the reasons that you cited is that as teams were able to work more independently, right?
They did a lot of reduction in force in project managers so they can hire more developers. So that is absolutely a thing. And I think this is kind of res-killing is absolutely necessary. But what I think Safe recognizes and takes head on is like, hey, for major efforts like at the program portfolio management level, there is a ton of things we don’t want developers or even product owners to be doing. We want team people who are more broadly connected. It’s not a track action, and that it because we need people who have the networks that can enable a certain types of work and can escalate things quickly. So I think project management is defined by PRINCE2 and PMM are certainly disappearing. But there is certainly a role for large scale program management and that is… I mean if you look at some most successful executives in the planets, I mean they’re doing… the reason why they’re senior people is because they’ve done massive program management at a massive levels with huge consequences if things go wrong.
So hopefully that nine, 10 minute answer addresses that question. Great question and thank you because this is something I wanted to write about for a long time and this transcript will certainly help. Okay. And yes, Kersten is definitely a force to be reckoned with. A director of project management means that almost all the major initiatives in the company he was actually responsible for. So a great observation.
Question number nine, Kristian, “How do you prevent having two departments in the same organization from having two drastically different practices and operations?”
Great question. Yeah, I guess you can say there’s kind of two schools of thought, right? One is you don’t care, right? They’re running a different cloud providers, they have two different deployment pipelines and maybe that’s not a problem if they both meet the business goals maybe that’s okay. But I think it does get to be a problem if every team is on a different cloud provider and then a different route to production and has different tools.
And I would point you at these great presentation and I think I called out last session, which was the kind of the architecture presentations from HP. So this is Ralph Laura, their CIO and two of his lieutenants that really brought DevOps principles and practices into HPE before the split up. In fact, this was even before HPE, this is before the spin out of HPE and HP consumer. And what was the other presentation I had cited? Oh, it was the presentation from Levi Geinert from Target about the next generation architecture at Target and just to repeat it. The goal was kind of slot technologies into three categories. Recommended, meaning strongly recommended we have teams that love it, getting great results. What was the middle one? No recommendation or a neutral and then avoid meaning we are actively getting rid of this vendor, this technology. And Scott Prugh talked a lot about that in his presentation last year about their goal to eliminate what they called a hostile dangerous vendors from their stack.
And so they’re called out by names. It was actually kind of blurred out about what the target ones were. But I think that was their kind of compromise in terms of how do you sort of enable teams to select tools and practices to get done what needs to get done, but also sort of bring enough standardization so that the diverse choices aren’t actually jeopardizing goals. So a great question. Two minutes 40 seconds.
Question number ten, Kieth, “Dedicated cross-functional teams, that is from Keith, gets the best attention to Accelerate. Yeah, in the case study written by our good friends, that describe the squads chapters, et cetera, at ING Netherlands. I wonder if the good product team described by Mark Cagan at SVPG, right. Who wrote the inspired book, I was just referencing a couple of questions ago it was a construct that can be identified or proven not to be critical.”
In my experience such teams which are optimized for speed trump other forms. In fact, you know what? That’s been a while since I’ve seen that. Yeah. Okay. So thank you for posting the link. I’m scanning it right now and he cites Ben Horowitz’s Good Product Manager, Bad Product Manager. Gosh, this is more than I can… Let me ask. Let me take an action item to study this. This is interesting, and the reason why I’m hesitant to give an answer right now, yeah, I think even the state of the art has changed even since this blog post in 2014. By the way, one of my big lessons I learned in last couple of years is how even product owners sometimes take companies astray. The notion that I think the more modern way to say good team focus on reference customers, bad team focus on competitors is we want companies to be focused on the category, not a customer, right? Because the longterm goal is not just customer acquisition, but as you get to a certain point in time, you want them focused on the category that will become, right?
Because eventually the constraint will be the size of the market that the company competes in. And so the opposite practice is you focus on competitors as Marty Cagan to say… And then you build up a bunch of technical debt just to sort of one offs, right? So the opportunity costs of chasing that one customer is horrendous because it’s coming at the expense of the category, it’s not just the other customers, but the category of which all the customers are part of. And that’s only one of the 20 points of good team bad teams. Let me get you a more articulated response when my vision comes back, but that’s a great question. I’ll say right now that product management product is not really my area of expertise. But Alex, I have a great idea, I would love to do a podcast with John Cutler, whose area of expertise is actually this. I think that would be so interesting because I think he would have a far more pointed opinion on this.
