How to build an ideal customer profile from evidence
Build an ideal customer profile from observed customer outcomes, turn evidence into fit criteria, and test those criteria against future wins and losses.
Free app. Runs on your Mac with the Claude or Codex plan you already have.
What an ideal customer profile is
An ideal customer profile describes the kinds of accounts your product can serve well. It connects a business problem and the conditions for solving it to observed customer outcomes. For a founder, the practical result is a set of rules sales and marketing can apply to an account, with a reason for each rule and a way to handle evidence you still need to collect.
Start with a decision such as which inbound leads deserve a discovery call or which accounts to include in the next sales push. Review how customers bought, used, and benefited from the offer where those facts are available. Include examples that challenge your first impression. The result is a fit rubric, a set of criteria with instructions for deciding whether each one passes, fails, or remains unknown.
What it does for the business
- Make account selection easier to explainA rep can point to the evidence behind outreach and show which fit criteria still need a discovery question. A manager can review the same account using the same rules.
- Direct research toward the gaps that change a decisionExplicit unknowns turn a broad request to research an account into a check for a workflow, requirement, or constraint. The team can stop once it has enough evidence to choose a next action.
- Bring product limits into sales earlierDocumented exclusions help reps identify deals with strict requirements the current offer cannot meet. Product teams can review recurring exclusions when they consider where to invest.
- Write messaging around work customers actually doObserved use cases give marketing a basis for choosing which workflow and outcome to describe. The team can then check whether new prospects recognize the problem in their own words.
Ideal customer profile, step by step
Choose the decision and collect comparable cases
Decide what customer success means before selecting which accounts to study.
What to do
- Specify the offer, period, and business decision. Decide whether you are studying qualified deals, completed onboarding, ongoing use, retention, or another outcome tied to that decision. Track those outcomes separately if they answer different questions.
- Gather relevant account records from your customer relationship management system (CRM) and other approved sources. Use records of wins and losses where available, plus onboarding or product use notes if you are studying whether customers succeed after buying. Record which sources actually support each outcome.
- Define who could enter the review and how cases were selected. Include weak outcomes and losses where the sources support them. Record whether your sample comes mainly from one sales channel, region, or product version.
- Give each account or outcome a stable ID and preserve its source date. Mark follow-up as incomplete when a case has had too little time to reach the outcome you are studying.
Why
An ICP from CRM data can inherit the history of the sales process. If you already sell mostly to one industry, that industry will appear often among both wins and losses. A clear source scope lets you describe what the reviewed accounts show and where you need better evidence.
Compare customer needs and observed outcomes
A useful fit signal ties an observable account condition to the work the customer needs to do.
What to do
- Extract the customer's jobs to be done, meaning the work or change they want to accomplish. Describe the workflow concretely, such as reconciling stock across warehouses, and record where the need appears in a note or source record.
- Capture candidate signals such as workflow, operating model, technical requirements, or the resources needed for adoption. Separate what you could observe before qualification from facts that surfaced after purchase.
- For each outcome group, count the cases with the signal and the comparable cases whose signal value is known. Use the same definitions across groups so you can compare successful and weak outcomes. Record the numerator and denominator so a reviewer can reproduce the observation. Keep missing and conflicting values visible and explain how you handled them.
- Look for accounts with the signal and a weak outcome, and accounts that succeeded without the signal. Note other possible explanations, such as a different implementation path or support available only to certain customers.
Why
Traits that successful customers share can help you form a hypothesis. A contrasting case shows where it may break. Comparing the job, conditions, and outcomes gives you a stronger basis for deciding what to ask on the next call than relying on industry labels alone.
Turn the evidence into criteria you can apply
Each fit criterion needs evidence a rep can find or a question a buyer can answer.
What to do
- Write each criterion in observable terms. Replace a broad description such as operational complexity with a condition that matters for the workflow, such as whether stock reconciliation spans more than one warehouse.
- For every criterion, define pass, fail, and unknown. Specify a source or discovery question that can resolve each unknown. Use an observed pattern to support the rule or label the proposed rule as a hypothesis.
- Document disqualifiers that follow from confirmed product limits. If a required workflow falls outside the offer, record why that requirement overrides industry or company size similarities.
- Keep the few qualification criteria that affect what you do next. Include supporting cases and exceptions and state when an account needs research or human judgment before it can be prioritized.
Why
A company description can look plausible while leaving two reps with different decisions. A runnable rubric explains how each rep can interpret the same facts and shows which questions would change the account's path.
Review the rubric against the underlying cases
A reviewer should be able to reproduce each pattern and explain each exception.
What to do
- Trace each criterion back to the account records, passages, and product facts behind it. Recompute the pattern counts and confirm that unknown values and exclusions are handled as stated.
