Aug 21, 2018
Where will Scrum@Scale take the Agile industry?
Co-creator of Scrum Jeff Sutherland sits down with @AgileToolkit to explore Scrum@Scale at Agile2018.
Connect with Jeff Sutherland on Twitter.
Bob Payne: The Agile Toolkit.
Bob: I'm your host, Bob Payne. I'm here with Jeff Sutherland, one of the co‑creators of Scrum. Now you're working on Scrum at Scale. We're very excited about that at LitheSpeed. Just wanted to hear a little bit about Scrum at Scale and where you think that's going to take the industry.
Jeff Sutherland: It actually starts a long time ago. I was pulled out of medical school in 1983. For 15 years I was a research scientist, working with tools for Bell Labs. Learned how to write software for super‑computing, modeling the human cell. Really, all the basics of Scrum were formulated in that environment.
When I was pulled out of the medical school by a big banking company they said, "Over at the medical school you guys have all the knowledge, but at the bank we have all the money."
Bob: It's not an attractive feature of the banks.
Jeff: They made me an offer that my wife couldn't refuse...
Jeff: ...the bank.
Bob: Why do you rob banks? Because that's where the money is. [laughs]
Jeff: Of course, the first thing I realize I'm working on their new technologies such as the vice president for advanced technologies. I noticed their projects are all late.
They have hundreds and hundreds of programmers and we're running 150 banks all over North America. Our customers are actually CIOs of banks. I walked into the CEO's office one day and I said, "Have you noticed that all your projects are late?" He said, "Yeah, the customers are calling me screaming every morning.
Bob: The system's perfectly designed... [laughs]
Jeff: I said, "I'm just a medical school professor, a mathematician, a computer scientist, but I've looked at their Gantt charts on the wall. You have whole walls of Gantt charts and these tasks, names, and dates.
I calculated the probability of being late using this Gantt chart and it's 0.9999999. You're virtually certain to be late on every project. It's a completely inappropriate method to manage projects with lots of change like software..
Bob: Or uncertainty.
Jeff: He said, "Well, what should I do?" I said, "Well, you authorized me to keep this grant." When I came to medical school I had this grant from the Kellogg Foundation. A leadership grant where I had to with 50 national leaders travel around the world a third of my time for three years.
To my surprise the CEO said he was going to support it. He said he thought he was going to get his money back, basically. I said, "You know, I've been running around the world and a subgroup of our leaders are actually business school professors. I've been talking to them about the bank, and they say we need to create a completely new operating model within the bank, a little company within a company, an entrepreneurial startup." I said, "That's what we ought to do."
I said, "Here's the way these operate. I need everybody sales marketing, support, installation. I'm going to split them into small teams, four or five people. We're going to have product marketing come in on Monday morning and prioritize everything they're doing by value. Friday afternoon everything is going to be done and if it's software it's live.
I need you to give me the whole operation. You and senior management can be my board of directors. I will report to you the financials once a month. The rest of the time you have to stay out of my company because I don't want you messing it up."
He said, "We'll that's not going to be as much fun as looking at new technology."
Jeff: I said, "It needs to be fixed." He said, "OK, you've got it." You've got that headache.
He gave me the worst business unit in the bank. We split them into small team. I said, "We're going to throw away the Gantt chart." I said, "We're going to train you how to land a project just like I train fighter pilots to land an aircraft.
I'm going to give you a burn down chart instead of the Gantt chart. This is back in 1983. All of this was the first prototype. It wasn't a prototype of Scrum. It was a prototype of Scrum at Scale.
That's were it began. How do you run a whole organization with Scrum? How do you create an agile environment that is actually protected? I was protecting the environment from the Waterfall company.
How do you prioritize everything for the organization and then execute those priorities as small as sprints? The teams, how are they responsible for actually delivering at the end of every sprint, which is one of the fundamentals of Scrum. I experimented with that through several companies until I finally got to Easel Corporation in 1993 where we gave it the name Scrum from...
Bob: From the paper.
