What most founders actually write down as an ICP
Ask a founder to describe their ideal customer and you usually get a job title and a headcount band, pulled from whoever said yes first. That's not a profile. It's a sample of one, and an unrepresentative one at that. Early customers buy on relationship and personal risk tolerance, not on the pattern that will repeat once a media budget is behind it.
The gap shows up fastest in who actually has to sign. Gartner's research on B2B buying puts the typical buying group for a complex purchase at six to ten decision makers, each arriving with several pieces of information they gathered independently. An ideal customer profile that names one job title is describing roughly a fifth of the room. The other four fifths form an opinion from a landing page, a Slack message from a peer, or a competitor's sales deck, none of which the demand gen plan accounted for.
The tell that your ICP is a guess, not a target
The tell is simple, and most founders already have the data to run it. They just haven't looked. Pull closed-won accounts from the last twelve months and split them two ways: renewed or expanded, versus churned inside the first two quarters or never really activated. If those two lists share a firmographic pattern, the ICP was never wrong, it just wasn't written down anywhere marketing could use it. If they don't share a pattern, the company has been selling to whoever would buy, not to whoever should.
That distinction matters more once demand gen enters the picture, because the pool available to target is already thin. Per the Ehrenberg-Bass Institute's B2B research, the buyers actively in-market at any given time are roughly 5% of the addressable market. Spend against a fuzzy ICP and that 5% gets diluted further by prospects who were never going to close, no matter how clean the cost-per-lead looks in the dashboard.
Four inputs that replace the guess
None of these require new research. They require pulling data the company already owns and reading it honestly.
1. The closed-won pattern
Segment renewed and expanded accounts by industry, headcount, tech stack, and deal source. The pattern that repeats across three or more of them is the profile. A pattern that appears once is an anecdote a sales rep likes to tell.
2. The trigger
Firmographics tell you who could buy. The trigger tells you who's buying this quarter: a new head of the function, a funding round, a compliance deadline, a legacy tool being sunset. Without a trigger, targeting is a filter on a lead list, not a targeting decision demand gen can act on.
3. The buying group
Name the economic buyer, the champion, and the person most likely to block the deal, because that third one is where deals actually die. A campaign built around one job title is optimizing for a fifth of the room and hoping the rest goes along.
4. The expansion and referral signal
Accounts that grow, or send someone else in, are telling the company something the closed-won number alone won't. If the accounts that expanded look nothing like the accounts the sales team is currently chasing, that's the clearest signal available that the ICP needs to move.
Where the audit fits, and why it comes before the media plan
The audit is not a deliverable. It is the product. That's true of every Focus4ward engagement, and it's most load-bearing exactly here, because a founder bringing in a first fractional CMO usually wants to see activity fast, and demand gen is the most visible kind of activity there is. The instinct to launch campaigns in week one is understandable. It's also how a wrong ICP survives a full quarter before anyone tests it.
A founder can and should walk into that first engagement with a candidate ICP, built from the four inputs above. The audit's job is to test that candidate against actual closed-won, expansion, and referral data before a single campaign brief gets written, not to discover the mismatch in month three once the spend is already live and the report has to explain why cost-per-lead looked fine while pipeline didn't move.
What changes once the ICP is locked
Targeting narrows first: ad platforms, list-building criteria, and outbound sequencing all get built around the firmographic pattern and the trigger, not a job title. Messaging follows, because a message written for a buying group of six to ten reads differently than one written for a single persona card. Channel choice follows last, once it's clear where the locked ICP actually spends attention. The GTM strategy for B2B SaaS guide walks through how positioning, ICP, channel, and motion stack in order once this step is done. The B2B demand generation strategy piece covers what happens to a demand gen budget that skips straight to channel selection without it.
The campaign will optimize beautifully against the wrong target. Test the ICP before the budget goes in, not after the cost-per-lead report comes back clean.
Keep reading: GTM strategy for B2B SaaS · B2B demand generation strategy · Glossary: ICP and other B2B marketing terms
Frequently asked questions
What is an ideal customer profile (ICP) in B2B marketing?
An ideal customer profile (ICP) in B2B is a company profile, a buying trigger, and a buying group, not a job title and a headcount band. It names the firmographic pattern shared by accounts that renewed or expanded, the event that made them buy this quarter rather than some other quarter, and the roles inside the buying group who have to say yes. An ICP that doesn't disqualify a deal a sales team would currently take isn't an ICP, it's a preference.
How is an ICP different from a buyer persona?
An ICP is a company-level filter: which accounts are worth targeting at all. A buyer persona is a role-level profile inside an ICP-qualified account: how to message the economic buyer versus the champion versus the person most likely to block the deal. Skipping the ICP and writing personas straight from job titles is the most common version of this mistake, because it targets a role that exists at almost any company, ICP-qualified or not.
How do you build an ICP before running demand gen?
Pull closed-won accounts from the last twelve months and split them by outcome: renewed or expanded versus churned inside two quarters or never activated. Look for the firmographic and technographic pattern the healthy accounts share, the trigger event that preceded each purchase, and who sat in the buying group, naming the economic buyer and the likely blocker. Cross-reference against which accounts expanded or sent a referral. Only spend the demand gen budget once that pattern is written down and has disqualified at least one segment the sales team currently chases.
Should you define your ICP before or after bringing in a fractional CMO?
Before the media budget commits, and the first audit is where that testing should happen, not month three of a retainer after ad spend is already live. A founder can walk in with a candidate ICP; the audit's job is to test it against closed-won, expansion, and referral data before recommending where the demand gen budget goes.
Guessing your ICP, or does it hold up against who actually renewed?
Every Focus4ward engagement starts with an audit. Two weeks to test your current ICP against closed-won, expansion, and referral data, before a single euro goes into demand gen. No pitch, no pressure.
Book a call