Stop Boxing Your ICP So Tight You Kill Your Own Pipeline

I see companies box their ICP so tight they kill their own pipeline.

This isn't a hypothetical. I watched it happen at a company doing $50M ARR.

Their ICP was locked down. VP and above. Enterprise only. Specific verticals. Specific tech stacks. Their SDR team had targeting criteria that read like a 40-point checklist.

And they were wondering why pipeline was drying up.

The deal nobody expected

During a POC with MidBound, we surfaced a visitor on their pricing page. An analyst. Not a director. Not a VP. An analyst at a target account.

Their SDR team would never have prospected this person. Not in a million years. Analysts weren't in the ICP. They didn't pass the title filter. They didn't match the seniority threshold.

But this analyst was on the pricing page. For three minutes. Comparing features. Looking at case studies. Doing exactly what someone does when they're evaluating a purchase.

The account executive reached out. Had a real conversation. Turns out the analyst was leading the evaluation internally. The VP everyone was trying to get to? He'd delegated the entire buying process to this person.

That deal closed at $30K.

From an analyst. That nobody would have targeted. That MidBound surfaced because we show you people, not personas.

The ICP trap

Here's what happens when you over-index on ICP.

Your marketing team builds an ideal customer profile. Makes sense. You want to know who to target. So they define the titles, the company sizes, the industries, the tech stacks.

Then sales takes that ICP and turns it into a filter. A hard filter. If someone doesn't match, they don't get a sequence. They don't get a call. They don't exist.

The problem is that buying has changed.

The person signing the contract is rarely the person doing the research. The person doing the research is rarely the person with the title you're targeting. The person with influence over the decision might be three levels below the person you think matters.

And your ICP filter just deleted all of them.

Titles lie

I need to say this directly because I see it everywhere.

Titles lie.

"Director of Revenue Operations" at a 50-person startup has more buying power than a "VP of Sales" at a 10,000-person company. An "analyst" running an evaluation committee has more influence over a deal than the executive who'll eventually rubber-stamp it.

When you build your prospecting around titles, you're building on assumptions. And those assumptions get more wrong every year as org structures flatten and buying committees sprawl.

The $50M ARR company I mentioned? They had three SDRs dedicated to that target account. For months. Running sequences on the VP of Engineering and the CTO. Neither responded. Neither was involved in the evaluation.

The analyst was doing all the work. And nobody was talking to the analyst.

What your website traffic is telling you

Here's what most teams miss.

Your website is the most honest signal you have. Nobody visits your pricing page by accident. Nobody spends three minutes on your case studies because they're bored. If someone is on your site, they have intent. Real intent. Not "they match our ICP so they probably need us" intent.

The problem is that 97% of website visitors leave without converting (InsideSales research). They don't fill out a form. They don't book a demo. They just browse and leave.

And your ICP framework can't do anything with anonymous traffic. It can only work on the 3% who raise their hand.

Person-level identification flips that. You see who's actually on your site. Their real identity. Their title, company, LinkedIn profile, email. And then you make a decision.

Sometimes that person matches your ICP perfectly. Great. Fast track them.

Sometimes they don't match your ICP at all. They're an analyst instead of a VP. They're at a company you've never heard of. They're in a vertical you haven't targeted.

But they're on your pricing page. Right now. With real intent.

What are you going to do? Ignore them because they don't fit the spreadsheet?

The framework fix

I'm not saying throw away your ICP. That would be stupid.

I'm saying use it as a guide, not a gate.

There's a difference between "we primarily target VP+ at mid-market SaaS companies" and "we ONLY talk to VP+ at mid-market SaaS companies." The first is a strategy. The second is a blindfold.

Here's what I'd do instead:

Tier your ICP. Tier 1 is your perfect fit. Match on every criteria. These get the full treatment. White glove. Fast track. Personal outreach. Tier 2 is close enough. Right company, wrong title. Right title, wrong company size. These still get attention, just a different sequence. Tier 3 is anyone showing real intent on your site regardless of how they fit the matrix.

Let behavior override demographics. If someone's on your pricing page for the third time this week, I don't care what their title is. That's a buyer signal. Act on it.

Watch for non-obvious stakeholders. The $30K deal from the analyst wasn't a fluke. Buying committees are bigger and more distributed than your ICP model assumes. The person doing the research is often the person who decides.

The cost of being too precise

Pipeline doesn't die all at once. It dies in the margins.

It's the analyst you filtered out. The mid-level manager you skipped. The "too small" company where the founder would have bought in a week.

Each one is invisible in your reporting. You can't track deals you never started. You can't measure the pipeline you killed before it existed.

But it adds up. And six months later, your team is sitting in a pipeline review wondering why the numbers are soft.

The answer might be simpler than you think. You boxed your ICP so tight that the real buyers couldn't get through.

Open the box. See who's actually on your site. Talk to the people who are showing up, not just the people who match the spreadsheet.

Frequently Asked Questions

What is an ideal customer profile in B2B?

An ICP is your definition of who your best customers are. Title, company size, industry, tech stack. It's supposed to focus your sales team. The problem is when teams turn it into a hard gate instead of a guide. They filter out everyone who doesn't match perfectly and kill pipeline in the process.

Why does rigid ICP targeting hurt B2B pipeline?

Because buying has changed. The person signing the contract isn't the person doing the research. An analyst running the evaluation has more influence than the VP who rubber-stamps it. When your ICP filter only targets VP and above, you delete the people who actually drive deals. Pipeline dies in the margins.

How should B2B teams tier their ICP for better results?

Three tiers. Tier 1: perfect fit on every criteria. White glove treatment. Tier 2: close enough. Right company, wrong title. Right title, wrong size. Still gets attention. Tier 3: anyone showing real intent on your site regardless of how they match the matrix. Behavior beats demographics.

What role do non-obvious stakeholders play in B2B buying?

A massive role. Buying committees have 6-10 people (Gartner research). The analyst doing the evaluation. The ops person who'll own the tool. The technical lead assessing integrations. Your ICP model probably excludes all of them. But they're the ones on your website doing the actual research. Ignore them and deals die.

How does person-level identification solve the ICP problem?

It shows you who's actually on your site. Not who matches your spreadsheet. When an analyst is on your pricing page for 3 minutes comparing features, you see them. Their name, title, email, LinkedIn. You decide whether to engage based on behavior, not just firmographics. A $30K deal came from exactly that.

Can you identify website visitors who don't match your ICP?

Yes. MidBound identifies all visitors who can be matched, regardless of whether they fit your ICP criteria. You set the filters for what gets surfaced to your team. But the data is there. So when a non-ICP visitor shows strong buying behavior, you see it and can act on it instead of losing the deal.