What If I'm Wrong

Write the failure modes first, then the first 24 hours

My judgment is: [judgment], and I've already committed [money / time / promises] to it. Write the “if I'm wrong” plan: 1) Three most likely ways it's wrong, ranked by probability × loss; 2) For each, the earliest observable signal and how soon it appears; 3) What to do at 24 hours, one week, one month after detection; 4) What to prepare now — fallback, budget, what to tell people; 5) Which step is unrecoverable past a certain date, and what that date is. Don't reassure me, and don't say it probably won't happen.

How to use

Copy the “first 24 hours” section into your notes. When it goes wrong you won't be in the mood to think it through again.

Good for

The decision is made and the money is spent; what you need is an action list for failure, not another assessment of whether it was right.
You're presenting a high-risk judgment to a boss or investor and have to answer “what if you're wrong.”

Not for

Nothing is committed and you can stop at any time — do a pre-mortem or the shortest path to disproof; a contingency plan this early is fiction.
The downside is fully reversible and affordable — spend the time testing instead of writing a detailed plan.

Example

Input

Judgment: moving 60% of the paid budget from feed ads to search will lift ROI from 1.4 to 1.9 within three months. 40% has moved; ¥220k spent.

What you get

The likeliest error is underestimating search's marginal cost: there are only so many queries, and doubling budget pushes CPC from ¥3.2 past ¥5, leaving ROI near 1.5. You'd see that in week three — in CPC, not ROI; waiting for ROI means two months gone. Within 24 hours, stage the remaining 20% weekly, cap CPC at ¥4.5, and have a “move it back” schedule ready — if it hasn't recovered by week six, execute the rollback rather than waiting another month for data.

Common mistakes

Writing the plan as a disclaimer: plenty of “this may fail” language and not one trigger you could act on.
Covering only total failure — real errors are usually a two-month delay, not a wipeout, and the plan says nothing about that.

Try tweaking it like this

Swap “if I'm wrong” for “if I'm right but slow,” to handle the right direction with the wrong tempo.
Add “name an observer for each signal,” moving the plan off paper and onto someone's desk.