Ideal Customer Profile (ICP)

What It Is

ICP is not a demographic guess — it's a data-driven model built from existing customers and market signals. Build it from your last 20–50 closed-won deals: extract industry, company size, tech stack, growth stage, geography, and buying trigger, then find the patterns that distinguish them from poor-fit accounts.

The Criteria Layers

  • Firmographic — industry, company size (employees and revenue — they don't always correlate), growth stage, geography, buying department, seniority.
  • Technographic — what tools they already run that signal readiness or integration value (BuiltWith, Clearbit, Apollo).
  • Behavioral / intent — hiring signals, funding events, relevant job postings, content-consumption and website-engagement signals.
  • Negative criteria (anti-ICP) — too small to afford it, wrong vertical, locked to a competitor, procurement-restricted. Naming the anti-ICP is as important as naming the ICP.

Score and Tier

Score accounts by fit (firmographic) + intent (behavioral) + revenue potential, then tier:
- Tier 1 (top ~20%): personalized, multi-channel, high-effort
- Tier 2 (middle ~50%): standard, email-dominant sequence
- Tier 3 (bottom ~30%): low-touch/automated or excluded

ICP also scales with ACV: sub-$5K → founders at 1–50-person startups; >$100K → C-suite + procurement at 500–5,000+ employee firms.

How It Applies to Marketing Factory

The ICP model is the structured filter at the top of every agent workflow: enrichment agents build lists against it, scoring agents tier inbound leads against it, and outbound agents choose channel/effort by tier. Encode it as machine-readable criteria (including the anti-ICP) so agents can exclude bad-fit accounts automatically rather than burning send volume on them.

Referenced from: outbound-playbook