So that’ll be my promise to you, just for those of you who don’t know were in kind of the middle stages of planning a podcast series, and we had talked about this in the Unicorn Project launch. Basically all the sensei’s that I cite in the book, I’m planning on interviewing just to be able to have them teach us directly about the things that are highlighted in the Unicorn Project. So people like Mike Nygard, Dr. Steven Spear, Dr Nicole Forsgren, Cornelia Davis, all these people who I admire we are planning to have a podcast where we have them for 90 minutes and where I get to ask them the toughest questions in this topic. So for sure let us add John Cutler to that who spoke at the DevOps Enterprise summit last time in Vegas and let me post that link, and we will definitely at a minimum ask John that question. Maybe I’ll even a post on Twitter, as prep. Sound good Alex?
Okay. That was a four and a half minute answer to basically say to punt on it, but I didn’t want to sort of convey that one piece of wisdom that blew me away. The reason why it blew me away was I wish I had known that during my 13 years at Tripwire because how many times do we get suckered into chasing a prospect or very much in the bad team behaviors as referenced by Marty Cagan. Okay.
Question number 11, JT, “One issue that I’ve seen organizations that have not gone through a transformation is that multiple people are effectively doing the same role. This manifests most clearly in the product management function. When Business and Technology still are considered separate from each other, then you effectively get two (or more) people providing pieces of the product management function, sometimes with conflicting overlaps. What ways have you seen this problem overcome?”
That’s such a great question. I mean, I’m sure this thing exists still, right? At Tripwire back then in 2004 I just kind of always gravitate towards there just because that’s where I was spending the most amount of time with R and D. We had technical product managers, we had the product manager, we had the marketing, some other management role, right? And the product manager very much like the product owner role was split across… I mean, I even visualize the calendar in my head, I see that he was split so many ways. And so the product owner in the Unicorn Project, I mean I essentially was actually based on that person where not only were there multiple people, but the single person who was responsible for liaising with the technology teams were all over the place. They were doing competitive analysis, they were in sales calls, they were doing whatever and it left no time to actually enable the engineering teams. Dr. Mik Kersten, one of the things that didn’t make it into the book, but I thought it was amazing was we talked about like how many times is the product owner hit in a day?
And the answer was something like 10 to 20 times per day, and so when the product owner can’t respond quickly, it means that there’s like 10 to 20 decisions are being deferred every day where engineering work can’t be done because a product owner can’t respond quickly. And so that was one of the problems highlighted in the Unicorn Project is that led to Maggie essentially saying, all right, you can either be a product marketing manager and leave. And I’ll find someone to be the product owner or you can be the full time product owner and I’m going to get rid of all those other responsibilities. I thought that conversation with Mik just blew me away about how critical we need that expertise, and how we need quick decisions, and availability of that person. So yeah. What was the question?
Yeah, so I would say the big harm when sort of you have ambiguous roles or this role split across a bunch of people is when you can’t get what you need quickly. In fact, I think one of the things that I took a lot of personal offense out of, and it came up over and over and over again in the SAFe training was product owners not showing me up for the demos or the business analyst or the business stakeholder. And I just thought that was like the most offensive demeaning thing you could do, right? It’s like you’re the person asking for this and you can’t even show up to see what was built for you. I just saw myself getting so angry and I love the countermeasure, which was, hey, if you can’t show up for the demo, then we’re not going to work on anything else from you, right? We’re going to focus our work on… if it’s not important enough for you to show up for the meeting to see what was built, then obvious, it’s not very important at all, right?
And so we’re going to always prioritize people, the features that actually have product owners that are engaged. So I would measure the lead time required to get things that you need. And with a very heavy focus on acceptance and those final things that are required to actually… or demo review these very important milestones that are required to sort of get to the next stage of work. And if it’s high I think that needs to be elevated, right? And the causes will be many, right? Ambiguous roles, prioritization, work burden, et cetera. But I think it will be that lead time that drives all those right discussions. And the novel piece of knowledge that I would like to impart is that those interactions, it’s probably in the order of magnitude of tens of interactions per day, right? That are happening in Slack channels and so forth. So that’s the kind of responsiveness that I think most software teams need to get what they need done, done. Great question.
