Most B2B companies have an ICP definition that reads something like: "Mid-market SaaS companies with 50–500 employees, primarily in the US." That's not an ICP. That's a database filter. It tells you almost nothing about who to target first, what to say to them, or why they buy.
An ICP that actually drives GTM decisions is specific enough to be exclusionary. If your ICP definition describes a universe of 50,000 companies, it's not specific enough to guide channel selection, messaging, or outbound targeting.
Why ICP Definition Fails in Practice
Three patterns produce bad ICPs:
Pattern 1: Built from assumptions, not customers. The founding team defines the ICP based on who they think should buy, rather than analyzing who has actually bought, adopted quickly, and retained. The best ICP data is in your CRM — not in a strategy session.
Pattern 2: Deliberately broad to avoid excluding prospects. Sales teams often push back on tight ICP definitions because they don't want to miss deals. The result is an ICP so broad that the messaging has to be generic to apply to everyone — which means it resonates with no one at full strength.
Pattern 3: Missing the situational layer. Firmographic and technographic criteria describe a population. Situational criteria describe the individuals within that population who are in market right now. This is the difference between targeting an audience and targeting buyers.
The Four-Layer ICP Framework
Layer 1: Firmographic
Industry, company size (revenue and headcount), geography, business model (SaaS, services, marketplace), and growth stage. This is the entry-level filter — the companies that could theoretically buy.
Layer 2: Technographic
What technology does the company currently use? This tells you two things: the sophistication of the buyer, and the integration requirements that will make your product either easy or hard to deploy. A company running Salesforce, HubSpot, and Slack has different expectations and requirements than one running spreadsheets and email.
Technographic data also tells you about adjacent pain. A company running a legacy CRM is probably frustrated with it. A company that just deployed a new analytics stack cares about data quality. These are entry points for messaging.
Layer 3: Situational
What has to be true for this company to be in market for your product right now? Situational triggers include:
- Leadership changes (new VP Sales, new CMO, new CEO)
- Funding events (closed a Series A, Series B)
- Headcount growth signals (actively hiring for specific roles)
- Technology changes (just implemented a new platform)
- Business events (entering a new market, launching a new product)
- Pain signals (public complaints about a problem you solve, job postings for roles your product replaces)
These signals are actionable and time-sensitive. A company that just hired a VP of Revenue Operations is actively evaluating revenue tech stack. They're in market in a way they weren't three months ago.
Layer 4: Psychographic
What does the buyer care about? What are they measured on? What does success look like to them, and what failure would cost them? This layer is built through customer interviews, not database analysis. It's what drives message resonance — the difference between a message that generates a response and one that gets ignored.
How to Build Your ICP From Existing Customers
Pull your 10–20 best customers — highest ACV, fastest time to value, lowest churn rate, most expansions. For each one, document:
- Industry and company size at time of purchase
- Title and function of champion (who drove the deal)
- Title and function of economic buyer (who signed)
- What triggered the purchase decision (what changed?)
- What alternatives they evaluated and why they chose you
- What problem they were solving and how they're measuring success
Look for the patterns that appear in your best customers but not in your average or worst customers. Those patterns are your ICP. The things that appear in your worst customers are anti-ICP signals — use them to exclude, not include.
Testing Your ICP Definition
A well-defined ICP passes three tests:
- Exclusion test: Does your ICP definition exclude more companies than it includes? If it describes most of your market, it's not specific enough.
- Messaging test: Can you write a cold email or ad that would resonate specifically with your ICP and would not resonate with companies outside it? If the message would work for anyone, the ICP is too broad.
- Targeting test: Can you build a list of companies matching your ICP using available data sources? If your ICP is only definable through a conversation, it's not operationalizable for outbound.
Treetop begins every fractional CMO engagement with ICP definition. Without a precise ICP, channel selection, messaging, and outbound targeting all operate below potential. See how the engagement works →
Frequently Asked Questions
ICP stands for Ideal Customer Profile. It's a description of the type of company most likely to buy, adopt quickly, retain long-term, and expand. An effective B2B ICP includes firmographic criteria (industry, size, stage), technographic criteria (current tech stack), situational criteria (what triggers them to be in market), and psychographic criteria (what the buyer cares about and is measured on).
Start by analyzing your best existing customers — the ones that adopted fastest, retained longest, and expanded most. Document the patterns across those customers: industry, size, tech stack, what triggered their purchase, and what alternatives they evaluated. The commonalities across your best customers define your ICP. Use your worst customers to identify anti-ICP signals to exclude.
An ICP describes the company (firmographic, technographic, situational characteristics). A buyer persona describes the individual within that company — their role, goals, motivations, and objections. Both are necessary for complete GTM alignment. The ICP determines which accounts to target; the buyer persona determines how to message to the people inside those accounts.