The Ideal Customer Profile

What an ICP is, and what it is not

An ideal customer profile describes accounts, the companies or organizations you sell to, in terms you can check before you ever speak to them: industry, size, geography, the systems they already run, the regulatory regime they work under, a trigger event such as a funding round or a new executive, and the specific problem you solve for them. It answers 'which organizations should be in our pipeline?'

It is not a buyer persona, which describes the people inside those accounts and what each of them cares about. Personas matter enormously once you are in a deal (Module 2 is about the buying group), but a persona without an account filter just produces well-written emails to companies that will never buy.

It is also not your total addressable market. The market is everyone who could conceivably buy. The ICP is the subset where you win often, deliver well and keep the customer. A useful test: an ICP should exclude the majority of companies that superficially look like customers. If yours excludes nobody, it is a market description, not a profile.

Why the profile has to come from evidence

Customer value is far more uneven than intuition suggests. The Institute of Management Accountants' 2010 statement on customer profitability management, summarizing Robert Kaplan and V. G. Narayanan's 2001 work, describes the typical 'whale curve': about 20% of customers generate 150% to 300% of a company's profits, roughly 70% are around break-even, and about 10% destroy 50% to 200% of profits. The losing customers are not obviously bad at the point of sale. They are the ones who demand heavy customization, discounting and support relative to what they pay.

If profitability is that skewed, a profile built on 'accounts that look impressive' will reliably steer sellers toward some of the expensive tail. A profile built on accounts that were profitable to win and serve steers them toward the whale's head.

Retention data points the same way. SaaS Capital's 2025 benchmark of private B2B software companies reports that higher net revenue retention correlates with higher average contract values, and that the companies with the largest contracts have the highest gross retention. Your own segment patterns may differ, which is precisely why the profile should come from your data: the point is to find where you keep customers, not where an industry average says you should.

Building a first profile in an afternoon

Start with three lists from the last twelve to twenty-four months: deals you won, deals you lost to a competitor or to no decision, and customers who are still paying and those who have left.

For each won and retained account, write down the checkable facts: industry, headcount band, region, the system or process you replaced, the event that started the conversation, and how long the deal took. Do the same for the losses and departures. Then look for the attributes that are common among the winners and rare among the losers. You are looking for perhaps three to six attributes; more than that and the profile becomes unusable at the moment a seller has to decide whether to research a company.

Write the result as a short paragraph and a list of disqualifiers. The disqualifiers are often the most valuable part. 'Not companies under 50 people, not companies without an in-house finance team, not companies on system X' saves more selling time than any description of the perfect account.

Finally, write down how confident you are. A profile built from twelve won deals is a hypothesis. Revisit it every quarter with the new wins and losses, and expect it to sharpen.

The profile is what makes outreach relevant

The commercial case for an ICP is no longer only about seller time. Gartner's 2025 buyer survey found that 73% of B2B buyers actively avoid suppliers who send irrelevant outreach. And 6sense's 2025 Buyer Experience Report (vendor research) found that the winning vendor was on the buyer's day-one shortlist 95% of the time. Put those together and the profile becomes a reputational asset: every message to an account outside it risks teaching a future buyer to avoid you, while concentrated, relevant presence inside it is how you get onto the shortlists that decide most outcomes.

In Kavanah, the profile becomes operational in three places. The Accounts page at /clients holds the companies you work with, with a stage (Onboarding, Active, Expansion, Paused) and a health rating (Healthy, Watch, At Risk), which gives you the retained-customer half of the evidence. The Deal Room's deal records hold the won and lost history, and custom fields on a deal (written through update_deal) let you record the profile attributes you decided matter, such as industry or headcount band, so they can be filtered and counted later. And when you prospect, Lesson 3.1's research practice starts from the profile: research the accounts that fit, ignore the ones that do not.