Question number twelve, Jerreck, “Have you ever considered opening your own dockside bar?”
Not really. Yeah, I guess maybe like why does Eric own the Dockside bar? I thought that was actually really funny if Eric could show up in another unexpected way, right? And that was I had actually written the… he was kind of a behind the scenes a story. So there was a draft probably about a year old. Is that right? Maybe a year and a half old. So the way I wrote the Unicorn Project was you kind of write out the storyline first and as I had told before, right, and Mik told again, in his… Alex, what are we calling these? These sort of like invite the authors to consult on Parts Unlimited, what do we call in the series?
Yeah, the consulting call series or the-
So in Mik’s makes the consultant call he told the story about in January, we were meeting the night before we hang out with Chris O’Malley at Compuware in Detroit. And he’s like, how’s it going? And I was like, terrible. And he’s like, why? Because I wrote 130,000 words for the Unicorn Project and I don’t know what the point is. It’s just like a big pile of words that say nothing. And what was amazing is like 24 hours just hanging out with Mik and Chris O’Malley, I mean, it was just so clear what the Unicorn Project was about. And that’s where the five ideals came from, and so it was probably June of 2018… No, no. Sorry, wait. So that means that was January of 2019 and so the five ideals did not exist yet. Maybe even. Eric was just… Now Eric’s role was to be able to teach the five ideals. So it was probably only a few months before that probably Eric wasn’t even in the book yet. The first pass was really to sort of put down the story.
And I remember sometime, and I’m guessing it was probably October-ish of 2018 is when I merged the Dockside bar owner and Eric, and it was one of those things where I just sort of giggled endlessly writing because I’m like, no one’s going to see that coming. Although, a friend of mine just said that he called it when he was reading the book which actually shocked me. Anyway. Okay. So why does Eric have the Dockside bar? I think even in the Phoenix Project he wears a union hat, he’s very much a person of the manufacturing workforce. I just thought it’d be great if he’d had a bar where all the manufacturing workers hung out. And so it was really Curt who was actually the first sort of technology professionals hanging out at the dockside bar. It does occur to me that really DevOps Enterprise is kind of the dockside of the DevOps Enterprise community. It’s where all these technology leaders who are doing these brave courageous things, they go to gather more case studies, share the stories, and I’ve often heard get just enough inspiration to make it one more year.
I think it’s very taxing in every meaning of the word to be able to push these initiatives. And I’ve heard over and over that it was great to see your peers, be inspired by their achievements and rejuvenate to a point where you can make it just one more year. So I would say I think DevOps Enterprise really is kind of the cross industry dockside bar. I’m sorry one last thing, and I think this was in the end notes for the book. The Dockside bar was actually inspired by the Adidas story where the team that actually went on to create the platform team at Adidas was Café Intención and so the similarities between the dockside bar and Café Intención were astounding. In fact, so much part of the story I actually brought in and put front and center into the Unicorn Project. So some more trivia for you about the dockside bar. Good question. Jerreck.
Question number 13, Kristian, “You always seem to reference books in your responses, do you have a recommended book list somewhere and if so, an order in which to read those books?”
Oh man, great question. I would never suggest an order to them that’s too hard. I think it’s so dependent on the person. So I think… I would say there is a somewhat exhaustive list that is almost complete. So there’s the resource guide to the Phoenix Project, which I wrote in 2013 so I think there’s like five parts to that. Those are the kinds of recommended reading lists for the Phoenix Project. There’s also the end notes to the DevOps Handbook that’s in a GitHub repo. And Alex, if we can sort of post those out and then the part one of the Unicorn Project, a reference list and I think that’s going to be part one of three.
I’m like getting part two up. Man, those things take a long time to write. The Clojure love letter took I think 40 hours to write the part one of the Unicorn Project resource guide took 30 hours to write. That’s in my list, and so those are meant… Yeah. I’ll just state the goal that when it’s done… and Beyond the Phoenix Project, those reading lists, and the Unicorn Projects list, and Phoenix Project, and DevOps Handbook and notes. I mean I think that should represent kind of the totality of books I would recommend kind of forcefully and… I think those are just the books that certainly have influenced me the most. And DevOps handbook just because there was so many of them, maybe not all of them rise to that level, but yeah, take it for what it’s worth. But certainly, if you put the word cloud, if you create a word cloud from that, that will certainly generate a pretty complete population of books that I’d recommend.
