Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Articles Questioning the Retrospective Prime Directive

Questioning the Retrospective Prime Directive

This item in japanese

Imagine inviting a group of intelligent, thoughtful people over for drinks and having the discussion turn to the Prime Directive, a keystone of the retrospective process. If you’re not familiar with retrospectives, I can recommend Norm Kerth’s book on Project Retrospectives or Esther Derby and Diana Larsen’s book on Agile Retrospectives. Other similar processes are post-project reviews or postmortems (why don’t I like that word?) where teams examine what happened over the duration of an activity. The Prime Directive, as practiced by those who follow the practice outlined in Norm’s book, is a pre-requisite for contributing to the learning that is the heart of a retrospective. Many who are new to retrospectives and others who struggle to come to terms with the Prime Directive often raise questions like the ones in the following discussion. This is a good chance for all of us to learn from the comments of these concerned individuals.

* * *

Philippe was the one who started it all. Thanks, Philippe!


The Prime Directive says:

"Regardless of what we discover, we understand and truly believe that everyone did the best job he or she could, given what was known at the time, his or her skills and abilities, the resources available, and the situation at hand."
Really? I have met subversive, obnoxious, really destructive people during my career as a developer and consultant. In most cases, I could not do much about them. Work around or without them.

Just saying that I "truly believe," is not acceptable to me. I have fired one person (in my all career) for gross incompetence, or more exactly for being competent but totally uncommitted to the project. So I do not “believe.” I look at the facts. The Prime Directive sounds pretty naive to me, in terms of the reality of human nature. I am willing to get started with people with giving them full credit, but they have to prove their worth.


It took me awhile to really understand the purpose of the Prime Directive. I had the same struggle with a lot of the "pieces" of a retrospective. The purpose of a retrospective is learning. It’s not a performance review. The way people are, however, it's not enough just to make these "purpose" statements. You have to enlist things like the Prime Directive and the other "pieces" to reach your goals.

The Prime Directive is not about reality. It's about belief. It's about holding an idea for the sake of the retrospective. It's about enabling the brain to focus elsewhere, just for a short time, in order to maximize learning. It's a "game" -- "let's pretend" -- it's not a fact about the workplace or the people in it.

I know that when I ask people to "sign up" for the Prime Directive, they are thinking exactly what you are thinking and that's OK. They are just pretending for a short time, but it's enough to put those judgments aside so that the team can learn.

It's astounding to discover that you sometimes have to do this -- that you have to "trick" your own brain into taking the path you want to follow. Our brains are constantly subject to tricks like this.

I've had a lot of people come up to me after a retrospective and say pretty much what you have said, but then they add, ‘Now that I have signed up for the Prime Directive and gone through a retrospective, I think maybe the Prime Directive is the truth.’ I never say anything to this. I just give a little nod of agreement.


As someone with some experience, I know that people, even with the best intentions and best abilities, can sabotage the teamwork sometimes. I agree with Philippe that it’s frustrating to work with them or around them. I’d also pretty much prefer working without them. But the point is that in their own mind they are not sabotaging at all! They do everything possible to salvage the ill-managed project. From this point of view the Prime Directive is exactly correct. It even throws a “given… their skills” lifebelt, and team skills are just as important as technical ones.

Since usually these people are either strong-minded mature specialists or informal leaders or both, it is our duty as managers to a) build our team in such a way that we reduce this “internal resistance” and b) “sell” our approach differently to different people based on their level, their personality, our relationship with them. This is how we build cohesion in the team. My belief is that we should never expect or hope that everyone will rally round the leader. We will never have perfect team members. We have to learn how to work with the “people material” at hand building and shaping a cohesive team. It’s not easy even for a mature manager. Sometimes we achieve this cohesion, sometimes we don’t, and you see the difference in the team spirit and in the results. I’d say this is the most difficult but the most important “inward-aimed” task of a manager.

With this in mind, I think that it’s vitally important to approach the retrospective armed with the Prime Directive. It is the only way to create an atmosphere with potential for a successful retrospective, and it also gives us another chance to add to the team cohesiveness.

