How to define your ICP before you spend a dollar on ads

ICP, persona and positioning: the difference

• Ideal customer profile (ICP): the type of customer who gets the most value and is most likely to buy. For B2B, often a type of company plus a role.
• Buyer persona: a description of the person inside that customer, with their goals, worries and habits.
• Positioning: how you describe your product so that ICP sees it as the obvious choice.

The ICP comes first. Persona and positioning are built on top of it.

Five questions that define an ICP

1. Who has the problem? Role, type of business, size, location.
2. What triggers the need right now? A hire, a season, a failed tool, a new rule.
3. What is the current workaround? Spreadsheet, agency, doing nothing.
4. Who owns the budget? The user and the buyer are not always the same person.
5. What is a deal-breaker? The reason a customer who looks right will still never buy.

Write the answers in one paragraph. If any answer is “I’m not sure”, that is your next conversation.

Pull it from your CRM, not your head

After your first ten deals, won or lost, tag each one:

• Industry or type
• Size
• Source
• Trigger
• Won or lost, and why

Then look for patterns. The winners usually share a trigger and a source. The losers often share a deal-breaker. That pattern is your real ICP.

Turn the ICP into targeting

An ICP is only useful if it changes what you do:

• Ads: audiences that match, and exclusions for the deal-breakers.
• Outreach lists: filters for role, size and trigger.
• Landing page: the headline in the ICP’s words, about the ICP’s trigger.
• Qualification: the questions you ask on the first call.

ICP template

Our ideal customer is a [role] at a [type of business] with [size], who is [trigger]. Today they [workaround], which costs them [time/money]. The budget sits with [buyer]. We are not a fit when [deal-breaker].

When to rewrite it

Revisit your ICP when:

• Your best customers this quarter look different from last quarter’s.
• Win rate drops in a segment that used to work.
• A new trigger keeps appearing in discovery calls.

Worked example: an ICP from ten conversations

Illustrative example with a fictional product. A founder builds a scheduling tool and believes the ideal customer is “small businesses that book appointments”. That is everyone from dentists to dog groomers. But ten conversations tell a sharper story. So the ICP changes before a single ad runs.

What the conversations showed:

• Four of ten were mobile service businesses: cleaners, groomers, repair techs. All four described the same pain: driving time between bookings.
• Three were clinics. They liked the product but already had software tied to insurance billing.
• Three were solo consultants. They were happy with a free calendar link.

The ICP that came out of it:

Owner of a mobile service business with 2–10 technicians, who is adding new service areas this year. Today they schedule by phone and spreadsheet, which wastes driving time and causes double bookings. The owner decides and pays. Not a fit: clinics tied to billing software, and solo operators who book fewer than ten jobs a week.

The product didn’t change. The target did. Every ad, email and headline that follows is now about driving time and double bookings, for one type of owner.

A script for ICP conversations

Keep the call about their world, not your product. Twenty minutes is enough.

1. “Walk me through the last time [problem] happened.”
2. “What did it cost you: time, money, a customer?”
3. “How do you handle it today?”
4. “What have you tried that didn’t work?”
5. “Who else is involved when this goes wrong?”
6. “If it were fixed, what would change for you?”
7. “Who decides on a tool or service for this? Who pays?”
8. “What would make you say no, even if it worked?”

Write down their exact words. The phrases people use to describe their problem become your headline.

Score fit with a simple rubric

Once you have an ICP, score every new lead against it. It keeps you from chasing everyone.

A score of 8–10 gets a call this week. A score of 5–7 gets a short follow-up. Below 5, a polite “not a fit right now”. Store the score in your CRM so you can check later whether high scores really close more often.

The negative ICP: who to exclude

Knowing who not to sell to saves as much money as knowing who to target. Write a short negative ICP:

• Customers who need features you won’t build.
• Customers whose budget is far below your price.
• Customers who churned quickly in the past, and why.

Use it as exclusions in ad audiences, filters in outreach lists and a check on qualification calls.

ICP for consumer apps

The same questions work for consumer products, with small changes. Replace “type of business” with a life situation, “budget owner” with who pays (sometimes a parent or partner), and “trigger” with the moment the need appears: a move, a new job, a deadline. The goal is the same: one specific person in one specific moment.

Common mistakes when defining an ICP

1. Too broad. “Small businesses” is not an ICP. It is a market.
2. Built from wishes. The customers you want are not always the ones who buy.
3. Demographics without a trigger. Age and job title don’t explain why someone buys now.
4. Never updated. Your best customers change as the product changes.
5. Kept in a document nobody reads. If it doesn’t change targeting and qualification, it isn’t working.

Where to find ICP evidence when you have few customers

Ten deals is a small sample. But you have more evidence than you think, spread across places you rarely look. So collect it in one sheet before you write a word of targeting.

• Your inbox and DMs. Who wrote to you without being asked? Inbound interest is the strongest signal you have.
• Lost deals. The reasons people said no often describe your negative ICP better than any brainstorm.
• Signups that went quiet. Who signed up and never came back? What do they have in common?
• Competitor reviews. One- and two-star reviews describe what buyers expected and didn’t get.
• Community threads. The words people use when they describe the problem, before they know your product exists.
• Your own notes from calls. Triggers and workarounds you heard but never wrote into the CRM.

Put each piece of evidence on its own row with three columns: who, what they said, and where you found it. Patterns show up faster in a table than in memory.

Turn the ICP into messages: a trigger map

An ICP becomes useful when each trigger gets its own message. Here is an illustrative map for a founder selling a quoting tool to event rental companies.

One trigger, one message, one channel. A page or ad that tries to speak to all four triggers at once speaks clearly to none.

Test your ICP in 14 days

An ICP is a hypothesis until real people respond. Here is a two-week test you can run alone:

1. Days 1–2. Pick two candidate ICPs that differ on one dimension, such as size or trigger.
2. Days 3–4. Build a list of 25 people for each.
3. Days 5–10. Send the same offer to both lists, with the first line written for each group’s trigger.
4. Days 11–12. Count replies, positive replies and booked calls for each group.
5. Days 13–14. Keep the group that replied more and booked more. Write down why you think it won.

Fifty emails won’t give you statistics. But the difference between groups is often large enough to see. So treat the result as direction, then confirm it with the next fifty.

A 20-minute ICP review agenda

Run this once a quarter, or whenever win rate shifts:

1. Won deals since last review (5 minutes). Do they still match the ICP paragraph?
2. Lost deals (5 minutes). Any new deal-breaker that belongs in the negative ICP?
3. Triggers (5 minutes). Did a new trigger appear in calls more than twice?
4. Decision (5 minutes). Keep, narrow or rewrite the ICP. Update targeting and the landing page the same week.

Glossary

• Ideal customer profile (ICP): the type of customer who gets the most value and is most likely to buy.
• Buyer persona: a description of the person inside that customer: goals, worries and habits.
• Trigger: the event that makes the need urgent now.
• Negative ICP: customers who look like a fit but should not be targeted, with the reason.
• Fit score: a simple number that shows how closely a lead matches the ICP.
• Win rate: the share of qualified opportunities that become paying customers.

Questions and answers