Bio Esther Derby is well known for her work in helping teams grow to new levels of productivity. She helps teams and organizations make the transition to Scrum, and coaches technical people who are making the transition to management. Esther is recognized as one of the world's leaders in retrospectives, and is co-author (with Diana Larsen) of the book Agile Retrospectives: Making Good Teams Great.
The Agile 2010 conference is created by a production team of highly respected Agile experts and practitioners to present a program that spans the whole spectrum of agile practice.The Agile conference series is a organized as a program of the Agile Alliance, a non-profit organization dedicated to uncovering better ways of developing software inspired by the values and principles of the Manifesto for Agile Software Development.
How far back do you want me to go?
I’ll start with my professional life then. I started as a programmer and I programmed for many years. My first language and the one I still love best was Assembler because it was just like solving puzzles, so I had a lot of fun writing Assembler. I was really good at finding problems in systems. If the symptom was over here, I was really good at realizing that the actual fault was over here. Since I was so good at finding problems in code, they made me a manager.
I got to get good at something else. The skills that you learn looking at code as a system and computer systems (there is a whole system) don’t directly translate to being a manager, but knowing how to see something as a system, whether it’s a human system or a computer system is a useful skill for a manager. I went from being a programmer to a manager and eventually I decided that corporate life was not for me and set up as an independent. I’ve been head of my own company and also janitor at my own company since 1997.
Exactly. I find that any time you ask a question it opens one avenue of inquiry but it closes many others. If we don’t examine our assumptions very carefully, we are always asking the questions from one set of assumptions. Most of us have been trained to ask questions that relate to budget and cost and individual performance. If we only ask those questions, we’re missing one the major areas for improving the results in our organizations. These are questions about opening up a different avenue of inquiry and looking at your organization in terms of its capacity, in terms of what its customers desire and in terms of creating work systems so that everybody can be more effective.
They can be. Let me tell you a little story about a vice-president who once came in a company where I was working and she was supposed to turn things around - she was a "No nonsense" person. She came in and her first question in endeavoring to improve the results in this organization was "Where is the dead wood?" She asked this in a very public meeting, she had an all-hands meeting, and she said "I’m going to be finding out where the dead wood is! My job is to get you to work harder. How can I get you to work harder?"
Interesting set of assumptions there, right? That there is dead wood in the company, because I don’t think they had hired any dead wood. I was curious as to what had happened that had turned people into dead wood and I was interested that she thought people weren’t working as hard as they could. Those were dangerous questions in that they led her down one path of inquiry that was fundamentally fueled by blame. - "It’s you, people!"
It feels powerful, but it does get in the way of solving any kind of problem. The blame always gets in the way of solving problems and she in fact was not very effective in solving the problems of the organizations and helping it become more effective. What she was very effective at was getting the best people out the door.
They don’t have to take it. They just looked at it and said "It’s going to be ugly around here for a while. This isn’t worth it." So a lot of people left and it wasn’t the people she probably wanted to leave, it was the people who had other options. The people who didn’t have options were often the ones who stayed. They can be very dangerous.
Let’s see. I might look at one of my cards and tell you about one of these questions. I’m just going to take one at random: "How does your company create value?" I was talking to somebody the other day and they had been looking at their ability to produce software. They had discovered that one of the biggest impediments to actually getting work done in their company was the fact that they had very strict rules about the sorts of office space people could have, about what they could have on the walls, about what height walls were permissible, based on your grade level, whether you could have certain kinds of paneling.
So they had all these rules that were really getting in the way of creating value in the company. The people who put those rules in place are often called "the furniture police." They are actually well-intentioned people who are doing what they’ve been directed to do, but they lose sight because management lets them lose sight and helps them lose sight of what the company is really trying to do and how the company creates value. So, that’s an interesting example of where that question is missing. "How does our company create value and how can we support people to create value?" When that particular question is missing, then various parts of the organization start working against creating value.
This particular set of questions is really focused on management, because they are the ones who really design the work systems. They are the designers of the work experience, of the work system, although they usually don’t think of their role that way.
Absolutely. And it’s their mental model of management that leads to structures that we work within. A lot of my questions are designed at shifting mental models because it’s not very effective to go argue with someone and tell them they’re wrong. "You have the wrong mental model how management works." That’s not effective, but questions can get people to think about this differently. On the flipside of this particular card it says "What would be different if you focused on creating value rather than counting costs?"