Where I disagree with Philippe is: “I am willing to get started with people with giving them full credit, but they have to prove their worth...” I think it’s either one or another. Either we give people 10 points at the start and deduct if they do not perform (define “perform” any way you want), or we give them 0 at start and they have to prove their worth. I have seen both approaches in action, and strongly believe in the former one. Moreover, I think that the latter approach may be appropriate for McDonalds (don’t know, never worked there), but it never works for IT. I interpret Philippe’s statement to say that full credit has to be evaluated from time to time, and I totally agree with this.


I understand the purpose. It’s the "truly believe" part that puzzles me. And what you are saying now is that "for the sake of having a healthy and productive retrospective, we will ASSUME or PRETEND (your words) that everybody did the best job they could..."

If I know for a fact that Linda was goofing off, telling us she was sick while having been spotted skiing at Whistler, being confrontational with her peers, being late in delivering her bits, accusing other people for her delays, you do not want me to enter the retrospective truly believing that she did her best, do you? What you are saying is that what we see of Linda is the best she can do, and we will not question during the meeting why or how she did not did not do her best?

I'd rather rewrite this directive: Regardless of what we discover, we assume that everyone did the best job they could, given what they knew at the time, their skills and abilities, the resources available, and the situation at hand.

But why shouldn't Linda’s performance be an outcome of the retrospective? Why should we use nice metaphors and flowery language for 3 hours when the real problem is visible to anyone (and now maybe not to Linda).


It is not the purpose of the retrospective to do a performance evaluation. An organization certainly needs to evaluate performance, somehow, and perhaps, some of the things that come up in a retrospective can help the person who must do a performance evaluation


it’s difficult to proceed when members of the team begin blaming each other or others outside the team. In many cases, it's difficult to say why there was poor performance. I think that when this is part of a retrospective, participants will consciously or unconsciously protect themselves and either not tell the truth or hide anything that might make them look bad.

It's difficult when there have been bad times for a team to focus on learning. I used to think it was an "engineering mentality" to continually find fault, but now I am learning that this is a human quality. Things can quickly deteriorate, especially if the project has "failed," and it takes considerable effort to learn in a setting like that. I think we do that in our own lives.

Words are important in any language. Things are not always as clear-cut as we might imagine. What does it mean to "do the best you can, given..." I think you can say that someone's performance was not up to expectations. I don't think you can say that they were not doing their best. It’s possible, of course, to, in good conscience, fire someone who was doing his best.

Finally, I think you are free to re-write the Prime Directive or any of the pieces of a retrospective in any way you like :-) !

Most of the time people are struggling with problems they never talk about. No one lives very long -- and I have lived a long time so that includes me -- without having some times in their lives when others might have said they weren't doing a good job. That certainly doesn't mean they weren't doing their best. Judging whether people are doing the best they can is perhaps impossible. Doing a performance review is much easier, although I'm glad I don't have to do that :-)!

Linda, continued:

This brings up something I have seen only because I have facilitated many retrospectives, so I realize it’s a rare event, but a retrospective can be a time for change. I have seen individuals change and teams change, sometimes in startling and unexpected ways. That cannot happen unless people sign up for the Prime Directive and take advantage of some of the other pieces in the process. I remember so many retrospectives where things turned around. I would hate to lose that for the sake of holding on to what we think we "know" about other people.

The penalty for not having the Prime Directive is getting stuck. It's so easy to do -- pulling out the same rubber stamp for our interactions -- I see that in myself all the time.


Just to re-iterate some of the points that Linda made. The Prime Directive is a very deliberate request for us to suspend disbelief. Let's pretend, as Linda said, that for the next two hours, or the duration of the retrospective, that we will openly accept that everyone in the room has acted with the best of intentions in a way that was in line with their best efforts and abilities. Of course, this is naive, but let's consider it a mental challenge. Let’s take all of our suspicions and judgments about the others in the room and try to turn them on their heads and try to figure out how someone could behave in that way and yet be in line with their best efforts. It’s challenging since the core of our being believes the opposite.

In the example that Philippe gave of the seeing the employee skiing when she should have been working. Instead of assuming that she was just playing hooky, how can we imagine that she is going skiing and still be contributing her best effort to the job. This encourages us to take a positive perspective on her motivation. Maybe she was skiing with an important client. Maybe her brother who was a ski fanatic has just died from cancer, and she is going skiing to deal with her grief.

