Ideal customer profile template: build it from who activates, pays and stays
A one-page ICP template you can copy, with steps to fill it from activation, payment and retention data, plus an illustrative worked example.

An ideal customer profile (ICP) template is one page naming the segment most likely to activate, pay and stay, with evidence beside every field. On Tuesday, a founder opens the sign-up dashboard, sees freelancers on top, and asks whether next quarter's ads should chase them.
The one-page ideal customer profile template
Copy this table into a shared doc and fill it one row at a time. The example column is illustrative: it describes a fictional B2B tool, using a segment from the worked example below.
| Field | What to write | Where the evidence comes from | Example (illustrative) |
|---|---|---|---|
| Segment | One group, named in a phrase your team will repeat | Segment comparison from your funnel export | SaaS teams, 51 to 200 employees |
| Firmographic traits | Industry, headcount, revenue, tech stack, location | CRM records plus enrichment | B2B software, 51 to 200 staff, already runs a CRM |
| Use case or job | The job the product does for them, in one line | Sign-up answers and onboarding calls | Report pipeline numbers to leadership every week |
| Activation evidence | Your first-value event and this segment's count | Product analytics | 44 of 80 sign-ups connected data and shared a report |
| Payment evidence | How many pay, how much, and how fast they buy | Billing system and CRM | 24 of 80 sign-ups became paying accounts |
| Retention evidence | How many still pay at a fixed month | Billing cohorts | 21 accounts retained at month 6 |
| Disqualifiers | Observable traits that predict churn or poor fit | Cancellation reasons and sales notes | Solo users with no team to share reports with |
| Where we find them | Places this segment already searches or gathers | Attribution data and "how did you hear about us" answers | Integration directories, referrals from retained accounts |
| Owner and review date | One named person and the next date on the calendar | Leadership decision | Head of growth, first Monday of next quarter |
Fill only the fields you can prove today. Leave the others blank until you reach the steps below, or mark them "assumption" if you have too few customers.
ICP versus buyer persona
An ICP describes a company, and a buyer persona describes a person inside it. Clearbit's template guide draws the same line: the ICP covers the businesses that fit best, and personas cover their decision-makers.
The company view usually comes first in the work. A persona only helps once you know which kind of company it sits inside.
Move any job titles in your current ICP document into a separate persona note this week.
Why profiles built from closed deals mislead
A list of past buyers shows who found you, which is a different question from who was worth keeping. Clearbit's history is a clear example, because its old ICP came from a list of all its customers and their shared firmographic traits.
In its own account, Clearbit says that profile "simply mirrored all our closed/won opportunities." The list was built on people who had "knocked on our door," so sales kept talking to lookalikes of past inbound leads. Clearbit used that version up until early 2020.
The company then measured revenue after the close and found that "around 86% of Clearbit's long-term revenue comes from around 18% of the leads that enter our funnel." That finding led it to rebuild the profile around retention, and it calls the result an "LTV ICP." Read the number as one company's data about its own leads. It is not a benchmark and predicts nothing about your funnel.
Clearbit's advice can look inconsistent here. Its story says "we significantly narrowed our target customer," while its template guide warns "don't get too narrow." Read together, they give a workable rule: narrow where retention evidence supports it, and keep the page open to revision.
Open your current ICP this week and mark every trait that came only from a list of closed deals.
How to fill the template from funnel data
Work through six steps in order, because traits become criteria only after the funnel comparison.

Step 1: export accounts with traits and funnel events
ChartMogul's 2025 panel write-up says a company with an existing customer base can define its ICP from data. Enrich each account with industry, headcount, revenue, tech stack and location.
Then add product and billing data: primary use case, product usage, price paid, speed of purchase, ease of implementation and churn. Include accounts that never activated and accounts that left.
Starting from customers is not the mistake Clearbit described. The mistake is treating every customer as equally good, so keep outcomes beside each account.
Step 2: define a good account
Write the exact event that counts as activation, what counts as paid, and the month you use for retention. Pick the month from your own retention curve, at the point where it starts to flatten.
Clearbit's guide ties the definition to your goal. A retention goal points toward high lifetime value, and a revenue growth goal can point toward high annual contract value. Choose one before you look at the segments.
Step 3: compare segments on the full funnel
ChartMogul's write-up advises looking for the companies that churn least, pay most and have the easiest path into the product. If firmographic groups fall apart, the same write-up suggests grouping accounts by use case or job. Reading activation, payment and retention per segment is the growth marketing work Kreto does with startup teams.
Here is the comparison on invented numbers: one quarter of sign-ups at a fictional B2B tool.
| Segment (illustrative) | Sign-ups | Activated | Paid | Retained at month 6 |
|---|---|---|---|---|
| Freelancers, 1 to 10 | 160 | 72 | 19 | 7 |
| Agencies, 11 to 50 | 120 | 78 | 31 | 26 |
| SaaS teams, 51 to 200 | 80 | 44 | 24 | 21 |
| Enterprise, 201+ | 40 | 14 | 8 | 6 |