- Try the rubric on different cases from the evidence. Include a strong fit, an otherwise attractive account with a disqualifier, and an account with unresolved inputs. Check that a colleague using the same source facts reaches the same proposed action.
- Review what the evidence can support. If cases come from one sales motion, flag the resulting limits. Keep relationships that could reflect channel, opportunity selection, or follow-up time as hypotheses.
- Write down the review changes and gaps. If a proposed hard rule depends on an unresolved fact, lower it to a testable hypothesis or hold it until the source can support the decision.
Why
A rubric shapes who gets attention and what you learn next. A weak rule can steer reps away from accounts that would have succeeded. Checking evidence, exceptions, and coverage helps you keep useful distinctions without treating the sample as a picture of the whole market.
Hand off the profile and set up the next test
A customer profile improves when future outcomes can challenge its rules.
What to do
- Give sales and marketing the rubric, examples, and source scope. Define which decisions it can support now, such as research priority, and which require more evidence before the team changes a qualification rule.
- Before outcomes arrive, record fit decisions and unknowns for new target accounts. Keep the later outcomes in a separate field so you can test the original judgment without rewriting it after a win or loss.
- Review the results after accounts have had a reasonable time to progress. Compare like outcomes, record exceptions, and decide which criteria to keep, revise, or remove.
Why
You need observations made before the outcome to test how well a rule helps with a live decision. Carrying the initial rubric and its assumptions into future reviews gives the team a record of how its definition of customer fit changes.
Common mistakes
- Studying only strong customersInclude weak outcomes and relevant losses so favorable traits have a comparison.
- Treating a missing value as a negativeKeep unknown separate and name the check that could resolve it.
- Picking a trait you can see only after purchaseUse account facts available at qualification time for rules applied at qualification.
- Letting an appealing profile override a product gapApply confirmed disqualifiers whenever the buyer needs a capability the offer lacks.
- Extending a local pattern to the marketDescribe the cases reviewed and test the proposed signal on future outcomes.
What a good result looks like
- The offer, decision, outcome window, and source coverage are clear.
- Fit criteria have source evidence or explicit hypothesis labels.
- Observed patterns have traceable counts, denominators, and contrary examples.
- Rules explain what passes, fails, and remains unknown.
- A follow-up test connects each uncertain rule to future wins, losses, or customer outcomes.
And then what? Using the result
- Use the rubric in your next account review and compare how different reps handle the same unknowns.
- Give onboarding and product teams the recurring fit gaps so they can investigate where customers struggle.
- Use evidenced workflows to brief one campaign or outbound list and track which accounts meet the intended criteria.
- Review new wins, losses, and customer outcomes against the original fit decisions before revising the profile.
When to run it
- When the founder hands qualification to a sales teamTurn working knowledge into observable criteria that the team can apply and challenge.
- After a change to the offerUpdate the use cases and exclusions that depend on current capabilities. Compare older cases with the version those customers actually used.
- When wins and adoption tell different storiesCompare purchase and usage separately. Look for conditions that distinguish a good sales conversation from a customer the product can serve well.
- Before choosing a new account segmentUse the profile to frame what you need to learn from the segment. Keep untested fit signals as hypotheses until you have relevant cases.
Questions people ask
What is an ideal customer profile?
An ideal customer profile describes the accounts your offer can serve well and gives criteria you can use to assess fit. Build it from the work those accounts need to do, the conditions they face, and evidence about their outcomes.
How do you build an ICP from CRM data?
Choose a decision, export current account records with known outcomes, compare fit signals and exceptions, and convert the evidence into pass, fail, and unknown rules. Add relevant onboarding or usage evidence if the CRM covers buying better than adoption.
What is the difference between an ideal customer profile and a buyer persona?
An ideal customer profile defines account fit, while a buyer persona describes a person or role involved in the buying decision. Use the profile to choose accounts and use role evidence to decide whom to involve in discovery.
How much evidence do you need for an ICP?
You need enough comparable cases to explain each criterion and show where the evidence runs out. Use a small or narrow sample to propose rules you can test, and reserve firm conclusions for criteria that the evidence supports.
How do you validate an ideal customer profile?
Record fit decisions before new outcomes occur, then compare the original decisions with observed wins, losses, or adoption under a defined window. Review counterexamples and unknowns before changing your rules.
What should an ICP say about a poor fit?
Name the observed condition that rules an account out, such as a required workflow the product cannot support. Give incomplete or conflicting evidence an unknown status and name the check that could resolve it.
More jobs in Leads
Define an ideal customer profile from evidence, on your own Mac.Free app. Runs on your Mac with the Claude or Codex plan you already have. Your files stay on your computer.
Download for Mac