That’s another question that looks at changing the point of view just a little bit and getting people to do a little thought experiment about "What if we focused on getting the work out the door rather than containing the costs of office furniture or containing the costs of office moves? What would be different if we did that?" Shifting that mental model through asking questions enables people to see the system and make different choices.
I’ll take the one out of the top of the deck this time: "What decisions are you making that the team could make as well or better?" That’s another question because in hierarchies there is an assumption that the people the higher you go are more qualified and they do generally have better access to knowledge, which is sort of a problem. But the assumption is they are more qualified, they have different qualifications, not necessarily that they are better qualified to make certain decisions.
But traditionally that’s been the manager’s role: to make decisions, which in many cases creates a huge bottleneck. It’s useful to look at all the parts of the decision and where the different parts are best taken care of. We typically think of decision as a choice, the moment of choice, but decisions actually have a number of parts which is who sets the boundaries. Saying we need to be making a decision about hiring.
That may be a joint discussion between the manager and the team, but there is another part to it, which is "What are the criteria?" which maybe decided between the manager and the team. "How do we select the person? How are we going to vet the person?" Which in a self-organizing team really should be a joint decision and then "Who makes the ultimate choice?" The manager is acting as an agent of the company, so they have a legal authority to enter into a contract. And I think most companies don’t want to have teams making salary decisions (I’m just thinking).
But traditionally that’s been the type of decision that a manager has done, so rethinking that I think is very useful when you are trying to create a self-organizing team and when you are trying to get the benefits of the team effect. One thing that people often overlook is the fact that if the manager makes all the hiring decisions alone, then the team really has little investment in making that hiring decision work. The manager has a lot of investment, the team doesn’t, but if the team is part of the selection then they have an investment in making it work and helping that new person work out.
That’s just one example of a decision that traditionally managers have taken as their decision, but could very well be taken as a joint decision. You just have to be clear on who has which parts. Tactical decisions often fall into this realm where if the manager hasn’t been deeply involved in the technical aspects of the work for a while, they may not have the right knowledge or the most current knowledge to make a decision. They may have the budget authority to make it, but they may not the technical expertise any more. Getting the team involved there is helpful.
They are all open-ended. I don’t think there is a single yes/no question in this stack because closed question close off inquiry. Sometimes they are useful if you want to confirm specific information. Some of these are asking for a little bit of self-reflection and some of them are asking to gather some data. There is one question in here "What message does your reward and performance management system send? How do you know? Based on what evidence?"
This is a question that’s saying: "Go find out. Go talk to people and find out. Get some data!" Because this is an area where people have a lot of assumptions. The whole matter of rewards and performance management it’s a case of people know something that isn’t so. We have a lot of beliefs about this particular arena, but if you actually look at the research and the data, out beliefs are often not correct. Some of these are saying "Go out and get some data, test your assumptions."
Yes. Early on, when people first started talking about Agile and moving to Agile methods and then moving through the organization, there was a lot of talk about "We don’t need no stinking managers" and that might be a fun thing to say, but it’s not helpful and it’s not realistic. There are still lots things that managers have to do in organizations, because they are acting as agents on the company, they have legal responsibilities within the organization. Most companies (not all --there are some exceptions to this) don’t want random individuals or teams making higher offers or firing people or making expenditures of a million dollars and so on and so forth.
There are those fiduciary responsibilities and the legal responsibilities that managers have in organizations, to they cannot go away, there is a need for them. I think their role needs to shift. In traditional hierarchical organizations where the assumption is as you said that people higher in that organization are better qualified, there is an assumption that managers have to direct people what to do, tell them what to do. And then make sure they do it, which also has some assumptions about people’s motivation and people’s capability to manage themselves.
We hire adults and then we treat them like children: "We must monitor you, we must tell you, we must make sure you don’t make stupid and foolish mistakes." So that role goes away, if you want an empowered team that’s going to be creative and highly productive. What managers focus on then is the broader organization. They start looking at "How does this company function as a system? How do we produce value? What’s the throughput of our organization? What’s getting in the way of people doing their best work?"
Because managers are the creators of the work experience and the work system, so that’s where they need to focus their attention, along with supporting teams rather than telling teams what to do. Of course, teams are always in a context and they have to be working on what’s important to the organization. Self-organizing teams don’t get to go off and decide "We’re going to do something completely different."
