An Interview with Jeff Sutherland

Haegwan Kim (HK): I am going to talk with Jeff Sutherland, the inventor of Scrum software development process. Thank you very much for your time. First I want to talk about your personal life. Why did you start working in the world of engineering?

Jeff Sutherland (JS): I’ve had three different careers. I was a fighter pilot in the air force for 11 years. I went to the US military academy and then changed into the air force. That actually impacted the creation of Scrum. My last tour at the air force was teaching at the air force academy. I started to do a PhD program at the University of Colorado School of Medicine and I got out of the air force to join the faculty. Then I was a professor at a medical school for 11 years doing cancer research; I was funded by the National Institute of Health collecting cancer patient data and doing mathematical modeling of biological systems. In particular, I was modeling how a cell gets altered such that it grows into a tumor.

Then it was through the mathematical modeling and the cancer research that a large banking company came by and said, “You guys have the best technology team for what we’re using actually in the banking system for rolling out automated teller systems.” And so a company that had 150 banks in North America said, “Why don’t you come and join the bank? You have all the ideas at the medical school but not a lot of money. Over at the bank we don’t know what we’re doing and we’ve got lots of money.”

So they made me an offer I couldn’t refuse, and I wound up at the bank. They had hundreds of programmers, but I noticed that their projects are always late. When the projects are late the management is always upset. They put the developers under pressure. No matter how much pressure they put on them the projects just get later and the customers are really unhappy. So I was coming to this from a medical research background, having never done commercial software. I looked at what they were doing and I said, if I were to model this complex system as I did in my cancer research, I would note that systems stabilised at certain states in a state space. Being late was a normal state for traditional project management. So I said, “The management is just making it worse because they don’t understand what’s going on.” They put more rules, more reports and it just makes it worse and worse.

So I spent about five companies sorting out how to get the system to work right because I wanted the projects to always be early. I wanted developers to be having fun. I wanted the customers to be ecstatic. I wanted the managers to be making lots of money. So the goal of Scrum was to create so much software the managers would actually ask the developers to slow down

HK: That sounds really interesting, because the beginning of your work was not management or engineering. Were there demands for ‘Scrum’ from your boss and other people or did you just start it spontaneously?

JS: I proposed changes when I came into the bank. I actually had a grant from the medical school for a national fellowship program from the Kellogg Foundation. That was a leadership training program. So from that program I brought in ideas for the bank. In that group there were other people from business schools. They actually helped me with the business side, and of course I worked the technical side. We actually initiated within that bank a little company within a company where we started doing things that eventually became Scrum.

HK: The development of the IT industry is so fast. What kind of ability is required for IT companies?

JS: Well, as you say, things are going faster and faster. Scrum was one of the fast processes, which are now called Agile development methods. These are becoming dominant in the software industry, because on the average 65% of requirements change during the time you are developing the software. That’s industry wide data. 65% of what you are doing changes in the middle. So you have to be able to respond to change. Obviously my Air Force experience, flying fighters, and flying over North Vietnam in the Vietnamese War, trained me to respond to whatever happened. Bad things are always happening. No matter what happens, you have to change what you’re doing. That is built in to Scrum.

HK: Can that work for other companies as well? For instance, retail stores?

JS: Yes. I spent a lot of time working with a venture capital group, and they have found that Scrum works everywhere. Now they are working primarily with software companies, but they move it out of development into marketing, sales, support. The venture capital group itself used Scrum for all their work including non-software work. Scrum is designed for getting stuff done. We now have several companies, obviously it’s in the early stages compared to software, but we have several companies in the world that are not software companies that are using Scrum to run their companies.

HK: Now I know the good of Scrum but can you tell me weakness of the Scrum development process?

JS: The bad thing about Scrum is that it’s a continuous process of improvement. Scrum is driven on surfacing problems, prioritising problems and removing problems. It’s very difficult for companies to change and remove problems. So if they won’t do that they get stuck, and it becomes very painful for them because Scrum will surface all your problems.

HK: Do you call leader in Scrum as ScrumMaster right?

JS: Yes.

HK: Can you tell me what is required for them in terms of leadership?

