It Already Failed

Surface the risks that will sink your plan — before you commit.

It Already Failed

What this fixes

You’re about to commit to a plan - a sprint, a release, a migration, a launch - and asking “what could go wrong?” never surfaces the real risks. Once a team has built a plan, it stops evaluating it and starts defending it, so the doubts people privately hold stay unsaid. Then the post-mortem opens with “I had a bad feeling about that.”

It Already Failed flips the question. You assume the plan has already failed and explain why - which turns defending back into investigating and surfaces the concerns already quietly living in the room while they’re still cheap to prevent.

How this works

Imagine: tomorrow, your team is committing to a major customer migration. Before anyone signs off, imagine it’s six weeks later, and the migration failed badly - and explain why.

Imagine the project has already failed. Now explain why.

That single shift - from predicting to explaining - is the whole tool. Instead of guessing what might go wrong, the team writes the causes of a failure that has “already happened”: what did somebody already suspect, but nobody acted on? Every cause goes up on a wall, one card each, until the room can see just how many ways this could fail.

Then you thin the wall. Keep only causes you could actually have acted on before starting, narrow it down to the top five that matter most, and for each write one preventive move with an owner. The wall is the point: seeing fifty ways to fail is what makes “we can only protect against five” land.

Team: one person acts as the facilitator, running the activity from the Facilitator Guide. Everyone writes causes; the room builds the wall together.

Time required

Team ✓ · About 20–30 minutes, with a facilitator.

The time required is indicative and will depend on the size of the plan and the team.

The facilitator should read the short facilitation guide first - about 8–10 minutes. It is a downloadable PDF on this page.

When to use

Bring one real plan you’re about to commit to - the tool works at the point of commitment, not on a hypothetical.

Use it whenever:

  • you’re about to commit to a plan with real stakes - sprint, release, migration, launch
  • “what could go wrong?” keeps getting polite, shallow answers
  • people seem to be defending the plan rather than testing it
  • you sense there are unspoken doubts in the room but no one’s naming them

What you get

  • A wall of named failure causes - the risks the team can actually see
  • The top five that matter most, chosen by forced elimination
  • For each: one preventive move, with an owner and a time

You’ve done it well when: the team commits with its top failure modes named and prevented - and leaves with owned preventive moves, not just a list of worries.

What you’ll notice

  • The doubts people were holding quietly are suddenly on the wall
  • Naming a risk stings for a second - then it’s a relief, because now it’s preventable
  • The room stops defending the plan and starts investigating it

Why this works

Once people invest in a plan, they become its advocates, not its evaluators - so a forward-tense “what could go wrong?” feels like an attack on something they now own, and the risks they suspect stay silent. The past tense removes the need to defend the plan: you’re examining a fictional failure, not criticising a colleague’s plan, so people investigate freely. And the real reason it works: we are far better at explaining failure than predicting it. Forward tense demands prediction, which humans do poorly; past tense demands explanation, which we do well. (Based on the premortem technique.)

What you need

One real plan. A plan you’re about to commit to, with genuine stakes - a sprint, release, migration or launch.

The cause cards and the board. One blank card or sticky note per failure cause, and a whiteboard, wall or large flip chart to build them into a visible wall, then thin to the top five. Supplied with the activity.

Markers. One per person, so everyone writes causes at once - the wall should fill fast.

A facilitator. They hold the obituary frame, keep the wall filling before any sorting, run the actionability filter and the forced cut to five, and make sure every survivor leaves with a preventive move and an owner.

Steps

  1. Set the frame (2 min). Team: appoint a facilitator first. Then create the fictional world: “It’s six months on. This plan failed. We’re not asking if - we’re explaining why.” Write the failed plan at the top of the board.
  2. Build the failure wall (8–10 min). Everyone writes causes of the failure, one per card - especially the things someone already suspected but nobody acted on. Work silently while everyone writes; discussion comes later. Put every card on the wall as soon as it’s written. Don’t sort, don’t debate - just fill the wall until the room sees how many ways this could fail.
  3. Apply the actionability filter (4–5 min). For each cause, ask one question: could we reasonably have acted on this before we started? If yes, it stays. If it’s vague blame, or unactionable, it comes off the wall.
  4. Force it to the top five (4–5 min). Twenty equal risks is no output. If two cards describe the same failure, merge them first. Then converge - argue, group, cut - down to the five failure modes that matter most. The cutting is what makes them a priority.
  5. Assign a preventive move and owner (4–5 min). For each of the five, write the prevention, not the problem - a real action with an owner and a time. “Build an API spike before Sprint 2”, not “API instability”. Prevention is the output.

Output / Takeaways

You leave with:

  • The top five failure causes, named and ranked
  • One preventive move per cause - a real action, owned and timed
  • A plan you can commit to with its worst risks already pre-empted

Important rules

  • Run it in the past tense - “it failed, explain why”, never “let’s brainstorm risks”
  • Build the wall before you sort - let the room’s own concerns emerge; don’t pre-sort into categories
  • Force it to five - the actionability filter keeps only what you could have acted on, then the cut makes them matter
  • Only then assign prevention - a risk with no preventive move is just a worry

Planning to facilitate this activity? Access the resources below to get started.

More Tools
MIndn
Execution must be smooth. Not chaotic or stressful.
Built by Witmer.