ICP fit on LinkedIn: how to turn an ideal customer profile into an audience that works

Picture this. Sales says the ideal customer is "mid-market IT companies in Europe". Marketing nods. The LinkedIn campaign goes live with exactly that sentence, translated into filters.

Three weeks later the leads come in. Half are freelancers. A quarter are students. The rest work at companies that would never buy.

Sound familiar?

An ICP on a slide is not an ICP in the platform

Most ideal customer profiles are written for people. They describe a company in sentences. LinkedIn does not read sentences. It reads fields: industry, company size, job function, seniority, skills, groups.

Every field is a rough proxy for what you actually mean. The gap between your sentence and the proxy is where budget disappears.

Translate the ICP, field by field

Take the sales sentence and break it into the questions LinkedIn can answer.

Who buys? Not "IT leaders". A job function plus a seniority level. Often that means two or three titles, not twenty.

Who influences? The buying group has more than one person. A technical evaluator, a budget owner, sometimes a security or compliance voice. Each of them deserves a separate audience, with a separate message.

Where do they work? Company size ranges are blunt. A company with a few hundred employees can be a perfect fit or a bad one, depending on how it sells. Use size as a boundary, not as proof.

The three filters that quietly break the match

Industry is self-reported by the company. Plenty of software firms file themselves under something generic. If you filter on industry alone, you lose real buyers and keep irrelevant ones.

Job titles are free text. The same role is written five different ways across five countries. Use job function and seniority to catch the variants, then use titles to tighten.

Skills and groups sound precise but are thin. A skill on a profile says what someone once added, not what they buy today. Treat them as a supporting signal, never as the core of the audience.

Start with the account list, not the filters

If your sales team can name the companies they want, upload them. A matched account list beats any combination of filters for fit, because it starts from your own knowledge instead of a platform guess.

Then layer function and seniority on top. You get the right companies and the right roles inside them.

The trade-off is reach. A list is smaller, and LinkedIn needs a minimum audience size to deliver well. If the list is too small, widen the roles first. Widen the companies last.

A quick fit test before you spend

Pull ten people from the audience preview or from your last campaign. Ask sales one question: would you take this call?

If fewer than most of them get a yes, the audience is wrong. Fix the audience before you touch the creative, the offer or the bid. Nothing downstream rescues a bad match.

Review the audience like you review the ad

We see this pattern often. The ads get tested every week. The audience gets set once and never looked at again.

Put the audience on the same review rhythm. Check who actually clicked, which companies showed up, which roles converted into conversations. Exclude what does not fit. Add what does.

What to do this week

Write the ICP in one paragraph. Split it into buyer, influencer and company boundary. Map each part to a field. Upload the account list if you have one. Run the ten-person fit test with sales.

It takes an afternoon. It saves the next quarter of budget.

If you want a second pair of eyes on how your ICP translates into LinkedIn audiences, that is a conversation we have often.