Jeff: That was the first team that was called a Scrum team. Then in 1995 I pulled Ken Schwaber in. Ken Schwaber was leading a CEO of a project management software company selling waterfall methodologies. [laughs]
I said, "Ken, come on in and look at Scrum. It actually works. That stuff you're selling doesn't work."
He came in. He spent two weeks with the Scrum team and at the end of it he says, "You're really right. This is really more or less the way I run my company." He said, "If I used the Waterfall methodology to run my company I'd be bankrupt."
Bob: ...respond to the marketplace.
Jeff: I said, "Well, why don't you sell Scrum instead of selling all this Waterfall stuff?" He thought about it and he said, "Yeah, why don't we do that."
Then we decided, OK, I want to make opensource so that it can be widely deployed. We need to write the first paper on what it is. Formalize it, which we presented at OOPSLA'95.
Then Ken started selling Scrum out into the market. I was head of engineering inside companies implementing this. One of the very first companies we go together into was a company called Individual, an Internet startup that had just gone public.
We had 60 million dollars to spend. We were hiring people like crazy. We had multiple Scrum teams. What do we do? We implement Scrum at Scale.
Bob: Scrum at Scale.
Jeff: These guys went from they were delivering every six months, the first quarter we implemented what we call today Scrum at Scale, they were doing multiple releases of their flagship product and delivered two new products in three months.
Bob: That's great.
Jeff: Scrum at Scale is designed to incrementally deliver quickly with a whole organization involved in making that happen. That's the challenge of Scrum at Scale.
Bob: One of the largest early project that Sanjiv Augustine and I did we brought in for stabilization recovery from a classic Waterfall fiasco.
At the time we were using XP or Extreme Programming, but interactive and incremental. We'd executed it on a small team. We were like, "There's no reason these ideas shouldn't be simply scalable up to a larger team."
We had about 300 people on that program and we just said, "Look, we're going to integrate and demo the system," which took them two months to build before their first testing phase, "every two weeks."
Of course, they didn't do it for the first four iterations, but after that every two weeks we got a demoable system up, we had some cross‑organizational planning, and a daily stand up that then the team member would step out, and we had it in a big long hallway. We would do the Scrum and then the Scrum of Scrums on a daily basis because things were changing so fast.
Jeff: Absolutely, yeah. The other interesting thing in those first two prototypes with Scrum at Scale is that today DevOps is a big buzz word, but that was fully implemented in both of those prototypes.
In fact, in the Internet company they were having trouble with deployment. I said, "I want the deployment server and the developers cube and you guys will deploy multiple times a week. That's the way it's going to be. There isn't going to be operations blocking deployment."
It was funny because the operations team, of course, screamed, but I was the head of engineering. I said, "You guys are too slow. We're not using you anymore."
A week or two later they came back by and they said, "We've implemented Scrum in operations and we want to become part of your Scrum at Scale. We'll meet in your Scrum of Scrums daily meeting.
Jeff: Can we have out server back if we do that?"
Bob: You said, "Maybe. We'll see."
Jeff: "Only if the developers can deploy multiple times a week with the server in your server farm, if not we're taking the server back to the development area."
Bob: Multiple times a day, yeah.
Bob: That's funny. Other than the deployment what sort of engineering practiced were you guys using for test automation and that sort of thing in those early days.
Jeff: Actually in the first Scrum a team testing product was part of the Scrum because it was a small token environment so we had a fully integrated piece of code running all the time.
From the very first sprint we deployed internally. The tooling was such that our consultants could actually use it in the field to get feedback.
Jeff McKenna was working with us today as one of the Scrum trainers, wanted to start a testing company. We were particularly interested in component architectures. Move way beyond the unit testing levels. How do you test for a component? Which, today has turned into micro‑services, right? [laughs] .
How do you set up a test environment that tests the micro‑services and then make sure it's ready for deploy? All of this stuff has been around for a long time.
Bob: It has. The pattern have been there for a while. Now you've pulled it together. You're doing trainings and certification around that.
