About icp-personas
icp-personas is a strategy skill for marketing and sales teams to define target customer profiles, account-list size, and the people to reach within each account. It combines account scoring, market sizing, prioritization, and buying-committee mapping, with ColdIQ endpoints used where data is needed.
Key features
- 01Define ICP across firmographic, technographic, and intent criteria
- 02Score accounts and assign priority tiers
- 03Estimate addressable market size
- 04Select accounts using layered criteria
- 05Map buying-committee roles within accounts
Use cases
- 01Derive ICP criteria from closed-won customers
- 02Estimate target-account counts and priorities
- 03Find distinct buying roles at target accounts
- 04Plan stage-appropriate messaging by persona
Install
Copy the command and run it in a terminal.
npx skills add https:/After installing, type/icp-personasin Claude Code to confirm it loaded.
Details
From the project READMEICP, Accounts & Personas
The strategy layer: decide WHO to target, HOW MANY, and WHICH people inside each account — before you spend credits sourcing and enriching. Mostly methodology; a few steps source data from ColdIQ.
ColdIQ Marketplace Endpoints (where strategy touches data)
| Task | Method | Path | Credits | Endpoint ID | Notes |
|---|---|---|---|---|---|
| Validate filter values (ICP criteria) | POST | /v1/limadata/references/autocomplete | free | limadata.references.autocomplete | Confirm titles/industries exist |
| Size the addressable market | POST | /v1/ai-ark/companies | per result | ai_ark.companies.search | How many fit the ICP |
| Layer intent for prioritization | GET | /v1/signalbase/funding-signals | unknown | signalbase.funding_signals | Bump in-market accounts |
| Map the buying committee | POST | /v1/limadata/prospect/employees | 25 | limadata.prospect.employees | Find personas at an account |
1. Define the ICP (3 layers + scoring)
- Firmographic: industry, employee size, revenue, geography, growth, funding stage.
- Technographic: tech stack, CRM, marketing automation, competitor tools.
- Behavioral/intent: hiring, funding, leadership changes, content engagement, website visits.
Reverse-engineer from your best 10–20 closed-won customers. Score 100 pts (example weights): industry 20 · size 15 · revenue 15 · geography 10 · tech fit 15 · growth 10 · intent 15. Tiers: A 90–100 (1:1 ABM) · B 70–89 (1:few) · C 50–69 (programmatic) · D <50 (exclude).
Validate your criteria translate to real filter values before sizing:
→ POST /v1/limadata/references/autocomplete · free · limadata.references.autocomplete
(Hand the scoring rules to tam-scoring to run on an actual export.)
2. Size & select the account list (ABM)
Revenue reverse-engineering — work backward from the target through funnel benchmarks: Identified→Aware 55% · Aware→Interested 32% · Interested→Considering 18%. (e.g. $1M ARR target ≈ ~3,367 accounts identified.)
Check how many companies actually fit before committing:
→ POST /v1/ai-ark/companies · per result · ai_ark.companies.search
4-layer selection: firmographic fit · technographic indicators · CRM intelligence (closed-lost,
lost-to-competitor, churned) · lookalike modeling from best customers. Then prioritize the
in-market ones:
→ GET /v1/signalbase/funding-signals · ? cr · signalbase.funding_signals (unverified)
Track stage progression: Identified → Aware → Interested (5+ clicks / 10+ engagements) → Considering (site visits, content, demo interest).
3. Map the buying committee
| Role | Function | Budget priority |
|---|---|---|
| Champion | Internal advocate driving evaluation | 40–50% |
| Economic Buyer | Signs the check; cares about ROI | 20–30% |
| End User | Daily user; cares about UX/workflow | 15–20% |
| Technical Evaluator | Integration, security, compliance | 5–10% |
| Blocker/Gatekeeper | Can veto, rarely initiates | monitor |
For each persona capture: title patterns + seniority, function, jobs-to-be-done, pains, success
metrics, content preference, buying role. Find the actual people once you've defined the personas:
→ POST /v1/limadata/prospect/employees · 25 cr · limadata.prospect.employees
Messaging matrix: per persona, different JTBD + pain + stage-appropriate content + role-appropriate CTA (champion → demo/ROI, end user → trial/ease-of-use). Hand off to crawford-method / cold-email-copy to write it.
Where this hands off
ICP scoring rules → tam-scoring · sourcing companies/contacts → apollo-search / coldiq-search-enrich · personas-into-ads → ad-audiences · the full build order → campaign-delivery.