The ICP Precision Test: How to Know If You've Found Your Real Buyer

A vague ICP is more comfortable than a precise one. If your ICP is 'B2B SaaS founders,' you have permission to talk to any B2B SaaS founder, which means no conversation is obviously wrong. If your ICP is 'technical co-founders of bootstrapped SaaS products under $10K MRR who are managing their first customer success function,' you have to turn conversations away. You have to say no to inbound that doesn't fit. That feels like leaving money on the table, even when it is the correct decision.

The false precision problem is equally common: founders who have a detailed ICP document but whose actual customer list doesn't match it. If your ICP says Series A fintech founders but 60% of your customers are bootstrapped e-commerce operators, one of those is wrong. Either the ICP needs updating or the customer mix needs filtering.

Both problems — vagueness and false precision — have the same root cause. The ICP was defined based on who the founders wanted to sell to, not based on which customers actually retained, expanded, and referred others.

A precision ICP has three attributes, each of which individually eliminates a meaningful portion of the people who might otherwise seem like prospects. Broad demographic categories (company size, industry, geography) don't count unless they alone eliminate 50%+ of the total addressable market.

The first attribute is a specific trigger event — something that happened in the prospect's world that makes the problem urgent now. Not 'they have a scaling challenge' but 'they just hired their third person into a role that has no documented process.' Not 'they need better analytics' but 'they just missed a board target and the CEO asked for a metrics dashboard in the next two weeks.'

The second attribute is a specific behaviour that proves they are actively solving the problem. They have already tried something: a spreadsheet workaround, a competitor product, a manual process, an internal tool. Passive awareness of a problem is not qualification. Active problem-solving behaviour is.

The third attribute is a constraint that rules out otherwise similar-looking prospects. Budget range, technical sophistication, team size, or decision-making authority. The constraint must eliminate real opportunities, or it is not functioning as a filter.

Narrow when you have customers in multiple segments and the retention, expansion, and referral rates are significantly better in one segment than others. The data will tell you where the real PMF is. The difficult part is accepting that you need to de-prioritise the other segments even if they represent more near-term revenue.

Expand only when you have evidence that the core ICP is saturating — when the win rate in your primary segment is high, retention is strong, and you are running out of obvious targets. Expansion based on wanting more opportunities is almost always the wrong reason. It dilutes the product roadmap, confuses the positioning, and spreads customer success resources across segments with different needs.

The right sequence is: narrow to where the best customers are, achieve clear PMF in that segment, then define the next-closest segment with similar trigger events and constraints, and verify before scaling into it.