Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Interviews Jeff Sutherland: Are Agile Teams Truly Agile?

Jeff Sutherland: Are Agile Teams Truly Agile?


1. My name is Michael Floyd I’m the chief editor InfoQ and I’m sitting here with Jeff Sutherland today and Jeff is, as everyone knows, one of the 17 original signatories in the Agile manifesto. We are very pleased to have you here first of all. It’s great to have you here. Just maybe to kind of start things off, I’d like to ask each of the manifesto people what are you finding interesting currently? What maybe are you working on that might be of interest in a general way?

I think the interesting thing about what’s going on is what’s next what’s of the next 10 years. We had a meeting to discuss that at the original manifesto site in February. And the key thing is as we said last night we meant what was in the Agile manifesto and there were a lot of people that are not doing it. For example getting working product at the end of a sprint is challenging and impossible for about half of the Agile teams out there and as we discussed that that is so critical it’s fundamental, you’re not even Agile until you can do that.

That a group of Agile leaders met from all over the world and they said the top focus for the next 10 years has got to be demanding technical excellence. People have got to deliver working product at the end of a sprint. That’s where it starts and it doesn’t begin until then and we are not doing our jobs as leaders until we get people on board with that; so that’s kind of the next biggest challenge. The interesting thing about that discussion is that about half this the group at snowbird felt that Agile software and the other half felt that Agile went beyond software. And the way beyond software people won the argument and so the objective is for the next 10 years is to get people with working product at the end of a sprint everywhere not just on a software team.

Now to do that we found that the thing that’s really impeding progress is you put up a development team and they start going fast and they immediately bump into other parts of the company that are slow. And so we concluded it is not enough to promote agility at that individual or the team level. Agile leaders need to lead agility throughout the organization with the catalyst for change, and many people have not stepped up to that responsibility although many people have but that’s the message that we need to communicate as Agile leaders.


2. So this is the management issue really?

So that brings you to the management okay how are we going to deal with that? Now actually what I’m finding I’ve worked with a venture of capital group now for five years we have 14 companies doing scrum and we are about to make four new investments and one of the things we do is really train the managers. I’m also finding out and just today I bumped into systematic software engineering where I’ve done work in the past there are a venture level five company in Denmark, their CEO is having a daily standup meeting every day.

This last year he finally got it that Scrum is not about the projects it’s about the company executing. I bumped into a guy that was Motley Fool where he did some workshops a couple of years back, "well how is going Motley Fool?" Well they moved Scrum throughout the company the CEO is having a daily standup. I would say a third to a half our venture company is have a daily standup with their CEO. So we are seeing that’s kind of the next wave is getting let’s get Agile out of the software and actually get it into the companies where the revenue begins okay?


3. So it’s about the communication, the transparency - those things?

We are finding that we have to completely just like we had to retread the project leaders, we have to retread the management because it’s a huge change. Steve Denning is giving a talk tonight he’s written this book about radical management. And so I was just consulting with a big gaming company last month and the CEO says to me "What do you think of this book called Radical Management?" He’s reading the book, it’s all about Scrum right? And so this gets right back to where the roots of Scrum come from. Back in ‘93 we read this paper from two Harvard business school professors talking each other knock up.

They described how the highest performing teams in the world were put together. They weren’t software teams they were automobile manufacturing, consumer products and they described the model and we said that is exactly what we need in software. We implemented that directly in software and that’s what Scrum is it comes directly out of the ideas from Takeuchi and Nonaka. Now since then the Nonaka has gone on to be one of the leading management gurus on the planet, he holds Peter Drucker’s chair who, Peter Drucker invented the concept of management.

He holds his chair at his university now that Drucker’s dead, so he really had a lot of foresight when he put this together and I’ve met him for the first time 18 years after reading his paper in Tokyo in January. And fascinating guy he is a Yoda kind of guy and I realized we’ve only adopted about 15% of what the Nonaka has to teach us and he told me he said "You know my partner Takeuchi came back to Japan and he’s been he was dean of the Hitsu Vashi business school for 20 years but he had to retire he said he had to retire, he hit retirement age and he went back to Harvard. He’s only one mile from your house you need to go back there and see Takeuchi right away".

So I get together with him and he asked me to teach a section of his course in the business school and I just finished last month doing a Scrum half day for one of these Harvard business school workshops they do for leaders from all over the world - they bring them together for two or three weeks. And Takeuchi and I talked taught Scrum to the managers, and it’s really interesting from Takeuchi’s point of view he is a really he’s an innovator in the Harvard business school curriculum. He takes all the case studies that they use for teaching and he’ll say okay for this case study this company was successful you know if you read what this Harvard school guy said you know they did this and this they were in the right market they made the right decisions and now they got a whole bunch of money.