The content of the work comes from above. That needs to be within the organizational context, but in a self-organizing team the team is organizing their own work and monitoring their own progress.
Right. "Go figure out how to do it!" So the manager doesn’t have that role anymore, but he does have a role supporting the team, making sure they have what they need, protecting them from the interference, stepping in when they need help. It’s always a delicate balance between when you step back and when you step in. Sometimes mangers step too far back and they say "You are self-organizing, you take care of it all, not matter what the problem is." Sometimes they step in too quickly and deprive the team of the ability to learn something important and increase their capability.
So two aspects: looking at the system and supporting the team, renegotiate their relationship with the team. I personally, as a manager, was never in favor or comfortable with telling other adults what to do and tracking their daily progress, because they’re adults. We hire adults who are capable of making huge financial decisions like buying houses and major emotional life decisions like marrying or having children and then we expect that we have to be on top of them everyday.
I was never comfortable with that aspect and always took the approach of trying to be more supportive than directive.
16. One of the things that you said about the tasks and traditionally a manager’s job was to say "You go do this!" and you say the manager’s role in Agile changes. And I hear that they create the environment, they should be very focused on creating the best environment that helps their people attain the company’s goals, getting things out of their way. What about this thing about work? They are no longer parceling it out, but do they have anything to do with the work itself and communicating the work, setting goals? Because they used to parcel it out, to take it, parcel it out and if you had a hierarchy, it would be just parceled out every single time.
There is an interesting aspect of that in that when the manager was parceling out the work, the manager had the opportunity (didn’t always take it) to say "Well, we need to have more than one person who can do this job, so I’m going to assign this work in such a way that we are building some strength across the organization here, so that we don’t have just one person who is either a bottleneck or a risk factor if they just decide to leave the organization. The manager had the ability to do that, but you can also make that a constraint on the team.
Though it’s not like the manager has nothing to do with the team, the manager sets boundaries. So you could say "Here are the risks. We have only one person who knows how to use this particular tool and we need to have more people who learn that. So, when you are assigning your work to reduce that risk and to reduce the bottleneck, let’s make sure that you figure out a way to spread that skill through the team. Not everyone needs to know it, but let’s at least spread the skill, do some cross-training."
That’s an appropriate constraint. That’s certainly part of the manager’s job, it’s part of the environment and it’s part of being a steward of the assets of the company. I don’t want to talk about the people as assets of the company, although they are an asset, the most important one in some way, but it’s managing risk within the company, which is definitely part of management’s job.
No. Thank you for asking that clarification, because when the managers are completely divorced from the work, it’s very dangerous. When managers don’t have a sufficient understanding of the work and the dynamics of the work, they start looking at proxy measures. Managers who don’t understand the work look at the amount of time people are spending in the office or the number of documents produced or the lines of code produced. They look at all these measures that have nothing to do with actually creating value.
So it is crucially important that managers still understand the work and understand the dynamics. I can still be a good manager, even though I have not written code in many years, because I understand the dynamics of work, because I stay connected enough to what technical work really is. I don’t think you want me writing code at this point in time, but if you understand the dynamics of the work, then you can see when there are issues you understand, what questions to ask and you can make more appropriate measures about what is going on with the work.
So it’s really critical that managers are not divorced from it. I mean it goes back to the old Toyota saying: "Go and find out!" You need to know what’s going on, you can’t make decisions or adjustments in the vacuum. There are actually some really interesting stories from John Seddon, who’s quite an interesting thinker, about how management decisions made in a vacuum have affected work. He does a lot of interesting work in the service sector and looks at time series studies about how long it takes to process requests and so forth and so on.
Very often, when they have a spike and suddenly things are going along like this and then it takes a large amount of time to complete. Something spike’s up because of the management’s decision that was taken without deep knowledge of how the work works.
Right. I like to say "understanding how the work works".
19. I know I’ve heard from managers who are new to adopting Agile there is still a lot of fear. They may be not saying it that way. Should that fear be there? Are they overreacting? Because it sounds like you’re describing a different job.
I am describing a different job. I think when managers have the response of fear it’s because no one is talking to them about how their job is going to change and how it’s going to play out. I remember sitting in a session describing Scrum a few years ago and the guy sitting next to me, I could visibly see him tensing up and getting further back. Finally he leaned over to "there is no place for me in this." If people feel like that it’s going to be hard for them to embrace moving forward.
