How to build an ideal customer profile for your SaaS business

Executive overview

Founders consistently confuse their total addressable market with their ICP — and trying to serve everyone means resonating with no one. An ICP is not a one-time document; it's a strategic decision about which segment has the most urgent, important problem and is least served by competitors.

The framework covers three principles: narrowing your ICP to the next revenue stage (not forever), using win/churn data to choose between competing ICPs, and running a two-by-two discussion to define what your ideal customer is — and is not. A bonus principle covers how to operationalise the ICP so it doesn't collect dust.

The ICP that works is not the broadest one — it's the one where the problem is urgent, important, and underserved.

ICP is a stage-by-stage decision, not a permanent definition

• Your ICP is not your TAM. The TAM is the full market; the ICP is the specific segment you target to reach the next revenue milestone.
• Trying to serve three ICPs simultaneously dilutes messaging and targeting unless you have separate teams for each.
• Amazon started with books online; it expanded only after dominating that niche.
• Focus the ICP on where the problem is most urgent, most important, and where competitors are weakest.
• Once you reach the next stage, you can expand — more resources means you can go after additional segments.

Using data to choose between competing ICPs

• Most founders arrive with two to five viable ICPs. Choosing between them is a strategic inflection point, not a checklist task.
• Even at $100K ARR you have more data than you think — won deals, lost deals, and pipeline patterns all carry signal.
• Key metrics to compare across ICP candidates: win rate, churn rate, average contract value, LTV.
• Data narrows the options; qualitative discussion closes the gap and surfaces competitive dynamics the numbers miss.
• An ICP that looks strong on win rate may still be wrong if the TAM is too small or competition is intensifying.

The two-by-two discussion: defining your ideal and non-ideal customer

• The most valuable ICP work happens in the discussion, not the document.
• Map competing ICP options on a two-by-two using two defining attributes — different for every company.
• The two-by-two reveals which customers have both attributes (ideal), one (marginal), or neither (non-ideal).
• Firmographic, demographic, and psychographic data matter — but the real unlock is identifying the two attributes that together define your best customer.
• ToutApp's lesson: resisting the urge to stay broad (recruiters, HR, sales) and committing to salespeople unlocked a $5B+ category.

Operationalising the ICP so it drives behaviour

• An ICP that isn't embedded in go-to-market motions collects dust regardless of how good it is.
• At Marketo ($450M ARR), the ICP was hardwired into incentives: marketing was credited only for ICP leads, not all leads.
• Sales reps received full commission (10%) on ICP opportunities versus reduced commission (7%) on non-ICP deals.
• This alignment worked because ICP deals had higher win rates, shorter sales cycles, higher LTV, and lower churn.
• Earlier-stage teams should at minimum track ICP leads and ICP opportunities as separate metrics.