Freelancers bring 40% of sign-ups and 12% of retained customers. A dashboard sorted by sign-ups would send next quarter's budget toward the group that keeps 4.4 accounts per 100 sign-ups.
The middle columns show where each segment leaks. Freelancers activate at 45%, but only 26% of activated accounts pay and 37% of payers stay. Agencies activate best at 65%. SaaS teams turn 55% of activated accounts into payers and keep 88% of them.
Agencies and SaaS teams sit close together, at 21.7 and 26.3 retained per 100 sign-ups. Agencies keep more accounts in total, and SaaS teams convert a larger share. Break that tie with revenue per retained account and cost to serve. Pull both numbers and choose the segment that leaves more after serving it. The template above assumes SaaS teams won that comparison.
Enterprise looks respectable at 15.0, but 40 accounts are too few to read. A couple of accounts either way would swing the rate, so keep that row as a hypothesis.
Step 4: ask why they stayed
Call retained accounts in your leading segments and ask what they would use if your product disappeared. Then borrow a question from Superhuman's survey: "What type of people do you think would most benefit from Superhuman?"
Rahul Vohra's reason is practical: "happy users will almost always describe themselves, not other people, using the words that matter most." He groups answers by users who would be very disappointed to lose the product, then uses them to narrow the market. Our product-market fit survey guide covers the setup.
Vohra says results turn directionally correct at around 40 respondents, so a small team can run this survey. Count respondents, not sign-ups: the 40 enterprise sign-ups above left only six retained accounts to ask.
Step 5: write who to exclude
Read cancellation reasons for the segment that loses the most paying accounts. Geckoboard CEO Tom Randle, quoted by ChartMogul, puts the stakes bluntly: "Churn is totally governed by the kind of customers that are coming in and whether they're a good fit."
Treat that as one operator's view, because your churn may have other causes as well. Write each disqualifier as a trait a marketer or salesperson can see before the sale.
Step 6: fill the page
Copy the counts into the evidence rows and name a single owner. Set the review date before you close the document. Run the export this week for the latest quarter whose accounts have reached your retention month, even if the data is messy.
If you have too few customers
An early company can still write a profile, as long as the page says it is a guess.

ChartMogul's write-up calls a profile written before your first 10 customers "assumption-based." Its early path is to book discovery calls with people who experience the problem you are trying to solve.
ChartMogul does not name the point where data takes over, so we use 10 customers as a working line. Passing it does not make every segment readable, as the enterprise row above shows.
Write one best guess in the template and mark each evidence cell "assumption." Book five discovery calls this week, then replace each mark as real accounts arrive.
In a product-led motion, the ICP steers rather than blocks
A self-serve product cannot turn away the wrong sign-ups at the door. Ingrid Bonde Akerlind, Principal Investor at Oxx, said on ChartMogul's panel that "you can't really control who swipes a credit card on your website."
Marketing is where she draws the line instead: "this is who my product is for, and this is who it's not for."
Keep sign-up open and let the profile direct your spending. Point ads, onboarding and roadmap investment at the ICP. Freelancers in the example can still buy, but you stop paying to attract them.
Find one headline on your landing page that speaks to everyone, and rewrite it for the ICP this week.
Validate the profile, then put it to work
A profile earns its place when the market is large enough and each team changes something because of it. ChartMogul's validation step asks whether the addressable market is big enough to reach your next ARR milestone. Count the companies in your CRM or a prospect list that match the firmographic row. Multiply by your average contract value and compare the result with your next ARR target.
The write-up then asks how your roadmap, website and pricing would change, and says to "assess the delta between where you are now, and where you'll need to be." List that delta as tasks with names beside them.
Give each team one line:
- Ads: Target the profile's traits, exclude the disqualifiers, and judge campaigns by activated accounts.
- Site: Write the hero and proof for the segment's job, in the words retained customers used.
- Sales: Check disqualifiers first, and spend discovery time on accounts that match the profile.
- Product: Shorten the path to first value for this segment before serving edge requests.
One framing can land hardest with one buyer group. In Kreto's Peerbie repositioning, the case page says the Collaborative Insights framing "drove the strongest resonance with mid-market B2B buyers." Compare how each message performs by segment before you call one weak.
Revisit the page early if one of ChartMogul's four signs appears:
- Churn rises despite strong onboarding.
- Your best customers look nothing like your average lead.
- Sales cycles lengthen as close rates fall.
- Teams are unclear who the ideal customer is.
Add those four triggers under the review date, and send each team its line this week.
Frequently asked questions
How many ideal customer profiles should a startup have?
Start with one primary profile. Add a second only when its own funnel evidence holds up and you can serve it without splitting the budget. Each extra profile thins your ads, pages and onboarding.
What if our sign-up data has no company information?
Enrich what you can from work email domains, then add one required sign-up question about team size or use case. Keep older accounts in an "unknown" row until new data fills in.
Who should own the profile?
Give it to one person, usually the growth or marketing lead on an early team. Sales, product and customer success bring evidence, and the owner decides what changes.
Does an ICP replace a positioning statement?
No. The ICP says who to pursue, and a positioning statement says why that group should choose you. Use your ICP segment as the evidence for the statement's target customer slot.
Find the segment that keeps paying
Export one quarter of accounts old enough to reach your retention month, and build the funnel table for each segment. If your data sits across tools or the segments refuse to separate, talk to us about finding which segment activates, pays and stays.