He says but if you go and interview those people on how that company got there, it wasn’t they didn’t get there that way they started all off they didn’t know what they were doing they tried a bunch of things they got a big surprise they jumped at it and all of a sudden now they’re hugely successful. It’s all about its teams inspecting, adapting, changing and learning. He says that’s what Scrum is. Takeuchi and Nonaka, they’re not about software. They’re about Scrum, meaning teams evolving new knowledge new business and new revenues by working together inspecting and adapting and changing really fast.


4. So we are only getting 15% of the benefit?

Yes, so we are getting 15% of the benefit of this right now you know it’s like people say you only use 10% of your brain, we are only using 15% of Agile.


5. And Agile seems to be moving kind of slowly. Would that be your assessment at this point? The innovations within Agile and the community have kind of have slowed a little bit?

Take Scrum for example. The framework is very simple, it’s gotten simpler. The way that people are complaining about the latest version of the Scrum Guide that Ken and I wrote, because we took some things out its simple. The innovation comes around executing it and the product you’re building is the innovation, from that point of view we’re seeing a lot more innovation in Agile. The leading company in the world in terms of return on investment to the stock holders if you read Forbes magazine its and one of the major reasons is they are implemented of what they call the advanced development method which is a Scrum-based methodology.

They are getting a 41% rein of compound and annual return which is twice the average of what the V.C. companies make, and the venture company firm that I’m working with is interesting in doubling their return on investment using Scrum. They think Scrum as a key of doing that. So we are seeing the innovation now moving, they don’t need any more innovation they need to get working software at the end of the sprint. That’s what they need.


6. Are we failing at that?

The innovation you know there are great Agile teams out there okay and there are more and more of them, but as we move into the mainstream without we are way beyond the early adopters who were into the early majority, all the major companies. And as you do that all of a sudden it kind of gets watered down on people don’t really get it they go through the motions but as it happens.


7. It’s something they’ve been told to do and they are not really up on the philosophy behind it?

In about half of the companies it’s like that and the other half of the companies it’s great they are getting really significant benefits. So I don’t want to say bad things about anybody except there are people who tell, half of the companies I talk to tell me they’re not getting that working product at the end of the sprint. So that’s what they need to innovate: is to get it working. The real innovation is going to be new product, new business and companies that are so good they are going to put their competition out of business.

So I would say the threat is getting greater in companies that are not Agile or not implementing Agile well, the pressure is not getting less its heating up significantly. So the innovation is going to be getting it into getting, you know we find in the venture group 20% of the effort get the teams working. Teams double production immediately with us every time because we don’t mess around we don’t tolerate anything less. We then immediately go to the manager and say why isn’t any revenue doubling?

We spend the other 80% of our time fixing that because the business is broken. They can’t take advantage of the accelerated production. They don’t have the right market. They don’t have the right focus. They don’t have the right backlog. So that’s where the innovation is really going to happen is trying to get that to work but now it’s moving into the senior management getting the whole company organized.

Just as an example of things that are happening that I mean is so major. In March I was asked to come to the Pentagon because there are eight or nine committees from every part of the fence. There is a committee for all the health care in defense in every service. There is a committee for all the intelligent services so far, they have been mandated by the congress to implement iterative and incremental development in their procurement going forward. And in March they had six months to get the plan in front of congress and so the leader of DOD IT who had worked with me and had me there talking, he said "Jeff you need to come down here because the coordinating taskforce for all these defense has got to come up with this report" and we’re down there, I spent half a day with them saying okay this is what you need to do this is how you need to do it I mean we are talking about a few billion dollars of savings in the budget you guys alone could beat that this year if you could actually execute.

So how long is it going to take you and really soften the budget of the United States by eliminating the waste in the system? It’s huge. So I mean that’s the kind of thing we are seeing that in the management of the United States they’re saying "We can no longer not be Agile. It’s time to change everything we’re doing and we are ordering you to do it."


8. That’s very profound, and the government is a great example of different departments or different organizations that need to communicate and work together better. It sounds like they need to fix the departmental stuff before they can even think about the interdepartmental.

