Why you should perform a premortem on your research

Like all researchers, I’ve had my fair share of failed projects: papers that remained unfinished or never got through the revision process; collaborations that didn’t come to fruition; and research projects that never became the successes I anticipated. Some of this was necessary failure — such as the trial-and-error learning or hypothesis rejection needed to obtain new knowledge in research — the kind that is often fetishized in both Silicon Valley start-up culture and academia. Failure, the conventional wisdom goes, makes you better.
But I try to circumvent unnecessary failures — those made through insufficient preparation or execution. Why not try to avoid the avoidable barriers to success?
Over the years, I’ve grown fond of a simple tool that can help to prevent unnecessary failures in my research. This tool, which draws on the work of the cognitive psychologist Gary Klein, is called a project premortem. A premortem takes place before the launch of a project. To carry out the premortem, I imagine that a few years have passed since the project’s launch, and it has failed miserably. What were the underlying causes? As research by Klein and his colleagues shows, this simple method is extremely effective in revealing hidden problems.
The method makes it safe for knowledgeable objectors to voice their reservations about a project. Theoretically, such reasoning is based on prospective hindsight: imagining that an event has already occurred. The true value, however, comes from thinking through how to proactively reduce potential problems and thereby prevent a project from failing.
For instance, some years ago I dedicated weeks to learning the nuts and bolts of a statistical software tool, because I felt some pressure from collaborators to use it. I was initially reluctant to master it, fearing that this would take away much valuable research time. I have not done any research with the tool since that time, vindicating my initial hesitance. In hindsight, it’s obvious that a premortem could have been useful and prevented me from wasting my time.
I now feel I’ve learnt my lesson, and often create premortems for my own research efforts. Before submitting an article to a premier journal, I ask myself: if I presume this article will be rejected, what will be the underlying reasons for that outcome? Sometimes this makes me aware that my data are not strong enough or that the theoretical logic is incomplete or inconsistent. I also increasingly do premortems with my collaborators, and when designing new courses for students.
However, awareness of the issues that premortems raise is just the first step — there is also a need to identify remedies for these problems. In my own work, I’ve forced myself to spend extra time trying to make a certain theoretical logic more consistent with existing debates in the field, and intuitively understandable, before reviewers find my work lacking. In some cases, it might be difficult to address the issues, but at least you can think about how best to approach the concerns with a revision, thereby getting a head start.
A premortem often becomes even more effective when you include collaborators. In my experience, they help to overcome any biases or blind spots you might have and improve the overall quality of the input to the premortem. A group can also generate more ideas on how to proactively address different issues.
Recently, I asked a close collaborator if we should pull the plug on a project in which we’ve invested a lot of resources. In the end, we reasoned that the timing of the project probably wasn’t right (and wouldn’t improve in the near future), so it would be best to make the tough decision to stop the project altogether.

How researchers and their managers can build an actionable career-development plan

However, a group setting can also be a potential minefield. Failure could result from a collaborator acting as a bottleneck or not communicating sufficiently with the rest of the group. Such actions have the potential to disrupt group harmony or cause conflict.
Although showing sensitivity remains crucial, the premortem setting might provide a ‘safe space’ for people to air potential concerns, because it is framed as an imaginative thought-experiment with the premise that all potential causes of failure should be broached. Dealing with interpersonal issues doesn’t get any easier once a collaboration has failed, so it’s often better to have difficult conversations early on.
How else can you use the premortem in your research? In my experience, a premortem works best during the preparation stages of papers, projects and partnerships.
I’ve developed a quick checklist for conducting your premortem. The process has five simple steps, each with a related key question. Print out the checklist and keep it in your office or take it to your next meeting. Once you’ve conducted an individual step, tick it off.
The bottom line is that academic life might be full of failures, but we should still try to avoid the avoidable ones. The premortem method provides a simple approach for doing just that. And in case you were wondering: yes, this article survived my initial premortem.