Jeff: In recent years we've been pulled in Toyota, GE, Maersk, the world's biggest shipping company, 3M we're the global trainers for 3M. We've been pulled into these big companies. I've had to coach the coaches that have gone in there.
We need to formalize the method of scaling that works in really large. We work with SAP who's implementing this. It has 2,000 Scrum teams. These people with really large implementations have told us what we need to do to make it work, not only for one or two teams but for thousands.
That's the beauty of Scrum at Scale. It will work really well for one or two teams because it has no extra overhead. It's the minimal viable bureaucracy.
Bob: The Scrum framework...
Jeff: It scales up to thousands of teams. It works just as well as for thousands with a minimal addition of any bureaucratic overhead. High performance for an organization.
3M had the biggest stock price jump in history last November. If you read "The Wall Street Journal" it's because of several technology divisions, some of which we'd implement Scrum at Scale, because it is an organizational implementation to increase the value of the organization, not just make a project better. [laughs]
Bob: There was a heated article to really springboard off of.
Jeff: About a year and a half ago the Scrum Alliance came to us and said, "We are interested in participating in a scaling framework where we have ownership and our trainers and our membership within the Scrum Alliance can actually participate in the evolution of that framework. We can't do that with the other frameworks. Can we do it with you?"
It took about a year of negotiation to decide how we're going to set this up. We have a joint venture now called Scrum at Scale LLC. It's half owned by the Scrum Alliance. It's half owned by Scrum Inc., my company.
Its goal is to train the trainers and do the whole certification process for Scrum at Scale. Within the first few months, we have 42 trainers. They're training all over the world. We're off and running.
Bob: We're excited to be on that ride with you.
Bob: Louise, Q. Is it in the class? The joke was that he was your bodyguard or something because he was very pumped up.
Jeff: Yes. It was some guys from South America or something that said "Oh, Q, he must be his bodyguard," or something.
Jeff: Q looks like...
Bob: He works out a lot.
Jeff: He wears sunglasses all the time, dark glasses. He looks like a martial artist.
Bob: He's like Bono meets Steven Seagal.
Jeff: That's pretty funny.
Bob: That's very exciting. How do you see the growth rate? You said 42 trainers in the first few months.
Jeff: We just did two train‑the‑trainer sessions last month, one in Denver and one in Boston. We'll be doing one in October in London, probably another one in January in London. We're going to be doing them in Europe.
We'll be doing another one in Boston. Walking up here, there's a guy from Austin really twisting my arm trying to get us to come down to Austin and do it.
Bob: From the Scrum gathering or near the Scrum gathering in Austin?
Jeff: Yes. There's a lot of pent‑up demand for a better alternative for a scaling framework. It's growing really fast and I think it will continue to grow.
Our challenge, mainly, is to... we'll have plenty of trainers, how do we help them fill their courses all over the world? That's a major objective to Scrum at Scale. To really get those trainers successful wherever they are.
Bob: What's been interesting in your trip to San Diego? Anything on the personal front or just business?
Jeff: I have the opportunity where we're doing some training up in Sunnyvale. My wife was with me. I said, "Why don't we just come down the West coast? We'll do a little road trip, rent a sports car. Stop at all the nice locations. [laughs]
Jeff: That's what we did coming down here to San Diego. We're about to do that to go back up again. I have a Scrum at Scale training in Sunnyvale on Monday, Tuesday. That's been fun.
Bob: It's humid here which is an odd experience in San Diego.
Jeff: Yes. [laughs]
Bob: Thank you very much, Jeff. I really appreciate the time. I look forward to doing the train‑the‑trainer course with you coming up.
Jeff: Thank you very much for inviting me.
Bob: Thank you. The Agile Toolkit Podcast is brought to you by LitheSpeed. Thanks for tuning in. I hope you enjoyed today's show.
If you'd like to give feedback or be on the show, you can ping me on Twitter. I am @agiletoolkit. You can also reach me at email@example.com.
For more free resources, transcripts of the show and information about our services, head over to lithespeed.com. Thanks for listening.