There’s a fascinating I did on my blog on this you know the Iraq war turned around and there was an article on this. It turned around because the intelligent services got together and started sharing. So they could identify something about to happen and in less than 24 hours nail that terrorist group, that’s what turned it around. And the intelligence services were told to share until it hurt. And they were afraid that they might get sued for criminal activity or something. You know a lot of the stuff that they do is that you know necessarily.


9. Seal team six kind of stuff.

Right, right but they did and that completely turned the war around. As soon as the war turned around they went back to their fiefdoms and everybody said, "We don’t like this sharing business we don’t want to be upfront. We’re going to go back to our silos and hold onto our information". So one of the questions would be is that what’s going on in Afghanistan today? So, it’s not just at the Pentagon procurement that needs to get Agile, these guys in the field need to get Agile.

And they have the same problems that companies do. I was with a guy, a VP from Fidelity Investments was on a plane with me a while back and said he was starting Scrum he said well, "You guys at Fidelity have started Scrum in half a dozen divisions already and the projects are always successful then you go back to your old ways." He turns to me and says, "The problem is always the same and Scrum is always, always the answer".

It’s when the problem goes away that we revert to our old, bad behavior because everybody wants to build their own empire, you know, have their own silo.


10. I know you’re getting together with some of the manifesto signatories coming up and looking ahead for the next 10 years or the next whatever the future of Agile is. Is that one of the things that will be talked about?

As I said, we had a discussion about that in February. I think this event here is more of looking back on the last 10 years so we had the event last night. We are talking during this week.


11. How do we look forward?

Well, I think we need to go back. I think it was a reasonable set of objectives we came up with in February. Number one is we better get people to actually providing tactical excellence. Professionals deliver working product every iteration at the end of every iteration, and not to do it is not professional. That’s where we need to stand - number one. Number two: it’s not enough to promote individual change or even team change. The organization as a whole has to change. There was also a lot of discussion about training that we need to have a better as we get into the mainstream the training needs to be better the knowledge needs to be more carefully organized better presented. So we need to upgrade our whole training capability.

And the last one we argued for four hours about until we got a unanimous agreement and the unanimous agreement was we need to optimize the whole value chain from concept to cash which is a basically Lean concept. So that means Agile leaders are about optimizing the whole organization not just about their team. And that’s going to be enough to keep us busy for the next 10 years.


12. That last points might take up this problem of the organization reverting back to old patterns?

It also addresses this idea of Agile and Lean that you are asking about.


13. There was a discussion last night, and it was a spirited one. The basic question came down to "Is Agile Lean and is Lean Agile?" Where do you fall in that discussion?

Well as I said the roots Scrum come out of Takeuchi and Nonaka. These guys have written dozens of papers on Toyota, Honda - all these organizations. They’ve written books on Toyota. When you hear them talking though their focus is on building teams that generate knowledge through inspecting and adapting. So they are all about Agile teams. And they call that Scrum that was their term not our term their term. And so they think Lean is about Scrum. Now Ken and I took Scrum and implemented it in software and we used all our software knowledge but it turns out that if the Scrum isn’t Lean it’s a bad Scrum. As soon as you start Lean-ing it out you start removing waste from the system. Scrum is all about removing impediments which is the same thing as removing waste.

So what we find in the venture group we train all the management in Lean first because it’s very hard to train them you know stop bothering your teams, "They’ll never get anything done if we don’t bother them." So we have to explain to them you General Motors was bankrupt and Toyota was number one for a while although they’ve had some problems. What was the difference? Toyota was Lean. So as a CEO you need to decide whether you want to be number one or bankrupt okay that’s in the mind of the CEO of a startup.

They always want to be number one so then we say okay well you need to be lean. So we train them in that, and then we show them Scrum and we say, Scrum is all about getting your organization Lean-ed out. Look at what we are doing it’s a pull system it’s smoothing out the workflow doing all these things which will really Lean things, so don’t do anything that’s not Lean when you’re working with a scrum team. We have them read Takeuchi and Nonaka’s book on the Toyota production system where he says, "All I did was take what I learned at Ford I went back to Japan and I removed waste."

So we tell the CEO’s remove waste, when the Scrum guys say they have an impediment that’s waste you forgot to read help that remove that. So in my mind they’re all part of the same things.


14. That wasn’t in the universal opinion last night I don’t think. I think that one of the opinions anyway was that you could be Lean in Agile but you could be Agile without being Lean. So you are saying is that Lean is at the Scrum done better?

