The mistake most early-stage teams make: they write the ICP, frame it on the wall, and never touch it again. Then six months later they wonder why the funnel is full but conversion is flat.
- Written once
- Framed on the wall
- Never touched again
- Funnel full, conversion flat
- Rewritten every quarter
- Checked against closed-won every month
- One assumption tested every week
- Loop between belief and proof stays short
ICP isn't a deliverable. It's a practice. Here's the cadence that actually moves the funnel.
- Every quarterRewrite it from scratchDon't edit. Start from a blank page.
- Every monthAudit the closed-won listLast 30 days, one sentence per deal.
- Every weekKill one assumptionTry to disprove one belief.
Once a quarter: rewrite it from scratch
Don't edit. Rewrite. The act of starting from a blank page forces you to confront what you've actually learned, which jobs-to-be-done you're now winning, which segments turned out to be tourists, which titles convert and which just take meetings.
If your new ICP looks identical to the old one, one of two things is true: you weren't paying attention, or you're not selling anything new.
Once a month: audit the closed-won list
Pull the last 30 days of closed-won. For each one, in one sentence each:
- What were they trying to do?
- What were they doing instead?
- Why did they pick us, in their words?
- What were they trying to do?
- What were they doing instead?
- Why did they pick us, in their words?
If you can't answer those for the last cohort, you're guessing, and your sales motion is downstream of guesswork.
Once a week: kill one assumption
Pick one belief you have about the ICP and try to disprove it this week. Talk to a customer, look at the data, run a small test. Most ICPs rot from the inside, held together by unexamined assumptions.
The discipline is the same as any operator practice: shorten the loop between what you believe and what you can prove.
