How is ICP scoring different from a boolean title list?
Boolean search returns keyword matches. ICP scoring evaluates fit with written reasons after (or alongside) list building and signal targeting.
ICP Scoring playbook
You teach it who is worth the send.
A scoring contract, not a persona poster. Reviews first. Autopilot earned.
ICP for outbound is a written scoring contract: which people deserve a send, which are a hard no, and what evidence moves a lead from 2 to 4. Score every lead 1 to 5 with reasons. Train with disagreements before you unlock Autopilot. Prospecting and Signals fill the queue; ICP Scoring decides who reaches Voice Writer.
- Observable fields beat workshop adjectives.
- Intent and fit are different scores — do not collapse them.
- Closed-won URLs beat brainstormed personas.
- Write reasons in plain language so you can argue with them.
- Review false 5s and false 1s before polishing 4s.
- Autopilot stays locked until the training loop is real.
The system
Five steps from slide to contract.
If you cannot score a stranger’s profile in under a minute, the contract is not finished.
- 01
Write the five dimensions
Who (titles), where (size and industry), what pain, what stack, and hard nos. Use LinkedIn-visible evidence. Kill phrases like decision-maker and innovative mid-market.
- 02
Seed from closed-won
Collect five positive and five negative profile examples. Wins beat vibes. If you lack wins, use would take a meeting tomorrow accounts and label them provisional.
- 03
Score every new lead 1–5
Five is exact buyer. Four is strong fit. Three is plausible. Two is weak. One is hard no. Require written reasons — not a silent black-box rank.
- 04
Review disagreements every week
Prioritize system 5s you would never buy, system 1–2s that look like wins, and a sample of 3s. Record one-line reasons when you override.
- 05
Raise autonomy only when earned
Keep Reviews with you until disagreement rate is low and the product training gate is met. You can move ICP Scoring back to reviews at any time.
Fit scale
The only scale that matters: 1 to 5.
Same number family as intent scoring. Different question.
| Score | Meaning | Action |
|---|---|---|
| 5 | Exact buyer — title, size, industry, no disqualifier | Enroll / prioritize |
| 4 | Strong fit — minor ambiguity only | Enroll |
| 3 | Plausible — enroll only with strong intent or human OK | Conditional |
| 2 | Weak — missing must-haves | Skip outbound |
| 1 | Hard no — exclusion list | Never enroll |
Intent vs fit
Two questions. Two specialists.
Signals asks: are they in-market now?
Posts, comments, recommendation asks. Hot rant from a non-buyer still fails ICP.
ICP Scoring asks: are they the right buyer?
Title, company, industry, stack, disqualifiers. Perfect ICP with zero urgency is a cold sequence, not a fake saw your post opener.
High intent + low fit
Skip or nurture. Do not burn sends on clever replies to students, consultants, or vendors.
High fit + low intent
Route through Prospecting sequences. Save signal-referenced openers for real signals.
Which specialist
Six specialists. This one owns the step.
Named faces, not a prompt pack. The handoff is the product.
See the full teamPrimary
ICP Scoring
Scores every new lead against your ICP and learns from your reviews.
- ProspectingFills the list from saved searches and enrichment.
- SignalsAdds people who already showed public intent.
- Voice Writer + Follow-upOnly after fit clears the threshold you set.
Outcome: qualified leads — fit at or above threshold — before anyone is contacted.
Questions
The ones operators ask.
When does Autopilot unlock for ICP Scoring?
After it has learned from your reviews. The product keeps Autopilot locked until that training loop is done. You teach it who is worth the send.
What if my fit-pass rate is high but conversations are low?
Fix messaging, sequence, or offer — not the ICP. If fit-pass is tiny, widen carefully from closed-won evidence, not vibes.
What score should enroll in outreach?
Usually 4 and above. Allow 3 only when intent is also strong, or a human has approved the exception.