Your ICP tells you who could buy. Not who's buying right now.

Most ICPs describe a type of company. Industry, headcount, revenue band, maybe a country. It's a photo of who could buy from you. The problem is that "could buy" and "is buying" are two completely different lists, and your team is working the wrong one.

Two companies land in your CRM. Both software, both around 250 people, both in your sweet spot. On paper they're identical. One is hiring a RevOps lead, spinning up a new sales team, and just closed a funding round. The other is freezing software spend and quietly cutting headcount. A demographic ICP puts them in the same campaign, with the same message, on the same day. One is a live deal. The other is a dead account. Nothing in your ICP can tell them apart.

That's the gap. And it's getting more expensive every quarter.

Why the old way is breaking

The demographic ICP was built for a world where talking to a sales rep was how you bought things. That world is gone.

Gartner now puts the number at 17%. That's how much of the entire B2B buying journey a buyer spends with any potential supplier's reps. Split that across the vendors they're considering, and your team gets a sliver of attention, late, after most of the thinking is already done. 61% of buyers say they'd prefer to buy with no rep involved at all. And 73% actively avoid suppliers who send them irrelevant outreach.

Read that last one again. Irrelevant outreach doesn't just get ignored now. It actively disqualifies you. And nothing produces irrelevant outreach faster than a list of companies that "look right" but aren't doing anything.

So you get the pattern every commercial leader knows: connect rates drop, email engagement slides, SDRs burn out working lists that look perfect and convert like cold stone. The targeting feels precise. The results don't.

Here's the uncomfortable part. The list isn't wrong because the firmographics are wrong. It's wrong because firmographics describe potential, and what closes deals is timing. A VP of Operations at a 200-person logistics firm might own a budget that a CTO at a 2,000-person enterprise doesn't. Same category, completely different reality. The photo can't see it.

The shift: from who they are to what they're doing

A behavioural ICP answers a different question. Not "who looks like my customer" but "who is acting like a buyer right now."

That sounds abstract until you make it concrete. Buying behaviour shows up as observable signals, and they fall into a handful of categories:

• Intent: they're researching your category. Pricing pages, comparison content, review sites, third-party research spikes.
• Event: something just changed. A product launch, an acquisition, a regulatory deadline, a new office.
• Financial: they just raised, beat earnings, or announced a budget. Capacity and urgency in one move.
• Hiring: the roles a company posts tell you what it's about to build. A wave of SDR roles or a first RevOps hire is a buying signal you can read in public.
• Technographic: what's already in their stack, and what that says about fit and switching cost.

No single one is enough. A funding round without research is capacity without interest. A research spike without budget is curiosity. The signal you can trust is convergence: when three of these stack up on the same account in the same few weeks, that's not a lookalike. That's a buyer warming up, in public, where anyone paying attention can see it.

And almost nobody is paying attention. That's the opportunity.

What it looks like in practice

This isn't a theory about better lists. It changes how the whole commercial motion runs.

Instead of "we help logistics companies," the message becomes "congratulations on the new distribution centre, here's how teams setting one up usually handle X." Same product. The difference is that one references something real that just happened, and the other is wallpaper. Teams that wire trigger events into their outreach see it convert in the 10 to 14% range, against the roughly 2% that standard cold outreach limps along at. The numbers move because the relevance moves.

The win rate moves too. Teams that genuinely run on intent signals, with sales and marketing working the same data, report 38% higher win rates than teams that don't. Not because they work harder. Because they show up to conversations that are already happening instead of trying to start ones that aren't.

Now the honest part, because hype helps no one. Signals are a starting line, not a score to worship. Only about a quarter of intent signals turn into real opportunities, and plenty of teams drown in noise from tools that promise certainty and deliver a number with no story behind it. A signal that says "Company X is interested" and can't tell you why is almost useless to a rep. The ones that work tell you what happened, so your team can say something a human would actually say.

There's a timing layer too. A signal is worth most in the first 48 hours and decays fast after that. A VP of Sales hire you act on this week is a relevant conversation. The same hire six months later is a generic check-in. Most CRMs are full of signals that already went cold because nobody was set up to act on them in time.

A demographic ICP tells you the market. A behavioural ICP tells you the moment. You need the first to know where to look. You need the second to know when to move.

For European and Belgian scale-ups, there's a twist worth naming. Our buying groups are bigger and slower on purpose. A complex B2B deal now runs through 6 to 10 decision makers, and 74% of buying teams are in open disagreement with each other before you ever show up. Buying is consensus work here, across functions, often across Flanders, Wallonia, and three other countries with different norms. Signals don't override that culture. They aim it. They tell you which roles inside the committee are actually moving, so you can engage the whole group with the right message instead of betting everything on one title and one champion.

Where this lands

A behavioural ICP isn't a tool you buy. It's a layer you build into how you go to market: the signals that matter for your category, the rule for what "ready" looks like, the trigger that puts the right account in front of the right person while it's still warm. That's exactly the kind of thing a Growth Engine is supposed to do. Positioning and ICP first, then the signals and the funnel that turn "this account is moving" into a conversation, then experiments to learn which signals actually predict your deals and which are noise.

We build that Growth Engine for clients, and we run it on ourselves. The companies that win the next few years won't be the ones with the biggest list. They'll be the ones who notice demand the moment it forms, and move before anyone else sees it.

So here's the question worth sitting with: if your best-fit account went quiet and a worse-fit account started showing every buying signal in the book, which one would your team call first? And which one should they?