So let me know if there are books I’ve talked about that aren’t on those lists because they certainly should be and I’ll add them to the Uniform project list. Great question. One second.
Question number fourteen, Proctor, “For clarification on the “how do we find the Erik for air cover” question, that you wanted to think about some: — In The Phoenix Project: Bill quit because Steve wasn’t listening to his recommendations, and it was Erik that pushed Steve into getting Bill back and doing the Feature Freeze — In the meeting, Steve cites “Erik was right” (not the others) about the Horizon 3 projects — Also in the Budget Cut meeting Erik guided them on how to do the budget cut deeper and more correctly to meet the goal in regards to Core vs Contex. — Steve didn’t really seem to fully listen/accept what the others were telling him until Erik backed them up; which I am presuming was because his candidacy for being on the board of directors. — How does one find those that the executives will listen to until the rebellion has proven itself?”
So Proctor, as always interesting question. Now I’m not sure is he? So I guess Eric is being considered on the list of board of directors, which Eric rejects, does not want to do right because he wants to profit more directly from the surging fortunes of our Parts Unlimited as she says in the end of the Phoenix Project. But is that why Steve listens to him? I’d like to think that Steve listens to Eric because Steve trusts Eric’s judgment as zany, and what is it? It’s like… not oracular, but he speaks like Yoda, right? Very cryptically, despite being Yoda-like Steve chooses to listen to him. So I would say I think the goal is how do we find people that people like Steve listen to? Yeah, that’s a great question. I think this is very much a project and process.
Let me take a sip of my Starbucks, nitro cold brew to ponder that question for a moment because it’s something I’ve thought a lot about. By the way, the caffeine levels in these are outrageous. In fact, you can’t buy them in larger amounts like whenever you order larger, they’re like you can only serve them in this amount because apparently they limit the caffeine quantity delivered in a drink. Anyway, so I was on a call with my buddy Jeff Gallimore, who is an advisor to me, he’s my co-host at DevOps Enterprise, he’s part of the DevOps Enterprise programing community. And believe it or not, we actually went to high school together. We reconnected after the Phoenix Project came out and he’s great because he is a very much… he’s not just visionary, but he’s just a phenomenal operations person as well. And he’s a great person to copilot DevOps Enterprise. Because it’s always great to have someone like him around where I can focus on the things I need to focus on and then give everything else to him, which is…
Alex you’ll attest that he’s a great Ops person. Yes.
That we talked about for our 2020 goals. Okay, cool. You actually broke up for a little bit there. Thanks. So we were sort of discussing kind of strategic goals for 2020 and a lot of it was how do we get even more senior people speaking at DevOps Enterprise? The most senior person we’ve had speak at DevOps Enterprise was again with Levi Geinert, but now this time at US Bank and his boss Verner Loots. He’s an EVP at US Bank and I think I mentioned this before, and it’s just amazing to me, right, that these conversations happening in DevOps Enterprise interested him. And so I have an action item to have a call with him and just understand what interested, what was valuable, and then how do we get more people like him involved because I think it’s very important for this community is how to better appeal to people like Verner Loots.
So he’s responsible for, if I remember correctly, the digital transformation efforts at US Bank and he reports directly to… it’s either one step away from the CEO or two, which is I think one of the closest we’ve ever gotten to the CEO of a company. And so Proctor, to answer your question is, I don’t know. In fact… but I think by having people like Verner co-present at DevOps Enterprise, so he’s not… By the way, what makes him interesting, he doesn’t come from a technology background. And if I recall correctly, he actually comes from a strategy background in McKinsey. And the fact that he was willing to spend three days with us and share his story with the Levi Geinert about how they’re able to create these new offerings. For example, the ability to do small business loans and go from began my application to have money in the bank in 15 minutes that’s a critical capability that’s important to the business.