Well you need to be careful, though, because as I said last night that the Lean experts particularly in France where I bump into a lot them have said, Lean applied as a set of procedures to remove waste implemented like business process reengineering let’s just cut jobs and whack processes and forget about the people, that is not a pretty site and it’s not a good way to work with people so the view I’m hearing from Lean experts is Lean by itself is not the way to go.

Lean has got to become Agile and Agile is all about the people, so there is a difference. But what I’m saying is Takeuchi and Nonaka they view Lean as Scrum and for them it’s all about the people. So we have this kind of American or western notion about you know we find this, yes we’ve abstracted we went over to Toyota looked at all these procedures and we brought them back and implemented them and none of the American companies are really Lean. Every time they hire the Toyota consultants to come in, in six months they can double everything, because they are not really Lean. They don’t get it. It’s not about the process, the procedure. It’s about mindset that involves building better people better teams.


15. You talk about high-performance Scrum. Is this the same thing we’re talking about now or is that something different?

Scrum was originally designed and we benchmarked it as we were putting it together and we were not happy with it until we could get 10 times industry average production. And we used Capers Jones’s tools. He was probably the leading productivity expert in the United States. He had a company called Software Productivity Research that built a toolkit where people you know big traditional projects that...


16. So these weren’t arbitrarily determined.

So this was independent third priority tooling applied to creating a process with Scrum and when it was 10 times as fast then we said okay we got it. So good scrum is 10 times as fast and people have data that shows that. When you ask okay who has that maybe 5% of the teams out there do, maybe even less but there are teams that do. And we know consistently what their level of performance is. Capers Jones is seeing the same thing with his Agile teams as I’m seeing. And Capers Jones is a function point guy. So he says 15 function points per developers per month--that’s the number, the industry average is two.

So the problem is people are saying "Well, we are doing Agile." Many of them don’t even what their velocity or production is. And when you don’t know it’s always slow, so we’ve got a tremendous amount of work there. There’s another decade of work just fixing the basics.


17. Taking this kind of in a slightly different way, it’s been said that Scrum is not being done well by all our organizations. In fact the number, well we don’t know what the number is, but by a fairly high number of organizations. So they are happy that they’ve implemented, that they are doing it, but they are not doing it well. What are the essential ingredients for doing Scrum well?

First of all not doing it well means, Scrum has got a fail-safe. Yahoo and the financial group did analysis of their data several years ago after they did their initial implementation of 250 teams. 200 of them were not doing it well but they got a 35 percent reduction in cost and that turned into 200 million dollars of pure profit. Yahoo thought that was really good. They’ll take more of that thank you. However there were 50 teams that were doing Scrum well and they were getting 10 times the benefit they were getting 350% reduction.


18. So that’s the litmus test is throughput?

Well, Scrum was created initially to make lives better for developers and yet them having more fun. As part of that building better software that people would have more fun using and it turns out to do that you have to implement things that are related to Lean and it makes your production go up. So obviously production isn’t everything but I talk about it because at the end of the day I mean I work with a V.C. group. If their investment return doesn’t go up they are not interested.


19. So Jeff, given that, what constitutes doing Scrum well? Where are the essential agreements?

You know we look for several things but maybe the top two we can hit in the short time that we have. The first thing is that they need to get to the end of the sprint with software that’s shippable. That’s very hard for a lot of people to do. And the analogy would be if you’re going to get in a boat and try to cross the ocean to Europe and you have leaks in your boat, you may never get there. The analogy of the leaks in the boat is software. If it’s not working they have defects.

The other most important thing is the backlog coming into the sprint. If it’s not in a ready state, if it’s not clear, if it’s not sized properly, if the team doesn’t understand it, they can’t estimate it, if they don’t have acceptance test. Anything that comes into the sprint, it’s kind of like getting into your boat and going across the ocean, but you get halfway across and all the food is bad, then have to go back and get restocked.

That’s exactly what happens on the team. They get halfway into the sprint, they don’t know what they are doing they have to go back and figure out what were we doing? It’s a mess. So the first two things are get shippable code at the end of the sprint, and get the product backlog coming ready state into the sprint: it’s clear; it’s prioritize; it’s sized properly. People understand it and they can go fast.


21. Was there anything else you’d like to share with us before we wind it up today?

No I just like to say you know it’s been a great 10 years I mean it’s been great for all of us that have written the Agile manifesto. I think in terms of an actually transforming companies we’ve done a lot of team transformation but in the terms of transforming companies we’re just at the beginning of the wave of that and most of the work I’m working on now in consulting and training is taking whole companies not just software development teams.

Jan 04, 2012