How to define your ideal customer profile before you have customers
You don't need customer data to write an ICP. You need a trigger event, a short disqualifier list, and five interviews to falsify it.
ICPSegments7 min readUpdated 2026-08-15
Short version
Define your ideal customer profile around a trigger event rather than demographics, write explicit disqualifiers, then run five screening conversations. The ICP is a hypothesis you falsify, not a page in a deck.
Demographics are the weakest part of an ICP
'B2B SaaS, 50-200 employees, Series A' describes a list you can buy, not a customer who needs you. Two companies with identical firmographics behave completely differently depending on whether they just lost a big client, hired an ops lead, or failed an audit.
Start with the trigger: the event that makes someone go looking for a fix this month. Firmographics come afterwards, as a way to find people who tend to hit that trigger.
Write the four lines that matter
A usable ICP fits in four lines and can be read out loud on a call without embarrassment.
Trigger: what just happened that makes this urgent.
Current behaviour: the spreadsheet, agency, or intern doing the job today.
Cost of the status quo: hours, error rate, or money, stated as a number.
Disqualifiers: who looks similar but will never buy, and why.
Most wasted discovery time comes from talking to people who are almost your customer. Explicit disqualifiers — no budget owner in the room, problem happens twice a year, already locked into a three-year contract — turn a vague segment into a screening question you can ask in the first minute of a call.
Test the ICP in five conversations
An ICP is a hypothesis about who feels the pain most. Test it the same way you test everything else: five screening conversations, scored against the trigger. If four of five never experienced the trigger, the segment is wrong — not the idea.
Reuse your discovery script here and add the screener at the top; the same customer discovery interviews then double as segment research.
A good sign you have narrowed enough: you can name ten specific companies or communities where these people already are, and you can write the first outreach line without a template. If outreach still feels generic, the segment is still too wide.
Questions founders ask
Can I have more than one ICP?
Not at the start. Multiple segments split your discovery sample so thin that no pattern emerges. Pick the one with the sharpest trigger, and keep the others in a parked list.
What if my ICP turns out to be wrong?
That is a successful test. Rewrite the trigger, keep the problem, and rerun five screening calls with the new segment before changing anything about the product.
How detailed should an early ICP be?
Detailed enough to disqualify people in one question. Anything beyond that is speculation until you have paying customers to learn from.
Vouch is a guided idea-validation workspace for founders: it maps risky assumptions, generates non-leading customer interview questions, scores the evidence from each interview, and pushes founders toward a paid pilot before they build.
Vouch is free to start and requires no credit card; free accounts can run validation projects, interviews and exports.
Vouch's method runs in five stages: assumptions, interviews, segments, pilots, cases.
Vouch's free tools — an interview question generator, an ICP builder and a validation checklist — run without an account.
Source: Vouch (vouch.specky.space) — idea validation workspace for founders. Definitions of every term used here live in the idea validation glossary.