I mean, that’s amazing, right? And I think a very specific objective we have for DevOps Enterprise Summit for 2020 is have more people like him speak because… to directly answer your question is how do we find other people like him and other organizations to help better accelerate these efforts? Sorry that was a long answer. I’m basically saying I don’t know yet, but I will acknowledge that the fact that we were able to get Verner to speak says that we were doing something right in the years prior to be able to get to him. However, I can give you a better answer than that. And I think this question is so important that… and even though I’m at six and a half minutes, let me take another two minutes to say I think it is by finding the highest level business person and by that I mean the person with the profit and loss responsibility to care about it. So I think in this Unicorn Project, that’s why finding Maggie was so important, right?
She had a need, she knew what the mission was, she could articulate it clearly. She knew it depended on technology and she was looking for help. And I can imagine kind of you sitting in front of your org chart and scanning the people with P and L responsibilities. And just trying to find out who that person is, and in the beginning it will not be a very senior person. It could be… We don’t want a product owner. It’s the portfolio owner, the director of portfolio management or the director of a set of products. And we had one of those speak at DevOps Enterprise London. This was from capital one, it was Amy Bechtel and John Schmidt, the director of product strategy. So he owned all the big data offerings, so sometimes you can go up through the technology manager chain. So we’ve had a couple of CIOs co-present, and I think that’s an indication that they are very much champions, right? But ultimately kind of we want the people on the business side as well. And so you can imagine you and your collaborators conspirators really trying to figure out, all right, how high in the organization can we get?
And there’s going to be a boundary where they don’t care, and so understand where that frontier is, and invest in that relationship, make sure that they know you understand what their problems are, what goals they have, what are their quarterly objectives. Understand how you can help with that, make sure you take credit for that. Earn their trust, earn the right to be asked to do more things right. Understand what their boss’s goals are, which eventually informs how we can help advance not only their goals but their bosses. Most important goals, maybe even their boss’s, bosses most important goals. And I think that is how we increase sponsorship and organization. So Proctor, if you want to talk, let me know. You know how to find me. Thank you. Nine minutes, but worth it. I think it is without doubt one of those important missions of the DevOps Enterprise community and certainly reflected in the programming objectives.
Fifteen. I notice I have very bad posture when I was reviewing these Alex, so I’m very constantly trying to make sure I’m standing up straight and the good news is like the spots in front of my eyes have disappeared, which is… Thank goodness.
Question number fifteen, Philippe, “I ran a large scale systemic workshop this week on Diversity – I am starting an effort on #Diversity4Digital based on the idea that the world of Digital is very unbalanced at the moment towards White/Male/Educated, yet Digital needs innovation and innovation needs Diversity. I am promoting conversations about diversity for people to embrace diversity rather than be forced into it (very unskilful approach from Corporates). The workshop has received fantastic echos which indicates that something has to be done on this subject and done better too. — I was wondering if there was some intentions about making the case for Diversity in TUP (apart from the obvious of having Maxine as the centre character) or if it was more about aligning to the typical standard in IT/Digital/Corporate leadership (which in reality and at Parts Unlimited seem to lack diversity).”
And by the way, thank you for the work that you are doing. I do love the… there’s a school of thought that diversity is important because you need it to overcome hidden biases and which is a reason why I think there’s a lot of empirical evidence that says more diverse teams make better decisions. And the reason is because that the different perspectives overcome the biases that color every decision we make, and I love the work that Dr. Kahneman and Tversky do in their book Thinking Fast and Slow. And I’ve got to admit I haven’t read the whole book. I mean, it is a pretty big book, but it is breathtaking, right? And it’s like the whole notion of system one, system two in the brain, the fast part of the brain, the slow part of the brain, the fast is… I mean, to massively simplify, right? It is decisions to make very quickly and cheaply. Meaning it doesn’t need to engage the system two part of the brain. The slow part of the brain, which is actually it takes a lot more wattage.
So I think the brain is like 10… how many Watts to operate the brain? Alex, I think it’s like 15 Watts or is it 80 Watts? No, can’t be, if you could look that up. So the point is the brain consumes a disproportionate part of the metabolism of the metabolic needs of what the body can actually you operate on. And so you can’t be operating in system two all the time. And so system one is what allows us to be able to react quickly, flee from predators, and make good enough decisions most of the time without having to engage the slow part of the brain. And so there’s all these biases that color system one, the fast part of the brain that the slow part doesn’t 20 Watts a human brain… Yeah, so it turns out that one of the mitigations to make better decisions is that you have more perspectives and that actually kicks in… that and the combination of other things that is a good counter measure for all these implicit biases that are introduced by the fast part of the brain.
