Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Articles Three Keys to a Successful “Pre-Mortem”

Three Keys to a Successful “Pre-Mortem”

Leia em Português

Key Takeaways

  • It is to your advantage to acknowledge and prioritize the risks a project will face before actually facing them.
  • A pre-mortem is a useful structure to help identify and prioritize project risks
  • Including a diversity of voices covers the most ground in surfacing risks as well as problem solving
  • Problem solving is a necessary component for a pre-mortem in order for people to feel safe to explore the real risks
  • The learnings from a premortem are immediately applicable to influence the planning and execution of the initiative

When improving processes, there’s no doubt that post-mortems are helpful, but if they were perfect, there would never be any hiccups. Instead of exclusively looking back on what happened, why don’t we try to steer what will happen in the first place, especially for high-stakes projects with potentially costly issues?

The concept of a pre-mortem has been around for years, but only recently have we seen it pick up speed in the engineering community (though we have been doing this for some years now at Truss). This is an activity which is run before starting on a big stage in a project, but after doing a product mapping and prioritization activity. Rather than exploring what went wrong after the fact and what to do differently in the future, the goal of a premortem is to identify potential pitfalls and then apply preventative measures. It’s a great idea, but for those new to the concept, it’s easy to overlook some important aspects of the process.

To talk about what might go wrong is scary. It acknowledges many things are out of our control, and that we might mess up the things which are within our control. To talk about what might go wrong, and how to adapt to it, acknowledges the possibility of failure. As this is a rare thing in industry, if done initially outside of a structured activity, this can seem like trying to weasel your way out of work. But if premortems are held with some regularity, and always with a creative problem solving time at the end, it can build a safe space for adaptation in the face of adversity.

In our work with various organizations and missions--from digital overhauls for government services to disaster technology--these are the most commonly overlooked “must dos” for an effective pre-mortem:

Ask the right questions, the right way

Pose questions carefully, with a clearly outlined prompt. Conceptual questions are easier to process when given a specific context. For example, I like: “You are giving a presentation to the board of this company one year from now about why this project failed. What is in your presentation?” Being specific in this way prevents catastrophic or silly thinking such as “an 8-point earthquake hit San Francisco and we couldn’t finish the project” or “the military ran out of funding.” The specificity also helps your team shift their mindset to leaving no stone unturned at an early point in the project, when changes can still be made.

When I started managing a project to improve how military families relocate, one of the repeated themes in our first pre-mortem was the failure to sufficiently engage with the moving industry. To this day, we ask ourselves about new ways to engage, and push to engage sooner.

Make input fully inclusive

Too often, leadership seeks input exclusively from group leaders--even those not directly involved in the execution--and leave out the rest. Getting input from everyone involved--no matter how long they’ve been with the company--is crucial. These people were hired because the company felt they could do the job, and just because someone is more junior doesn’t mean they can’t bring a fresh perspective to the table. Think of this part of the pre-mortem as a book writing project: if we all thought we were going to write a book, and only the most senior person gets to write it, is that an accurate representation of the organization as a whole? Including more voices won’t just benefit the project, it will also create an environment in which people feel they will be heard, and can feel comfortable speaking up.

For example, in a previous job, I helped create a reference map and user interface for a gender-based-violence call center in Port-au-Prince. The user experience expert in the room guided the women who worked there to create their own interface, rather than making it for them. They had unique insights based on their experience as well as being the everyday users.

Group the concerns

After you’ve gotten each concern listed individually on sticky notes or note cards, group them. This will show which topics have “heat” around them -- many people thinking about the same concern, or many manifestations of the same sort of concern. It will also highlight outliers.

At a premortem for our military family moving project, many people expressed concern about the bureaucracy necessary to get something called Authority to Operate (ATO). But only one person worried about the complexities of overlapping the new system with the old system in allotting moves fairly. Does that mean that the individual sees something that others don’t, or that they haven’t been kept in the loop about that topic?

Weigh the risks with their impacts

Here’s a major regret I’ve heard about having prioritized a risk at many a post-mortem: “there was hardly any chance of that problem arising!” When you assess risks to a project, it’s not just about the severity of impact. The likelihood of that risk is equally important to consider. Think about insurance, for an example. Some people buy flood insurance for their houses. The chances that their house will flood depends on geography, location of their home, how their house is built, nearby infrastructure, etc. So a homeowner in a high-risk flooding area with excellent drainage might forgo insurance, whereas someone in a lower risk area with poor drainage might choose to invest. And someone in a very low risk area with a very expensive book collection in their basement might  

At a workshop around security for refugees and other vulnerable populations, I ran a risk analysis of the data flow on humanitarian aid workers. While something like wholesale shut down or takeover of the network would be devastating, the real concerns are in identity theft, extortion, and basic malware. Because of how unlikely wholesale takeover is, it wasn’t worth investing resources beyond basic security hygiene. The network admins should instead spend their limited time focusing on prevention, response, and recovery from identity theft, extortion, and basic malware. So, while some unlikely issues are still worth preparing for, effort to remedy them can be scaled depending on the degree of an event’s probability.

Solve theoretical problems playfully

Now that concrete concerns are expressed and prioritized, it makes sense to solve some problems. Because actions haven’t been taken yet (this is a pre mortem, after all), it’s possible to brainstorm wild ideas of how to stop the problems before they start. Can it be avoided entirely? Can a new group be involved in a unique way? While many of the ideas won’t be feasible, the creative juices will surface approaches that are.

At a recent pre-mortem, to creatively problem solve around some particularly onerous paperwork, a wild idea was entirely reorganizing branches of the military to increase understanding of modern devops methodologies. While we can’t do this, it did get us thinking about new stakeholders and ways we might lighten the load.

Those Who Asked “What If?”

The MIT Sloan Management Review published--“ahead of its time” in 2009--the story of an unlikely escapee of the Enron fall out: a credit union. The financial organization had run through scenarios that would be cause for collapse, and then added the necessary safeguards to their processes. In this case, the credit union had made sure to minimize their dependence on one corporation. That is a pre-mortem success story.

It may seem like another to-do on the list, but think about businesses that could have been saved from headline-making disasters if they had drawn up a comprehensive list of threats, and rolled with the punches? Perhaps Toys ‘R Us would have fared better had they considered the impact of e-retail when they chose to invest in their brick-and-mortar stores. Ultimately, introducing the pre-mortem process can determine how competitive a company remains in a fast-changing world.

In addition to avoiding massive issues, conducting a reasonable number of pre-mortems provides more data to help inform future projects--what problems are projects more likely to encounter in government, or healthcare, or finance? How much budget did preventative tactics take vs. the cost of not implementing them? This data can also help teams justify their approach on an issue to clients, making it easier to get buy in from all decision makers.

Introducing the pre-mortem to teams, and onboarding it as a regular occurrence, is likely to be more challenging with larger, established corporations as opposed to more agile startups that don’t have a library full of binders filled with branded guidelines (yet).

As teams improve on their own pre-mortem processes, businesses are seeing results.  Airbnb has done this on a broader level, asking employees to shift their mindset to think about the ways the company could go down. In the years since that’s been a practice, their business has seen great success. So, that leaves the question: what can take you down?

About the Author

Willow Brugh is Delivery Manager at Truss. They mastered the pre-mortem during their time working with disaster response technology. Willow has worked with digital tools to enable coordination between response agencies and emergent response groups in areas affected by fast and slow crises. This often means ad hoc project management with hundreds of volunteers across multiple timezones and cultures.


Rate this Article