ICP rebuild worksheet (30 days)

Working doc for rebuilding your ICP from real closed-won data, not founder intuition or marketing-offsite assumptions. Week-by-week sequence with deliverables. Drop into a Google doc or Notion page and fill it out alongside your CRM data.

Who this is for

• Outbound underperforming on forward rate or reply rate
• AE conversion uneven across segments
• Sales cycle stretching for unclear reasons
• Legacy ICP set in a planning offsite 12+ months ago and never validated

What you need

• Last 18-24 months of closed-won deals from CRM (minimum 30, ideally 100+)
• Last 12 months of lost deals
• Current open pipeline
• Access to AEs who closed the deals (for trigger reconstruction)

Week 1: data preparation

For each closed-won, capture:

Field	Source	Required
Company name	CRM	Yes
Industry	CRM / enrich	Yes
Employee count at purchase	CRM / enrich	Yes
Geography	CRM	Yes
Buyer title	CRM	Yes
Buyer tenure (months in role)	LinkedIn	Yes
Tech stack at purchase	BuiltWith	Optional
Lead source	CRM	Yes
Sales cycle (days)	CRM	Yes
Deal size (ARR)	CRM	Yes
Trigger	Reconstruct from notes/AE	Yes

Trigger reconstruction (the most important step)

The trigger field is usually missing from CRM. Reconstruct it:

1. Pull won-deal notes from CRM
2. Slack or email the AE: "What changed in [Company]'s world that made them buy now?"
3. Categorize each trigger:
   • Hire — new role posted or filled
   • Fund — funded round
   • Lead — leadership change (CRO, CEO, VP)
   • Stack — tech-stack change
   • Reg — regulatory filing or compliance event
   • Comp — competitor product issue / price hike
   • Acq — acquisition
   • Internal — internal pain crystallized
   • Unknown — can't reconstruct (mark, don't guess)

Most teams discover the trigger for 60-80% of deals.

Week 2: pattern analysis

Cluster closed-won by company size, vertical, geography, buyer title, buyer tenure, tech stack signals.

Answer in writing:

1. What size company tends to close fastest?
2. Which industry verticals are over-represented vs total pipeline?
3. What buyer tenure correlates with close rate?
4. What tech stack signals predict fit?
5. Which two triggers appear most often in closed-won?

Most rebuilds surface 2-3 real segments, not the 4-6 the legacy ICP claimed.

Week 3: lookalike disqualification

What almost-fits but doesn't actually close. These are where the team wastes the most outbound spend.

For each lost deal that demographically matches your primary ICP, what was missing?

• Wrong trigger (or no trigger)?
• Wrong tech stack?
• Wrong buyer tenure (long-tenured, no urgency)?
• Different decision-maker than expected?

Write the disqualification list:

Companies in [segment X] that lack [signal Y] are not ICP, regardless of how qualified they look on title and size.

The disqualification list is what makes the rebuild stick.

Week 4: validation and rollout

Test against lost deals

Apply the new ICP scoring to lost deals. The model should predict losses.

• 80%+ of lost deals score out-of-ICP — pass
• Lost deals that score in-ICP have a clear "what was missing" reason — pass
• Win rate on in-ICP segments is ≥2x out-of-ICP — pass

Test against open pipeline

• In-ICP open deals are progressing well
• Out-of-ICP open deals are the ones reps are struggling with

Deliverables (end of day 30)

1. Two-page ICP document with primary + secondary segments, triggers, exclusions
2. Scoring rubric the team can apply to new accounts
3. Disqualification list of lookalike segments to skip
4. Trigger taxonomy of high-yield signals to monitor

These go into:

• The team CLAUDE.md (Section 2: ICP definition) — see the CLAUDE.md template
• The outbound subagent's signal scoring rubric — see the TVA scoring rubric
• The lead scoring rules in your CRM

What surprises teams during the rebuild

1. The aspirational segment is dragging down win rate. Teams chasing enterprise often discover their close rate at enterprise is 3-5% while mid-market is 18-25%.
2. The trigger isn't what marketing thinks it is. Marketing claims content/inbound; closed-won data shows hires, funding, leadership changes.
3. The buyer title is wrong by 1-2 layers. Legacy says "VP Sales"; closed-won shows CRO or Director of Sales Ops.
4. 20-30% of the legacy "ICP" is lookalikes. Filtering them out raises win rate without hurting volume.

Where this fits

This worksheet is the second phase of the 90-day PE portco GTM rebuild. ICP comes after pipeline truth-finding, before outbound subagent buildout. Full deep dive here.