I might be getting system one and system two mixed up, so let me just state that to validate like how important I think that is. And then plus if there’s a whole societal part of it, right. Is that we don’t want absence of opportunity to… For a social just to have a better and more just society, we do want more diversity that’s a wonderful thing for society. As well as advancing the most necessary business objectives. So that having been said, that wasn’t… gender diversity is something that was sort of maybe in the back of my head, maybe not so much in the broader context. And that’s certainly something that was informed through the DevOps Enterprise Summit was that I wanted a cast of characters that actually represented the speaker population and the attendee population. So that really I did get a lot of satisfaction out of… But Philippe, to answer your question, I didn’t actually go through any more active modes in the Unicorn Project and not because I did care about it, but it was a something that I just didn’t budget the word count for.
I even cut Maxine’s home life from the book just to show how scares those words were and by the way, to say it again, the Phoenix Project was 90,000 words. And so every word above 90,000 it was very heavily scrutinized. And so the fact that we went 15/20% over was a considerable… what do you call it? Like over budget. It was over word budget for sure. Anyway, so Philippe I hope that answers your question and absolutely… in fact I think that is something I do want to study more is being able to say more… write up more articulately why diversity is important. Okay. Question 16 how are we doing on time? We’re do good, right? We have 45 minutes to go.
I think we have 15 minutes to go.
No, no. We’re doing terrible. Holy… This is awful. Okay. Yeah, it’s funny you get a certain kind of a tunnel vision when I’m just looking at a stopwatch. All right, thank you Alex.
Question number sixteen, Roman, “How do the five ideals relate to the three ways?”
That’s a great question. So I would say the three ways are the principles from which all the observed DevOps practices can be derived from. And I think it is… I wouldn’t say timeless, but I think they definitely do seem like good principles to me. Fast flow, fast feedback, and need for continual learning. So how does the five ideals… and I’ve thought about this a lot and I don’t really have a great answer. Well let me answer it in the micro and then the macro. What is the relationship of the five ideals to each other? And I don’t know, kind of in my head, I think that the first ideal of lookout and simplicity is a prerequisite for focus flow and joy. And to get there requires… greatness isn’t free, so therefore requires improvement of daily work over the other work itself, which thus in turn requires psychological safety.
To pull the end on cord in the physical world requires psychological safety, and then customer focus. What is all in the pursuit of, so I can’t even say what that is. I mean, prerequisites, a successful relationships, I don’t really know, but it does kind of speak to me. So yeah, I can’t even answer kind of what is the precise logical relationships with the five ideals to each other beyond sort of what I just said. So if I can’t do that, I’m not sure how I can even describe the relationship that in the three ways, but maybe if I can take 30 seconds to really think about it. I mean certainly not orthogonal to them and maybe… The best explanation I’ve come up with is that the five ideals are more the experiential outcomes in a more precise description of some of the prerequisites to get there. In my mind, really kind of the two that stand out are focus, flow and joy and customer focus.
To what extent, because as Mik Kersten says, so decisively our satisfaction engagement are so tightly interwoven with our ability to get work done. That is really kind of something that is not just about making people feel good. It is about what we need to aspire to and create the systems around that enable focus, flown joy. So yeah, that’s kind of the best answer I can come up with is consistent with… just like Agile and DevOps are consistent with each other. They’re certainly not mutually inconsistent, so too are focus, flow and… the five ideals and the three ways. And the five ideals… well let me answer it this way. Someone said is the Unicorn Project basically a novelization of what I’ve learned in the Accelerate work? And in some ways I think the answer is yes, it really is. So it’s not a book about numbers and benchmarking and cross population studies and R squared scores and clustering. The Unicorn Project is really a story about extremes. What does it look like when you can’t get anything done and how does it feel?
And the opposite extreme of like what does greatness look like and feel like and how does that really help the organization win in the marketplace? So that’s a pretty good answer. I’m happy with that answer even though it is imprecise. Thank you. Roman.
Question number seventeen, Rainer, “I have seen in many organizations a rush towards creating a solution before really trying to understand the problem. What are your recommendations for better understanding the problem and to cure such behavior.”
