GTMcraft Plays · The Dynamic ICP Play

Your ICP rotted since January; rebuild it from won and lost data, then split it into books your team works.

Play visual

rebuild-icp-signal.html

The 90-second read

Most teams write an ICP once and let it quietly rot for years. Targeting drifts from who actually buys, and win rates fall. A dynamic ICP gets rebuilt each half from your own won and lost signals. You add proactive intent, then split the definition into the books your team works.

• A static ICP is a dead ICP; markets and buyers move faster than annual planning.
• Your last half of won and lost deals already holds the real fit criteria.
• Extract the signals that separate won from lost, not firmographics you assumed.
• Rewrite the ICP from those signals plus intent, then split into books of business.
• Refresh every half, not never; the risk is drift, not refreshing too often.

Read time: about 9 minutes.

The #1 move this week

Pull your last half of won and lost deals into one sheet before Friday. That single sheet is the raw material for every other step. Most teams argue about the ICP from memory and stale decks. The deals hold the truth about who bought and who walked.

Why this move: you cannot rebuild a definition from opinion. You rebuild it from what closed.

Start here: export closed-won and closed-lost from H1, with reason codes and firmographics.

Ignore this week: do not rewrite messaging, do not rebuild scoring, do not touch tiers yet.