Ideal customer profile
An ICP is a filter you can apply in ten seconds
An ideal customer profile is a short, checkable description of the accounts you are best placed to win, written specifically so that most prospects fail it.
The test of an ideal customer profile is whether it excludes anyone. Most read like a description of the total addressable market with adjectives, which means every lead passes and the document has done no work. A profile that removes eighty percent of the people you were about to contact is doing exactly what it exists to do.
Situation beats firmographics
Industry, headcount, and revenue are easy to record and weakly predictive on their own. What predicts a win is a situation: they just hired for the role that owns this problem, they outgrew the spreadsheet, they got burned by a competitor. Firmographics tell you who could buy; situations tell you who is about to.
Write the exclusions explicitly
The most useful half of most profiles is the disqualifier list. Companies too small to have the problem, too large to buy without procurement, in a jurisdiction you cannot serve. Written down, these save a rep an hour a day. Left implicit, they get relearned account by account.
One profile per motion, not one per company
A tool sold both to solo operators and to teams has two profiles, and they disagree about almost everything. Trying to write one that covers both produces a description so broad it excludes nobody, which takes you back to the original problem.
It should be checkable from a public profile
If applying your profile requires data you only get after a call, it cannot filter the top of the funnel, which is where filtering is worth the most. Every attribute should be something you can verify from a LinkedIn page or a few posts.
How to do it
Start from your best ten customers
Not your biggest. Your best: fastest to close, cheapest to serve, least likely to churn. Those three qualities usually pick a different ten than revenue does.
Find what they shared before they bought
Look for the situation they were in, not the shape of their business. What had just changed for them is the attribute you can go and look for in other companies.
Write it as a checkable sentence
One sentence with three or four conditions, all verifiable from public information. If it takes a paragraph, it will not be applied consistently by anyone including you.
Write the disqualifiers next to it
Three or four conditions that mean no regardless of how interested somebody sounds. This list is what stops a quarter being spent on a deal that was never going to close.
Test it against last quarter
Apply it to every deal you worked. If it passes the losses and fails the wins, the profile describes who talked to you rather than who bought.
Common mistakes
- A profile so broad that every prospect passes it, which is the same as having no profile.
- Firmographics only, with no situation, so it identifies who could buy rather than who is about to.
- No disqualifier list, so the same bad fit gets requalified every quarter by a hopeful rep.
- Attributes that can only be verified on a call, which makes it useless for filtering inbound or discovery.
How Quillen handles ideal customer profile
Quillen takes an ideal customer profile description per campaign and scores every discovered post against it, so the profile is applied automatically to everything discovery finds rather than living in a document. A vague profile produces a weak score, which is the practical reason to write a narrow one. Discovery is a paid capability, and because each search and each ranked candidate costs credits, a profile that excludes properly is also cheaper to run than one that does not.
Common questions
- What is the difference between an ICP and a buyer persona?
- An ideal customer profile describes the account you want. A persona describes the individual you talk to inside it. You need the profile first, because targeting the right person at the wrong company is wasted effort in a way the reverse is not.
- How narrow is too narrow?
- Too narrow is when the resulting list is too small to fill your pipeline even at a high conversion rate. Almost nobody has this problem. The overwhelmingly common failure is the opposite one.
- How often should it change?
- Revisit it every quarter and expect it to change once a year. Frequent changes usually mean you are reacting to individual deals rather than to a pattern across them.