Oh gosh, I don’t know. In fact, I think it’s possible to go overboard. I mean if you look at the theory of constraints literature and kind of what Dr. Goldratt espouses, I mean this is all about study. You’re supposed to create these enormous current reality trees, you’re supposed to create the core chronic conflicts and identify the root cause that causes all this function on every area of the current reality tree. And I spent two years doing that in preparation to write the Phoenix Project and I wrote a blog post on your constraints to training, I got at Washington state university.
But yeah, that’s asking a lot for someone to do, we don’t need that level of analysis for everything. I mean I think it really is about experimentation and the book that I think does a good job about this, as in lean enterprise, the book that was written by Jez Humble and two fellow coauthors and it is that just like… In fact, this is actually emphasized in DevOps Handbook, Jez told me that his biggest learning is that there’s this incredible… his most incredible aha moment was experimentation is needed to satisfy the customer, right? It’s only through iteration, experimentation that we can understand and truly learn what the customer needs. It’s actually the same thing for process improvement is we can’t preordain the solution, we actually have to do many, many small experiments to see what actually moves the needle. And I’d like to think that kind of Unicorn Project, that was a story of like trial and error, everything was trial and error. One of the criticism of the Phoenix Project was that everything they worked on worked the first time.
And it should be pretty clear, in the Unicorn Project everything that they worked on didn’t work the first time. And it was about focused experimentation and learning. In fact, the book by Mike Rother not learning to see that was his first book, Toyota Kata. I mean that is really a book about doing rigorous experimentation. And the book Steven Spear High Velocity Edge about the scientific method to cast a hypothesis, doing an experiment about what will… hypothesize why the problem happens. Perform an experiment, evaluate the results, and repeat. I mean, that is the scientific method. In fact, the whole DevOps Enterprise summit experience reports are also kind of very closely mirror the scientific method. What problems did you set out to solve? What did you do, where did you start, what were the outcomes, what did you learn, what problems still remain? It kind of also mirrors that. So I would say that’s probably what’s most important is we don’t want someone taking two years to characterize a problem. We actually want much more rapid experimentation to see what works. So yeah, hopefully that… I was going to say something funny.
Yeah. And I think every plenary presentation at DevOps Enterprise matches… an experience report really matches that format, then I would sort of listen to how they talk about… how they choose to start. Six minutes. Okay. We’re not going to make it all the way through, but I feel okay about it Alex. Can you do one more check? How are we doing here everybody, if you could just maybe responds to the post that Alex reactions here? Just post an emoji would be great, because I love the feedback. Thank you. There’s a question about Patty. Okay.
Question number 18, Roman, “What happened to Patty? She did so much of the heavy lifting in the Phoenix Project.”
Oh, you’re so right. And she did kind of disappear. In fact, a couple of people on Twitter said, who’s Patty? They were like… and I got a couple emails, saying Patty disappears out of nowhere which kind of surprise me because I… Okay. So let me explain why Patty disappeared. I think she appeared a couple of times. She was the person when kind of Maxine goes to bizarre world instead of operations following Jared right before she gets really sick. Patty there trying to get approvals from Maggie to push it to production, push the fix into production.
I think she appears like two other times, but there was another Ask Me Anything question about like character composition. I can’t remember exactly what the question was. Alex will link to it, but it was a lot the characters… One thing that I learned from my novel coach is you can’t have very many characters, right? To be a named character. Like you have a couple of tiers. You have the protagonist, you have all the characters who have names and it turns out you can’t have very many of them because for the reader, if you have too many named characters, you’d lose track of them. So they’re really for the important people, and then everybody else gets demoted to an unnamed character. And so Patty is a named character, but… And so what happens when you have too many named character. You start combining characters, right? So that’s the reason the bartender at the Dockside got merged with Eric, which simultaneously reduced the character count and actually elevated the bartender to be someone important. And there just wasn’t enough people to merge Patty with because she’s kind of on the peripheries of the process side.
