Define an ICP before prospect data
Make account research testable with inclusion, exclusion and buying-context rules.
In this article
The situation
A target list is only useful when the team can explain why each account is in scope and what conversation would be relevant. Write the criteria before evaluating the size of any database.
Turn a broad market into a testable segment
An ideal customer profile is a working set of account-level rules, not a wish list of company logos. Name one use case, the kind of organization that experiences it, relevant size or process signals, and explicit exclusions. Add the reason a conversation would be timely for the buyer. A field is useful only if the team can verify it before acting.
Keep the buyer or role hypothesis separate from the account fit rule. A company can fit the market while the selected contact lacks responsibility for the problem. Conversely, a familiar job title does not make every company in an industry a good prospect. Write both decisions before comparing database search results.
Challenge the profile with real examples
Choose a set of accounts the team can examine manually, including near misses. Have two people apply the inclusion and exclusion rules and explain disagreements. A criterion that admits almost everyone has little screening value; one that excludes known good-fit organizations may be too narrow or use the wrong signal. Record why each accepted account is in scope and what fact still needs confirmation.
The sample should be small enough to verify thoroughly. Avoid calibrating the profile solely to a vendor's available filters, because that makes the tool dictate the market definition. If an important criterion cannot be searched, decide how the team will verify it manually.
Use the ICP as a control, not a slogan
Once the team agrees on the rule, use the same sample definition to compare data sources, messaging approaches and sales follow-up. Revisit the profile when deal reasons or customer needs change, but date each revision so a later pilot can be interpreted. Do not rewrite the ICP mid-test to make a poor data sample appear successful.
An ICP narrows research; it does not authorize contact. Check channel rules, contact basis, suppression and the usefulness of the proposed message separately.
Work through it
- Choose one industry or use case and specify company attributes, buyer context and exclusions. Avoid a target such as all small businesses.
- Review a handful of real accounts manually to see whether the criteria distinguish useful from irrelevant prospects.
- Decide which fields a tool must provide, what the team can verify and where accepted records will be owned.
What to measure
- Sample accounts meeting the written fit rules
- Fields that are present and usable
- Accounts with a justified next action
Boundaries
- A large estimated market does not prove that an outreach message will be welcome or effective.
- Keep lawful contact methods and suppression separate from account fit.
Questions for your own workflow
- Who has the problem now?
- Which accounts are explicitly out of scope?
- What fact changes the next action?