Kill the Postmortem
In a recent blog on stupid decisions, a reader asked about lessons learned processes. I had to defer the question since my reply would have been as long as the blog he was commenting on. So here we go: the entire class of retrospectives, postmortems, and lessons learned are a waste of time. Well, to be fair, I have never seen them work. They may have worked for others. Maybe the reason I never see them work is that I am involved only on disasters, you know, those projects everyone talks about for years to come, the ones people cannot get way from fast enough. Surely, the type of work I perform taints my experience.
Why They are Ineffective?
The primary reason the postmortem's fail is lack of executive commitment. Without the organization's management being behind the spirit of the retrospective and implementing the suggested changes, they are a waste of time. The event becomes drudgery. Attendees thoughtlessly answer questions while texting their friends or thinking about other tasks they have to perform.
In one retrospective, management told me not to use the project's audit report as a source of data and the facilitator restricted my comments to one item. The team quickly realized there was a gag order on the person that had fixed the project and concluded the meeting was to check off the box that the task was complete rather than to identify problems to resolve.
The next most common reason is lack of funding. Either there was no funding at the beginning of the project or to complete it budgets were trimmed to exclude anything that would not directly generate the deliverable. The concept of a retrospective was lost long ago.
Of course, there is the apathy component. No wants to prolong the pain by reviewing everything that went wrong. It puts salt in the wound. If this is a problem, try holding the postmortem at a bar. For some reason everyone becomes talkative.
Finally, apprehension will kill any objective involvement. No one wants to go into a meeting where they may be identified as part of the problem. Even though a properly run meeting does not assign blame, in companies that insist on finding fault, people will protect themselves and peers and the postmortem becomes a finger-pointing, interdepartmental blame game.
Are my observations isolated? I am afraid not. Last year, while writing my book, I read The Mythical Man-Month by Fredrick Brookes. The number of colloquialisms originating in the book surprised me—"there are no silver bullets," "How do projects become a year late? One day at a time," and "The Second System Effect." As I read through the development of the IBM 360 architecture, I was bewildered at the number of problems that he describes on a project in the mid-sixties—when I was a pre-teen—that are 100% applicable today. As the saying goes, the first kick by a mule is educational. If we have not learned from these lessons, how are we going to learn from our own problems?
A Different Approach That Works
Looking at these reasons above for why retrospectives fail, all of them have a common thread of management's attitude toward the retrospective. However, rather than try to change the attitude on the lessons learned process, focus your efforts on changing the culture of management as a whole. Management must proactively accept promoting change throughout the organization.
The solution is to make problem identification and resolution part of the recovery. Reflecting back on the article Recovering Projects in Four Easy Steps, you will see that the application of corrective actions is the first part of the Execute phase. This is critical since trying to run the project without fixing the problems that effect it is... well... silly. Anytime you find a problem, fix it; do not wait until the end of the project. Waiting prolongs the pain and probably will result in the problem never being addressed.
For example, in the article How Many Problems, there were three root causes generating nine major failure symptoms. Only one had to be fixed for the project to continue—defining the system's end user. The other two—improving executive management involvement and creating a maintenance group—were solved in parallel with the project. As might be expected, all three of these were intertwined in the failure. The fact that the end user was undefined was a failure of the project; however it also indicated that the PMO was not reading project charters. If they had, they would have seen two diametrically opposed end users.
During the audit, the head of the PMO was asked if the charter conformed to their standards. He indicated it did. When asked if the end user was correct, he reply was that the PMO did not have the expertise to understand whether document content was correct. I concluded they simply checked off the box that the charter had the correct sections. The recommendation to the CIO was that the people in the PMO either add value by reviewing the content or the PMO be disbanded. He implemented the former. Without doing this, the problem would happen on subsequent projects. Another half-dozen changes were also implemented in the CIO's executive committee.
What Are Your Experiences?
Now it is your turn. Have you had a great experience with a retrospective? Tell us about it.
- What was it that made it work?
- What kind of problems were identified and solved?
- What techniques were used to make sure no one felt blamed?