Kill-criteria template

Most founders kill projects too late. Not because the data is unclear. Because the data is clear and the decision is unbearable. By the time the metrics make the call obvious, the team has named the project, told investors about it, written a roadmap around it. The cost of the kill becomes social, not financial.

Kill criteria are the workaround. Decide the conditions for stopping before you start. Write them down. Reference them at every review. The point is to remove the founder discretion from the moment when discretion is most compromised.

Run this worksheet at the start of any initiative bigger than two weeks of work.

The bet

One sentence. The thing you are doing and the thing you expect it to produce.

• Initiative: ______
• Hypothesis: by [date], if we do [action], we expect to see [outcome] because [reason].

If the hypothesis takes more than one sentence, you have not yet decided what you are testing. Sharpen before continuing.

The metric

One number that, by a specific date, will tell you whether the bet worked.

• Primary metric: ______
• Where it lives: ______ (dashboard, sheet, query)
• Read date: ______
• Threshold for "worked": ______
• Threshold for "kill": ______

Two thresholds, not one. The "worked" threshold is the bar for doubling down. The "kill" threshold is the bar for stopping. The space between them is where you keep going at current investment without changing posture.

The cost ceiling

Money, time, and attention have separate ceilings. All three need to be set.

• Money ceiling: $______ across the read period.
• Time ceiling: ___ weeks of founder or senior IC time.
• Attention ceiling: this initiative gets the equivalent of [N hours per week] of leadership focus. If it needs more, it must displace something with a higher kill bar.

A ceiling hit is itself a trigger. Crossing it without a triggered review means the initiative is escaping the rules you set for it.

The pre-commits

Three statements. Sign them at the bottom of the worksheet, in front of a witness if you can.

1. If, on the read date, the metric is below the kill threshold, we will stop within seven days. We will write up what we learned within fourteen.
2. If we hit any cost ceiling before the read date, we will pause and decide whether to extend the budget or kill, with the same kill criteria applying to the extension.
3. We will not move the kill threshold mid-initiative. If we want to change it, we kill the current initiative first and start a new one with the new bar.

After the kill

A killed initiative is data, not failure. The write-up is one page: what we expected, what happened, what we now believe instead, what the next bet should be informed by this. File it under decisions. Reference it the next time someone proposes the same shape of bet.

The discipline is not in making the call when the data is clear. It is in writing down the call before the data exists, and then keeping the promise to yourself when it arrives.

Template from Marga Haus · margahaus.com/resources · Adapt and use freely. Attribution appreciated, not required.