Your Ideal Customer Profile is probably wrong. Not because you didn’t put thought into it, but because you defined it once — likely during a planning offsite or fundraise prep — and haven’t meaningfully updated it since. In a market that shifts quarter over quarter, a static ICP is a liability.
The result is predictable: your team is spending time and budget pursuing accounts that look right on paper but never convert. Meanwhile, your best customers — the ones who close fastest and expand most — don’t match the profile you wrote down.
The Problem with Firmographic ICPs
Most ICPs are built on firmographics: company size, industry, geography, revenue. These are easy to define and even easier to filter in your CRM. But they tell you almost nothing about whether a company is actually ready to buy.
Consider two companies that both match your firmographic ICP:
- Company A is a 200-person SaaS company in fintech. They just raised a Series B, are actively hiring for the problem your product solves, and their CTO has been posting about the exact pain point you address.
- Company B is also a 200-person SaaS company in fintech. They’re in a hiring freeze, their leadership team just turned over, and they signed a three-year contract with your competitor six months ago.
On paper, these companies are identical. In reality, one is a high-probability opportunity and the other is a dead end. Your ICP should be able to tell the difference.
Moving Beyond Firmographics
A modern ICP operates on three dimensions:
- Firmographic fit — The baseline. Industry, size, and geography still matter, but they’re the starting point, not the finish line.
- Behavioral signals — What is the company doing right now that suggests they’re in a buying window? Hiring in specific functions, adopting certain technologies, engaging with relevant content, and expanding into new markets are all signals that separate active buyers from passive matches.
- Persona readiness — Even if the company fits, is the right person in the right role with the right mandate? A VP of Sales who started three months ago has a fundamentally different set of priorities than one who’s been in the role for two years.
How to Fix Your ICP
Fixing your ICP isn’t a one-time project. It’s an ongoing process of hypothesis, validation, and refinement.
Step 1: Audit your closed-won deals. Look at your last twenty wins. What did they have in common beyond firmographics? What signals were present before they entered your pipeline? What was happening inside those companies when they decided to engage?
Step 2: Audit your closed-lost deals. This is where most teams skip. Understanding why deals die is just as valuable as understanding why they close. Look for patterns — were certain industries consistently stalling? Were certain personas unable to get internal buy-in?
Step 3: Build a signal layer. Instrument your ICP with the behavioral signals that correlate with conversion. This means monitoring job postings, technology installs, funding announcements, leadership changes, and content engagement across your target accounts.
Step 4: Create tiered segments. Not all ICP accounts are equal. Build tiers based on signal density:
- Tier 1 — Strong firmographic fit plus multiple active signals. These get your best reps and most personalized outreach.
- Tier 2 — Good firmographic fit with some signal activity. Worth pursuing with targeted sequences.
- Tier 3 — Firmographic fit but no current signals. Nurture with content, but don’t waste direct outbound effort.
Your ICP is not a document. It’s a system. The moment you stop updating it, it starts decaying.
Make It a Living System
The best GTM teams review their ICP monthly. They track which signals predicted conversion, which segments outperformed, and where their assumptions were wrong. They treat the ICP like a product — always iterating, never finished.
If your ICP hasn’t changed in six months, it’s almost certainly wrong. The market has moved. Your product has evolved. Your customers’ needs have shifted. Your ICP should reflect all of that.
Start by looking at the data you already have. The answers are in your pipeline — you just have to ask the right questions.