Ultimately, the motivations of others are unknowable -- at least until we verify them with the other person directly. Part of the Prime Directive is a recognition that our observations are subjective and influenced by our prejudices. It challenges us to attempt to discard them -- at least for the short time that we are in the retrospective. And if we do so, we just might learn something.


Bingo! Remember Yourdon’s book on structured design published some 20 or 30 years ago? He tells a story about a guy who left the office in the middle of a workday, then came back with a can of paint, painted the door of his office green without saying a word to anyone, and then left again for a day. Apparently he spent 36 hours in the office trying to fix a bug (unsuccessfully).


I found it took me a while to get it. The impression that I got from reading Norm's book is that it was just something that you just read at the start of a retrospective -- like an incantation. I tried it a couple of times but no magic. So I abandoned it. Discussion on the retrospective yahoo group was, for me, the key to understanding. I now try to include the directive in retrospectives -- but it requires more than just reading it out. You really need to have a discussion with the participants about what it means before the lights come on. If you do iteration retros, this is a great “iteration szero” retro topic :) the meta retro retro!


I have recently joined a new company and I have been asked to moderate a retrospective for a project that was not very successful to say the least. It had three announced release dates that were canceled and it was eventually pushed out of the door just because the company could not delay it further. As you can imagine this experience has seriously damaged trust and relations between business and development organizations. My goal for this retrospective is not only to learn what can be done better but rebuild relationships and trust. For this the Prime Directive is the only way to go.


The instances of the Prime Directive that puzzle me most are not those where someone is judging others, but those where someone says,

"I know I sometimes don't do my best because I'm feeling lazy or down or procrastinating or whatever, so I assume there are times when others are also not doing their best for the same despicable reasons.  And I am willing to take the blame for those times myself, so it should be okay to blame them too. We will all benefit and learn from ferreting out those instances and bringing them out in the open. Only then can we have all the data for problem solving. And maybe the humiliation will act as a future deterrent."
There's something about the high expectations and standards to which very bright people hold themselves and their willingness to examine behavior rigorously, that gets in the way of compassion and empathy for self and others. I personally do not believe that humiliation acts as a deterrent for anything. I do not find shame to be a motivator.


In the TQM days, both Deming and Juran observed that when a worker made mistakes, about 80-85% of the time it was due to a problem with the system, not the individual. I have found this to be generally true. It's not that individuals don't get tired or sloppy, because they do. But a system that does not ever allow a worker to get tired or sloppy is a bad system. For example, when you plug wires into a PC to assemble it, they are keyed so you can't make a mistake by putting them in upside down or backward. I recall in the early days of PC's it was very easy to install IDE cables wrong, and I did that several times. I didn't blame myself, but the un-keyed connection.

I likewise believe that most software defects are not the fault of the people doing the coding, but a fault of the system that, for example, doesn’t have testing in place to catch defects immediately. Just as factory workers get tired or sloppy, so do developers. The system should assume that this is the case, and put mechanisms in place to compensate for it. We should not expect every developer to be 100% alert 100% of the time. That’s an unreasonable expectation. We should design systems that allow people to be the people that they are.

Another related observation is that a team may lack the motivation to do a really good job. In my experience, this is often about 80-85% of the time, due to frustration with the system or lack of proper tools/expertise or lack of access to customers to find out what really needs to be done or a measurement or management expectation which gets in the way of doing the right thing (a very common cause of underperformance). My inclination, when people are not doing their best, is to look at the management system for demotivators, lack of appropriate expertise, locally optimizing measurements or expectations, or other blockages that keep people from doing their best. 

I don't assume that everyone is doing their best, but I do assume that if they are not doing their best, the underlying cause is far more likely to be a system or management problem than it is to lie with the individual. Certainly this is not always true. There are scoundrels in the world and underperformance needs to be addressed, no matter what the cause. However, trying to improve individual performance when the problem is systemic is not very useful. Thus, the first place to look to solving problems is not to place blame on individuals, but to dig deeper to find the systemic root of the problem.


