How to choose an ICP for B2B SaaS using evidence

By Julian Moore · Growth and marketing leadership for B2B SaaS and fintech · Published 30 Oct 2026

Compare sales, retention, product use and economics to choose a B2B SaaS ICP, make exclusions explicit and test the gaps before committing budget.

Choose a B2B SaaS ideal customer profile by comparing which customers get value from the product, keep paying and can be served profitably. Then check whether you can reach more of them at a cost and pace the business can support. The decision should name who gets priority, who does not, and what evidence could change your mind.

An ICP describes the organisations you want to pursue and the conditions that make them a good fit. A buyer's job title helps you find the right person inside those organisations, but it cannot settle which segment deserves your budget.

1. Define the segments you are actually choosing between

Start with plausible alternatives from your customer base and sales pipeline. Write each as a combination of a problem, a buying situation and the requirements your product must meet.

For example, a workflow software company might compare agencies coordinating recurring client approvals with in-house teams managing occasional approval projects. This is a hypothetical choice, but it gives the analysis something useful to test: whether frequency and working context explain differences in value.

Use company size or industry where they help explain those differences. Do not impose an employee cut-off simply because it makes a prospecting filter convenient.

Keep the initial definitions consistent while comparing results. If you redefine a segment around every successful customer, you will produce an attractive description without learning whether it predicts anything.

2. Bring five kinds of evidence together

Start with one account-level record joining CRM, billing, product and support data through a reliable customer ID. Include churned customers and lost opportunities, not just the accounts your team likes discussing.

Record the period, sample size and missing fields for each comparison. Where the product, price or sales process changed, separate the evidence before and after the change.

CRM: can these customers buy?

Compare opportunity conversion, sales-cycle length, deal value and reasons for loss or no decision. Define what counts as an opportunity, show wins and losses alongside percentages, and keep open deals visible rather than treating them as losses.

Check how each segment entered the pipeline. Strong conversion from founder introductions does not establish that a cold outbound campaign will perform similarly, while a segment sales rarely pursued may have too little evidence to judge.

Retention: do they keep paying?

Look at customer retention and recurring revenue retention separately. Gross revenue retention excludes expansion; net revenue retention includes it, so growing accounts can offset revenue lost elsewhere. ChartMogul's definitions explain this distinction and how its reporting handles reactivations.

Compare customers at similar points after purchase, with their contract and renewal terms visible. An annual customer who has not reached renewal cannot yet demonstrate renewal behaviour.

Inspect individual accounts as well as averages. If one large expansion explains the segment's apparent strength, the decision depends heavily on that account.

Product use: are they getting the intended value?

Choose an observable action that reflects the product's useful work, such as completing an approval workflow. Compare time to that action, repeated use and how much help customers need to reach it.

Look at the account as well as individual users, and interpret frequency against the job the software does. A monthly reporting tool should not need daily logins to demonstrate useful adoption.

Usage is evidence of behaviour, not proof of the business outcome. Check whether customers describe the benefit you think that behaviour delivers.

Economics: is the revenue worth pursuing?

Put revenue beside the cost of delivering it. Include onboarding, support and any custom work, then examine acquisition cost and the time needed to recover it separately.

Where time or cost records are incomplete, use an explicit estimate or range. Do not make a lifetime-value estimate look settled when the cohort has barely had a chance to renew.

If the segment looks unattractive only at the current price or package, record that as a separate hypothesis. The SaaS pricing guide covers that decision; changing price and target customer together would make the result harder to interpret.

Customer conversations: what explains the pattern?

Speak to retained customers, customers who left and prospects who chose another option or did nothing. Ask what prompted the search, what they were doing before, who had to agree and what happened after implementation.

Use those accounts to investigate the numbers. If customers say the product saves time but support records show repeated manual intervention, resolve the contradiction rather than choosing the more flattering version.

3. Compare the evidence without inventing precision

Use this comparison template before the decision meeting. For each evidence item, complete both segment entries with the finding, its source and date, the number of accounts or opportunities observed, and the main limitation.