People don’t resist change, they respond to change and if they don’t see a way to hold on to their value in the organization and their self-esteem, their identity then they feel fearful. It gets called "resistance" but resistance means that someone else is not doing what I want them to do with the speed and enthusiasm that I desire. That’s what resistance really means. What gets called "resistance" is a response and as you say, it’s often based on fear, because if people don’t see a place for themselves in the new order it’s a very natural response.
If you look at that as a source of information, then you can use that to help people navigate that change. I see a huge role for managers in organizations. There is a lot of work to be done to understand the dynamics of the organization, to understand how value is created and to really get to work on making our organizations as effective as they can be. Two of my favorite authors Jeffrey Pfeffer and Bob Sutton talk about the law of crappy systems and that says if you have a crappy work system, it doesn’t matter how brilliant the people you hire are, they won’t be able to do their best work.
If you have a crappy system the best you can hope for is mediocrity. I see that as a huge opportunity for the managers to do some really strategic work in creating work systems that support everybody to be more effective.
I would suggest that they start looking at Deming [W. Edwards Deming] and maybe read some Seddon to get a different sense of what their job could be, talk to me, start finding out about these 13 essential questions for managers. Because it will get them a different view on their organization and their possibilities within the organization. To some extent they may have to advocate for their own role and their own value in terms of "If I’m not doing this, this is how I can add value." And Deming and Seddon and I can give them some starting points.
Nobody’s useless. And it is incremental. Organizations do not change overnight, but you can start shifting your focus towards your team and then start looking at your horizontal level and doing some integration there and then start looking at the organization as a system.
Yes. Traditional hierarchies it’s all about looking down and then looking directly up. This is not forgetting your team, but looking more broadly across the organization and then looking into the organization as a whole.
I am. I have two that are in the queue right now. They recently changed position. The one I’m working on most directly right now is entitled "Team Traps and how to avoid them" because teams fall into a predictable set of traps. I want to lay out what they are and how to either stay out of them or dig out of them once you find yourself in them. It’s not one of those books that says "Oh, you’re in a trap! What a ninny. Tough luck!" There are a few that I’ll mention: one that we’ve talked about already is not having clear boundaries for the team when there is a revolving door of membership or people aren’t clear who the members are.
Or sometimes teams have a boundary that’s too tight and they forget that they are actually there to do the work of the organization. So that’s one trap. Another trap is withholding information - it’s not the kind of information that you would think. Because when I say that people usually think "Oh, they’re withholding important tactical information or task-related information" which is pathological. That’s obviously wrong. But the trap that teams fall into is members withhold information about their experience on the work.
I observed a group once that had a task to design a simulation and they were having a lot of trouble with arriving at options. They said they were going to do something, they said "We’re going to brainstorm" and they brainstormed for about two minutes and then some of them would start analyzing an option and then someone else would say "No, that won’t work!" Then they go out into another process. Then somebody would say "Wait! We said we were going to brainstorm" and so they’d go back to that and then they’d veer off into another direction.
It was becoming very painful for the people who were in this process, but nobody said anything until somebody said "I’m really confused about what we’re doing." And that got everybody to kind of reset and say "Oh, well, yes. We’re in a mess. Let’s just decide on one thing and we’ll stick to it." That was personal information about the experience of work that had he not said it, they could have stayed in that churn for a long time.
Can I tell you another story? Another team I was coaching, they were brainstorming something and they were getting to filtering and the person who was leading the brainstorming erased one of the ideas from the whiteboard. And I noticed this guy, they were all standing around the whiteboard and I noticed him cross his arms, which may mean that you are closing things off, it may mean you are cold, it may mean a lot of things. He crossed his arms and chin dropped and he stepped back.
I watched him for a second and I stepped closer and I said "What’s going on?" He said "They don’t value me. They just erased me from the board." I said "Yes. You have to tell them. Because they haven’t noticed it, they haven’t noticed that you stepped back." He said, "I know, they don’t value me." I said "Tell them" and he did. And the guy who had erased the idea was utterly unaware of the effect that could have. So they brought the idea back up there (it was something in a parking lot) and that guy was able to reengage in the work of the team.
But had he had withheld that information, they would have lost his participation and they would have lost his buy-in on the ultimate decision. That’s one kind of information that people withhold on teams. It’s huge.