JS: The leadership style of Scrum is a servant leadership style, a facilitated style where they are creating an environment where the team can manage itself. The best place to read about that is actually a book called, ‘Good to Great’ by Jim Collins, a business book where he looks at the best companies in the world. He finds that they have all been run by CEOs who have a servant leadership style. They try to build a great team and have the team figure out what to do, instead of telling them what to do. So that same style we want at a grass roots level with the ScrumMaster.

HK: You also worked as CTO at PatientKeeper. Are the requirements as CTO and as a leader completely different?

JS: The Chief Technology Officer of a company in the United States is responsible for the business side of things from a technology point of view. In most cases as at PatientKeeper I always like to be the technology visionary as well. So I own the development team. The CTO has to conceive of how the technology can be used to target the business needs of the company. Innovation is really important. He needs to be able to translate the technology into a sales capability for the company. So it’s a business technology transition. Plus he must be able to implement his vision.

HK: So not 100% engineer, or 100% geek? Rather being an intermediary between business and the technology?

JS: Yeah. He has to have both a business hat and a geek component.

HK: Hahaha, you also may have them. When you try to increase the company’s revenue, do you always introduce Scrum software development process? Or do you have alternative ways?

JS: Usually. For example, I’ve been VP of engineering or CTO of 11 different companies, and now with our venture group I am actually the coach for 12 companies. So I am coaching 12 CTOs as well as other people in the company. About half of what we do is work with the marketing and business side of the company. What is the business proposition? What is the marketing segmentation? Who are the users? What is the value offering to the market? What is the business plan for delivering that? That is absolutely critical. If you don’t get that right, it doesn’t matter how good your technology is. Scrum actually drives hard to create a prioritised backlog of software to be built that will maximise the business value for the company. Scrum directly approaches that. On the other side, the technology team has to produce very high quality software really fast. Scrum has been designed to produce it five to ten times as fast as a traditional team. The combination of the right business plan with a very fast, high quality delivery is what creates a winning company.

HK: It seems the process of decision-making becomes shorter and shorter. Can you give me your opinion on this progress for coming decades?

JS: Scrum is very similar to Lean development because Scrum is all about removing impediments; Lean is all about removing waste. Well, impediments are waste, so you wind up with the same thing. For example, one of the clients that uses Scrum in France is not a software company, but they are an automobile bumper manufacturing company. They have 42 plants worldwide. The time from the order of the bumper to the time it is actually installed on the car in the plant and painted the right colour and ready to ship is two hours. And their problem is that is not fast enough.

HK: Wow.

JS: That is where the world is going. In order to go that fast you need an iterative process of constantly improving both the quality and the timing. Scrum directs itself to that. So yes, things are going to get faster and faster. Things are going to get more and more automated

HK: Is there any limit for the development of administrative speed?

JS: Well there are several things going on. One is that the software component is becoming and bigger and bigger piece of the system. When I flew jet fighters, 20% of the expense of the system was on software. Now it’s about 90%. If you look at disk drives, one of my companies is EMC, one of the largest disk drive manufacturers in the world. They have thousands of people doing Scrum. What they’ve told me is the disk drives used to be 20% software components, and 80% of the value is in hardware. Now it’s reversed. It’s now 80% software and 20% hardware. The hardware is becoming really cheap; the software has to do more and more intelligent stuff to run those disk drives. This is at an imbedded systems level we are talking about. So software is going to become 95% of everything in the future. Being able to build it is going to have to become more and more automated, just like chip assemblies have become more and more automated.

HK: Can these processes be done by robots or AI? I am just afraid of the fact that all labour process will be replaced by them.

JS: Computers are getting more powerful and they are getting smarter. Computation has taken over certain tasks that people used to do. Think about the ocean, and the water is the stuff that the computers can do because they figured it out. Humans are on the land and doing stuff that computers haven’t figured out yet. Scrum has some roots in artificial intelligence and robotic theory. Some of my concerns about that are if the robots are built by traditional development teams which are under pressure, quality is poor, it could turn into a very bad experience for humans. If the teams that are building the robots, or even building the software systems, builds systems that people really like, that are really sensitive to people and they really empower people rather than turning them into slaves, it will be a completely different world. Scrum has a place to play long term. There are scientists in robotics who say by 2050 there’s going to be very little left that computers can’t do from a typical administrative side in a business.

Note: I am currently working with IBM on design of massively parallel computing systems of the future. Scrum solves the cache coherency problem on these systems. Every core memory has to understand what every other core memory is thinking. So the computers may be doing Scrum in the future.