So I don’t know what that means for Patty. Patty was the brains in the Phoenix Project, right? Every good idea in the Phoenix Project really came from primarily Patty and so it’s not like she’s not doing important work. And I guess interestingly, a lot of the responsibility and weight came from Kersten director project management and Maggie and so there was no plausible scenario to merge those characters when they were already both named characters. But yeah, so it just left… I too thought a lot about that and eventually decided it wasn’t a showstopper, but yeah, Patty is around. In fact, as you call Patty actually takes over Bill’s position when he gets elevated to CIO. So yeah, Patty has a great future ahead of her, even though she doesn’t show up explicitly in Unicorn Project. All right, Alex, three more minutes. That leaves time for one more question. Let’s do a fast one.
Question number ninteen, Saman, “Can you think of any instance where mainframe infrastructure team are practicing in a DevOps organization?”
That’s easy. Yes. Well I would go back to the Scott Prugh presentation from CSG last year and go to the video that Chris Barr gave.
He shows a… I think this is at the 20 minute mark-ish where Steve Barr he had to show the video clip of one of the engineering directors giving a talk, a celebration speech to the entire company celebrating all the achievements of the team. And they actually got their COBOL application in unit testing running in Jenkins and so that’s definitely is possible. And it’s just a great story and I think parts of the application are now running in side of GnuCOBOL in AWS, which is also amazing that actually the GnuCOBOL actually transpiles COBOL into I think C++ or straight C and is now running in the cloud. But anyway, just when you watch that video, realize that he’s actually referring to an application that is running on a mainframe that was probably written in 1974-ish.
All right Alex, we didn’t get to all the questions but I would love to tackle these in some way over the next couple of weeks to be prioritized with finishing up the Unicorn Project process list. But let me end by saying thank you so much for these awesome questions, and your participation, and your enthusiasm and your awesome questions. And maybe I’ll just say concretely why, a lot of these are questions I wanted to explicitly address but never had a venue to and like this is just an easy way to get them on the record, and my goal will be to actually get that into writing at least do the transcript, maybe in a blog post. So these are just fantastic, fantastic questions. And I think some of these questions are super, super important to me. So thank you very much. And then Alex I’m going to turn it over to you.
Yeah, thank you. Thank you everybody. We have quite the series of to do’s that I’ve logged through these 300 AMAs to be able to tackle those. It’s been amazing, we have three of these. We’ve got Gene for what? 270 minutes to just go question, after question, after question. This has been so engaging and all the questions for you guys have been amazing. There’s a question here I think that does a good rounding out towards the end is what happens now, does the book club end or? Realistically we look at this book club as a way of bringing together the community. It’s a place to appreciate these books, and to share with one another, and to learn together. So just like after the previous book club that we did, we don’t intend to close down these channels. So we do keep them open and the idea is that they can remain to be a community space, right?
Yeah. In fact, we just spent a bunch of money to actually get these all archived so they don’t disappear on us and so they will be Google searchable and so forth. So I think that’s pretty cool, right? Is that all of these will be Google searchable, et cetera. So the book club will stay open. And as soon as my… if you want me to explicitly answer anything, because I do tend to get behind tag Alex and we’ll figure out a way to… if there a backlog, maybe we’ll do another one or even just a recorded one, and then post that just because I think actually conducting these are actually the fastest way to process these questions. Yeah. So what happens to book club, it will keep ongoing until it dies, and it’s not… because we’re not going to kill it. So as long as there’s a purpose for it, it’s not going away.
Yeah. And to follow on that too, this will keep building into itself. Like Gene said, there’ll be a log. This is not the last book club either. We have more books coming out, there’ll be another one Agile conversations is coming out. And there is potential for our book club there as well, and so this Slack instance will remain open and will continue to regenerate with new book clubs. So that’s where it’s going, but today does mark the end of the Unicorn Project book club.
But yeah, perfect. I love these ongoing communities. They’re so awesome and just a little trivia fact, so the Clojure Slack channel also runs on a free Slack instance, and so someone built something to basically store all of them and be able to display things in history. So we actually engage that person with a consultant to build the same thing for the IT REV Slack channels to Ask Me Anything, the book club and the DevOps Enterprise ones. So it too is running Clojure being stored in datomic a Clojure database. So that’s pretty cool.
All right, Alex. Thank you very much. It’s up. It’ll be up very soon, within a week.
Well, we got some nice patches this year saying thanks in the AMA channel. Thank you everybody. And let’s sign off.
Take care all. Thank you. Bye bye.