It’s a trap that teams fall into and then people disengage and they pull back and they are no longer full participants in the team. The whole team could fall apart in those situations. I’m not saying you have to reveal everything, you don’t have to reveal the things that you would tell your therapist or your priest or your best friend. But your emotional experience of the work it’s an important part of getting the work done. The other piece of information people withhold has to do with working relationships.
I run into lots of people who feel some irritation with someone on their team, but they don’t tell them. If they don’t tell them, the irritation gets greater and greater, grows and grows until finally they blow up. I have a million of these stories, but one of them was they were in a shared queue and one of the guys had a new girlfriend and she would call him at work and leave a message and he would listen to his messages on speaker phone. So the other guy was finding out way more than he really was comfortable knowing about what was going on.
He told him "I’m really uncomfortable hearing all this information. I’m glad you’re enjoying your girlfriend, but I don’t need to know." So he found a way to talk about that, but had he not talked about it, it would have created a distance, he would have been more reluctant to talk about this guy, he would have started making up stories "This guy is an inconsiderate jerk. Doesn’t he have any manners?" And that’s what happens when people withhold that kind of information - they start making up stories in their own heads about the other person. It breaks up working relationships and it can break up a whole team. That’s team traps.
Yes. I don’t have a snappy working title for that one yet, but it’s based on my 13 essential questions for managers and those are related to the questions that we were talking about earlier, questions that will help managers start thinking about their role and thinking about the organization in a different way.
It’s been a pleasure.
Your answers in the context of Scrum
very interesting interview. Thanks for the all the insight. Personally I am currently working on how a company structure based on Scrum could look like. With the Scrum Master and Product Owner in place I don't really see the need for a manager any more. What are your thoughts in that context. You mentioned that the managers responsibility is going to shift. Thats what the SM and PO are in my opinion - former managers. I envision a CEO surrounded by Scrum teams which on its own could be Scrum of Scrums. This wouldn't scale to a company larger then ~200 person. So, if the company becomes to large you create a spin-off à la W.L. Gore and Associates.
Also, I would love to see the 13 questions.
Re: Your answers in the context of Scrum
Hard to say, based on the short description given.
Is the person with the title of ScrumMaster still doing budgets, dealing with HR on hiring and other personnel issues, and so forth? That stuff usually doesn't just go away. Changing a person's title may change the what and how of responsibilities--or not.
In my experience, it's sort of confusing (to the team and the manager) when there's a title change and the person who used to be manager is now ScrumMaster. Sometimes it makes it harder for the team to take up responsibility for managing their own work because they still look to their erst while manager. Sometimes it leads to friction because neither the manager nor the team are clear on where the new decision boundaries lie.
You can find the 13 original "essential questions" on my blog: www.estherderby.com/2010/02/13-essential-questi...
The questions have evolved a bit, and the set I'm referring to in the interview have a slightly different focus. I'll be writing about those on my blog, too.
Thanks for your kind words about the interview.
Re: Your answers in the context of Scrum
thanks for your answer. I guess we think alike. But, let me describe my thought process more in detail.
I see the following job and title changes when Scrum is introduced:
1. PM -> SM
2. PM -> PO
I assume that the PM used to be in charge of the people who will form the team.
As for 1. I see potential conflicts. You described those as well. In short the SM will still try to run the team. Depending on his/her old style more or less command and control. Therefore this is IMHO not the best option. I would rather see a team member to step up to become SM.
As for 2. That is the transition which makes the most sense for me. The PO still owns the responsibility for the product and has high visibility. All aspects which are important to most former PM.
Also, I see the team to handle and manage aspects like hiring etc. on their own. There could be a backlog item to hire a new team member. Sure there would have to be open position to be created. For that there could be a company wide staff department which handles bureaucracy stuff.
The budget and all things financial in regards to the product under development belong to the PO.
What I am trying to say is that most responsibilities of a classical PM will be handled either by the PO, SM or team. The remainder by the staff department.
The thought wrapped up: Do we still need managers?
I see the CEO in the center surrounded by Scrum Teams. Each of those teams could be a Scrum of Scrums. The CEO is to be considered the Over-PO in a company wide Scrum of Scrums. This wouldn't scale indefinitely (200?)
There would be no need any more to have hierarchies within companies. All Scrum Teams self organize and communicate with other Scrum Teams as needed and no managers at all.