top of page

Fail It First: How the CHANGE Framework Turns a Pre-Mortem Into Your Cheapest Insurance

Nov 13
6 min read

At kickoff, everyone nods. The timeline looks reasonable. The risks seem manageable. Six weeks later, in the post-mortem, everyone is suddenly brilliant: “We should’ve seen that coming.”

Hindsight is free. Rehearsing it in advance is the cheat code. You can be the one person in the room who already knows where this thing breaks, and that advantage costs you twenty minutes.

A pre-mortem asks you to imagine that a project has already failed, then work backward to identify why. The point isn’t to predict the future perfectly. It’s to spot the problems you can still fix while fixing them is relatively cheap.

Why your brain won’t do this on its own

When you’re planning something, it’s easy to picture the steps going smoothly: the people respond, the work takes the time you expect, and the launch goes as planned. You focus on your plan’s story, not on what similar work has actually been like.

That’s the planning fallacy. In their 1994 paper, “Exploring the ‘planning fallacy’: Why people underestimate their task completion times,” Buehler, Griffin, and Ross found that people underestimated how long their own tasks would take and often built a rosy, plan-based scenario instead of using their past track record. The bias shrank when participants were explicitly told to fold their past experience into their estimate.

There are important limits: the study used lab and student tasks. The planning fallacy is a tendency, not a law, and the study says nothing about whether you can eliminate it. Still, it points to a useful question: what happened the last time you did something like this?

Read the study or its PDF.

Why imagining failure can help

Psychologists Deborah Mitchell, Jay Russo, and Nancy Pennington studied what they called prospective hindsight: imagining an event as certain and already in the past, then explaining how it happened. In their 1989 paper, that framing produced roughly 30% more explanations than imagining the same event as merely possible.

Be precise about what that means: the 30% is about the number of reasons generated, not their accuracy or the quality of the predictions. Finding more reasons isn’t the same as finding the right ones. A pre-mortem gives you more possibilities to inspect, not a guarantee that you’ve found the crucial risk.

Gary Klein popularized the team version as the “premortem” in his 2007 Harvard Business Review article. One benefit of the exercise is that it gives people permission to voice doubts that politeness and hierarchy might normally silence.

A checklist with a project hazard highlighted in yellow

Run a 20-minute pre-mortem

Try this alone at your desk or with the people doing the work. For a team, have everyone write their reasons privately first. That gives people a chance to think before the loudest voice sets the tone.

1. Write the decision and finish line in one sentence. Include a date. For example: “We’ll launch the revised onboarding process for the whole team by March 15.”

2. Declare it failed, in the past tense. Be specific: “It’s April 15. The new process launched, but the team has stopped using it and new hires are still asking for the old instructions.”

3. Set a timer and list 10 possible reasons. Don’t filter or debate yet. Write concrete causes: “The checklist was stored in a folder no one checks,” not “communication was poor.”

4. Sort each reason into one of three groups.

  • Fix now: change something in the plan before launch.

  • Watch for: set a signal that would tell you the risk is emerging.

  • Accept: recognize a risk you can’t or don’t plan to prevent.

5. Turn the top two risks into action. Choose one concrete plan change and one early-warning signal. Name who will make the change, and when you’ll check the signal.

The timer keeps the exercise moving. The action step keeps it from becoming a longer worry list.

Add the outside view

Your imagined reasons are useful, but they come from your own imagination. That means they can miss problems you haven’t encountered before. Add a reality check with reference class forecasting, associated with Bent Flyvbjerg and built on Kahneman and Tversky’s outside view: ask what actually happened to the last handful of projects like yours, then use that range as a baseline instead of relying only on your plan.

That might mean looking at previous launches, similar projects at your organization, or comparable work done by others. Flyvbjerg’s work on reference class forecasting offers a way to put your project in the context of comparable outcomes.

This needs real comparison data. If you don’t have any, ask two or three people who have done similar work. Don’t invent a base rate. A rough but honest account of actual experience is more useful than a precise-looking guess.

A person carefully inspecting a bridge before it opens

Two quick examples

A freelancer launching a paid workshop. Imagine it’s a month after launch. A few people signed up, but the workshop didn’t cover its costs and the freelancer spent days preparing material that attendees didn’t use. Possible reasons: the sales page described the topic too broadly; interested people couldn’t tell what they’d leave with; the session time conflicted with client work; or the freelancer built the full slide deck before testing demand.

A plan change could be to share a clear outline and invite questions before completing all the materials. An early-warning signal could be people asking what the workshop actually includes. The freelancer can then clarify the offer before launch.

A manager rolling out a new team process. Imagine it’s six weeks after launch. People have quietly returned to the old handoff method. The new process requires updates in a tool the team doesn’t open during its busiest hours, and no one owns keeping the instructions current.

A plan change could be to walk through the process with the people who’ll use it and adjust the handoff to fit their workday. An early-warning signal could be a handoff that comes in without the new process’s required information. The manager can check the first few uses and fix confusing steps before they become routine.

These are examples, not predictions. The value is in turning vague worries into details you can test against your actual situation.

Three traps to avoid

Don’t use the exercise as a machine for reasons not to try. A pre-mortem isn’t a verdict. It’s a way to improve the plan. Name one thing you’re still going to do anyway and why. For example: “We’ll still run the pilot because we need direct feedback, but we’ll test the instructions with two users first.”

Don’t wait until nothing can change. Run it while the plan is still editable, before you spend the money, time, or goodwill you could have protected.

Don’t turn it into a performance. If nobody will name a problem with the boss’s idea, you’re not getting the full benefit. Give people quiet writing time, make room for candid input, and treat doubts as useful information rather than disloyalty.

A clean blueprint with a yellow-marked point for inspection

Where it fits in the CHANGE framework

A pre-mortem connects naturally to the six areas of CHANGE:

  • Communication: Say the doubt out loud before it turns into resentment or a last-minute surprise.

  • Habits: Make the 20-minute pre-mortem a regular step before a meaningful launch.

  • Attitude: Treat rehearsing failure as care for the plan, not pessimism about the people behind it.

  • Network: Invite the person who’ll tell you the truth, especially someone with relevant experience.

  • Goals: Define the finish line before you can imagine how the project misses it.

  • Education: Each completed project, including a failed one, gives you evidence to use in your next pre-mortem.

Brad Young’s CHANGE is built to be used on live decisions, not admired from a shelf. For the business side, explore Business Decision Making.

Small, cheap, easily reversed decisions usually don’t need a 20-minute ritual. Save the exercise for a decision where discovering a preventable problem early could make a meaningful difference.

Try it in the next seven days

Pick one thing you’re about to start. Set aside 20 minutes, imagine it has failed, and list the reasons while you still have time to respond.

Then change one line of the plan before you spend a dollar or a week on it. That’s it. No courage required, no crystal ball, just the willingness to look at the wreckage while it’s still cheap enough to move.

The people who look calm on launch day usually aren’t luckier than you. They ran the failure first, in their head, on a Tuesday afternoon, and fixed the parts that were fixable. Do that once this week and you’ll wonder why you ever launched anything without it.

 
 
 

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page