A Pre-Mortem, coined by cognitive psychologist Gary Klein, flips planning on its head. Before you commit resources to a launch, you imagine it is six months later and the project has already failed. Then you work backward and ask everyone to explain what went wrong. By treating failure as a fact instead of a possibility, you give your brain permission to say things it would otherwise suppress out of optimism or politeness.
It works because confidence is exactly when blind spots hide. When a team is excited, nobody wants to be the person raising objections. The Pre-Mortem reframes that dissent as the assignment, so the doubts come out before they cost you anything.
Running one is simple. Gather the people doing the work, state the plan, then say: "It is launch day plus three months and this was a disaster - why?" Have each person write reasons silently, share them, and turn the sharpest ones into owned action items. You might discover nobody owns the migration path for existing customers - so you fix it now, while it is still a five-minute conversation instead of a fire.
Common mistake
Treating it as a generic risk-brainstorm ("what could go wrong?") instead of assuming failure has already happened - the past-tense framing is what unlocks the candor, and skipping it produces polite, useless answers.
How to use it
Three ways to put it to work.
Everyday life
Booking a big family road trip this summer
Before reserving anything, picture the last day of the trip and it was miserable. Working backward, you name the culprits: the toddler melted down on 8-hour drives, the cabin had no AC in a heat wave, and you blew the budget on roadside meals. So you cap drive time at 4 hours, confirm AC in writing before paying, and pack a cooler now.
Tech
Shipping a payments rewrite next sprint
Imagine it is two weeks post-launch and the rewrite failed badly. Working backward, the team lists the causes: a webhook race condition double-charged customers, the rollback plan was untested, and on-call had no runbook. So you add an idempotency key to every charge, run a rollback drill in staging before release, and write the runbook this week.
Small business
Opening a second cafe location across town
Pretend it is six months in and the second cafe is bleeding cash. Tracing backward surfaces the killers: your best barista slipped covering two sites, noon foot traffic was half what it looked like, and rushes overlapped. So before signing the lease, count actual passers-by at 8am for a week, hire a second lead now, and stagger hours to share staff.
Reckon · Thinking tool
Pre-Mortem
Gary Klein, cognitive psychologist
What it is
Before launching, assume the project has already failed. Work backward to identify what went wrong.
When to use it
Before committing resources to a plan you feel confident about. Confidence is exactly when blind spots hide. Run the pre-mortem to surface them.
Work through it.
Print this (use the Save as PDF button up top) and fill it in on paper. Surface every way your plan can fail before you commit resources, so blind spots become action items instead of fires.
01
THE PLAN: In two or three sentences, describe the project or launch you are about to commit to.
02
THE FAILURE ASSUMPTION: It is one year from today. This project has failed badly. Set confidence and optimism aside and state that as fact.
03
THE CAUSES: Working backward from that failure, list every plausible reason it went wrong. Write fast and be specific. Aim for eight or more.
Include execution failures, market failures, team failures, and assumptions that turned out wrong.
04
THE UNDEFENDED RISKS: Circle the items on your list that have no owner and no mitigation in your current plan.
05
THE FIXES: For each undefended risk, assign one person and one action to reduce it before launch. If it cannot be mitigated, decide whether to proceed or stop.
Reckon · reckon.report · flip for the working template
Reckon · Working template
Pre-Mortem
Print double-sided. Read the how-to on the front; do the work here.
It is six months from now. The project failed.state it as fact