I have a way I’m comfortable leading into the Prime Directive. I say something like: “I try to stand in this space as a matter of my personal values. It’s not always easy, and sometimes when I see something that seems really “doh” (slap to forehead) I have to remind myself. Now, I’m not suggesting that you adopt my values, but from a purely pragmatic standpoint, it is much easier to influence someone if you haven’t written them off, and it’s virtually impossible to learn from someone you’ve written off as stupid.” Then I ask people if they can stand in this space, not forever, but for the time we’re in the retrospective.”


Even with a large group. I just hold up my hand or walk in a little circle or in some way indicate that I'm going around the room and looking at each person. It really doesn't take long. This is definitely one of those influence strategies -- making eye contact, getting each person to say "yes" somehow. Now you have much stronger commitment. Usually I don't have to do any "reminding" -- there's always someone in the group -- a retro facilitator-in-the-making :-) who will do that -- and since that person knows the folks in the group s/he can tailor the remark to the person and to the group.


The Prime Directive was developed late in my retrospective facilitator's career. It wasn't until I started to teach how to lead retrospectives that I found it necessary to scribe the Prime Directive. However concepts embodied within the Prime Directive were part of the first retrospective I ever participated in, and remain ingrained within me to this day.

As a teenager, I raced sailboats. During one heavy-weather race a friend of mine capsized and drowned. Top sailors, for whom I had a great deal of respect, led a fearless review of every aspect of the race, of every person’s actions, and of every decision made, for no other reason than to help the entire sailing community learn how to prevent another death. For this group of sailors to fearlessly retell such a painful story, our leaders made it clear we were not there to judge anyone's actions. There would be no scapegoat, nor will anyone be found to be at fault. Guilt would not be part of our ritual, and there would be no punishment as we all felt bad enough.

It was a fine example of a community collectively retelling a story for the purpose of learning. It's been more than three decades since that ritual, but the lessons I learned from that event are with me every time I step on the deck of a boat.

As I developed my course, I realized I couldn't say: "there will be no fault-finding, no judging, etc." because by saying it, I'd bring those concepts into focus. Instead, I needed to find a message that was a complement to these things. I first thought the complement was "we assume that everyone did the best job they could." But the word "assume" seemed too weak, so I changed it to "we understand and truly believe that everyone did the best job they could."

I think what's most important is for us, as facilitators, to understand the concepts behind the prose that is the Prime Directive. With this foundation, we can use our creativity to shift the perceptions of a group if we need to. I think Esther's Elaboration on the Prime Directive needs to become part of our story of who we are and part of our culture.


Remember to apply the Prime Directive to yourself! In my Personal Retrospective Workshop we discuss how our futures are greatly influenced by how we talk about ourselves and our experiences—to others and to ourselves, I use the Prime Directive as one example of context for self-assessment. Wonderful discussion usually follows, with many folks noting that they are more likely to apply generous interpretation to others than to themselves, and ultimately how this affects their work.


This has been a great conversation with all you great people. I found this tidbit on the yahoo group. Thanks to the contributor:

A television interviewer discussing the Challenger space shuttle explosion with Dr. Richard Feynman:

Interviewer: And yet your commission, and we just heard the Chairman Rogers, say, "We're not here to blame anybody."

Why not? Why is somebody not blamed?

Dr. Feynman: I don't know how to assign blame, and whether it does any good. The question is: how do we educate...

The MacNeil/Lehrer Newshour, June 9, 1986


Thanks to the following people who were willing to put their names on the line and add to this discussion in a number of ways: Steve Adolph, Paul Culling, Esther Derby, Geoff Hewson, Norm Kerth, Philippe Kruchten, Diana Larsen, Jaswinder Madhur, Ainsley Nies, Eugene Nizker, Mary Poppendieck, Owen Rogers, and Michael Vax.

About the Author

Linda Rising has a Ph.D. from Arizona State University in the field of object-based design metrics and a background that includes university teaching and industry work in telecommunications, avionics, and strategic weapons systems. An internationally known presenter on topics related to patterns, retrospectives, agile development approaches, and the change process, Linda is the author  Fearless Change: Patterns for Introducing New Ideas, written with Mary Lynn Manns, and editor of Design Patterns in Communications Software, The Pattern Almanac 2000, and The Patterns Handbook.

More on InfoQ: Interview: Linda Rising on Collaboration, Bonobos and the Brain

Rate this Article