HK: I presume technology will be a huge deal for those that make the platform for the market as now Google is doing. Can we create hardware side of platform as now Apple is trying to do with mobile device?

JS: Yes, we definitely can. The interesting thing about Apple from a Scrum point of view is that Scrum has a ScrumMaster that facilitates a team. But it is the Product Owner that’s responsible for building a business backlog prioritised by value that the team builds on. Steve Jobs is the chief Product Owner of Apple, and he is one of the best Product Owners in the world, maybe the best because he can create the right prioritised list of features. When it hits the market everybody just has to have the product. The combination of great Product Owner capability, great design and great engineering makes Apple, right now, probably the leading hardware and software company in the world as far as innovation and the quality of the products.

HK: In that light, ScrumMaster needs to, as Apple says, think different? Or to make sure that there is innovation?

JS: Part of the innovation has to be conceptual about the product. The other part has to be on the engineering side. But that’s usually coming up with different ways to do things so that from the product business side they can say, “Oh yeah, if we put it together that way then we have a really good thing.” A fascinating thing to read is the article in Wired on how the iPod got created. They actually hired a guy outside of Apple to create prototypes. It describes he brought in three different prototypes for Steve Jobs to look at. One of them he said was not going to be very good, he knew Steve was going to bomb it on that one.

HK: Hahaha.

JS: The next one’s a little better, so Steve can pick it apart, say this is better but it has got all these things wrong with it. The third one was the one he thought Steve would say, “Well that is almost it, but take the on/off button off it.” Do this, do that. And so the strategy of building multiple things from the engineering side that the product side can feed off of and make business oriented and design decisions from. That is what Apple does in such a really great way.

HK: Time moves on, so I want to have few questions about success. What is the definition of your success as an engineer?

JS: Well my definition of success has to do with making the world a better place. If I can make a contribution so that people are happier, better motivated, more successful then I’m successful. Scrum’s goal was based on work I have done with micro-enterprise lending. Professor Yunus in Bangladesh came up with a program where he’d give really poor people a little bit of money and they’d create their own business plans and then execute them. They eventually could boot strap themselves. Scrum was designed to do the same thing. Let’s give the team a little bit of capability and then let them figure out how to go really fast. So Scrum was designed to take development teams who were always late and under pressure and turn them into teams who were having a lot of fun, a lot of energy, building great products. And so the goal was actually to release the teams from their chains. They were under pressure from management; they were getting beat up all the time by management. I said to them, “You could live a better life. You could have a life where you’re innovating, where you’re having fun, where it’s a joy to come into work.” You can do it in a way that you can build so much software, you’d have five to ten times more software so you could do a lot more with a lot less work. It’s a much better life, and it empowers the people to take the initiative. That thrust is relevant to work, it’s relevant to politics, it’s relevant to the health and wellbeing of the people. That’s success. It’s to help people be great, that’s my goal. And Scrum is doing that for people. It’s amazing what’s happened. It’s a huge surprise to me. It works in every culture, it works in every country, and it works in every company. It’s just an amazing thing. And people have taken it and they’ve taken it there themselves. So it’s not me, it’s them doing it. So my success is that I’ve helped facilitate making that possible.

HK: Just mind-blowing. So your success is not about money.

JS: Here’s the secret and the core of real success. It comes from the heart, it comes from giving, it comes from expansively helping others. If you do that well you will have plenty of money. I don’t have any problems with money. I don’t have more money than Bill Gates, but I have enough money. Scrum has given me that. People all over the world want me to come and talk about Scrum everywhere all the time. Every time I need some money I just go someplace. People always want to hear about it.

HK: Can you give me your advice to be successful in general?

JS: Things today are built by teams. Obviously individuals are still a direct source of inspiration and innovation. But individuals can’t even execute really good ideas without building a really good team around them. So people need to learn to work with teams, that’s what Scrum is all about. They need to figure out how they can actually help people. We learned that in Silicon Valley. I spent some years working there and that’s affected Scrum as well. How do those companies in Silicon Valley get started? It’s always by a small team. It’s always with a great idea. It’s a world-changing idea. They are having fun at Google, and they have more money than they know what to do with. But it wasn’t for going after the money; it was actually a world changing idea and creating an environment where people could have a lot of fun executing it.

HK: Thank you so much for your time.