Walk into a B2B kickoff meeting. Ask to see the ICP slide. Nine times out of ten, what gets pulled up is a list: industry, company size, revenue band, geography. Sometimes a logo grid of "ideal" customers. Maybe a box of tech-stack tags. The slide is called "Our ICP." It is not an ICP. It is a firmographic filter wearing the wrong nametag.
This is the most common targeting failure in B2B. Teams build a firmographic profile, call it an ICP, and run their entire go-to-market on a filter that is missing 60% of the variables that actually predict whether an account will buy. Outbound targets accounts that look like fit on paper. ABM spends six figures on tier-one programs against firmographically-correct accounts that will never close. Demand gen optimizes for vertical and revenue band and ignores the dimensions that explain why two identical-looking accounts behave nothing alike.
This post is about why firmographic ICPs fail, what they miss, and what a real ICP looks like. I have written before about the framework I use to build them. What I want to do here is make the failure mode specific.
What firmographics actually capture
Firmographic data is the layer of company information that sits cleanly in your CRM and most third-party data tools. Industry classification (NAICS or SIC code, sometimes a sub-vertical). Revenue band. Employee count. Geographic location and operating regions. Often funding stage for private companies. Sometimes years in business or growth rate.
This data is useful. It is the starting filter that takes your TAM from millions of companies down to a workable target list. If you sell B2B SaaS to mid-market financial services firms in North America, firmographics narrow the field to maybe 8,000 companies. That is a manageable starting point.
The problem is what happens next. Most teams take that 8,000-company list and call it the ICP. They build content for it. They route leads against it. They size pipeline against it. They never ask whether the filter is doing the work it needs to do — which is to predict whether an account will actually convert. Firmographics, on their own, do not predict conversion. They only predict candidacy.
The story of two identical accounts
Here is a pattern I have seen at every B2B company I have ever worked with. Two accounts in your CRM. Same industry. Same revenue band. Same headcount. Same geography. Same year of founding. They are firmographically indistinguishable.
One closes in three months for a six-figure ARR deal, expands to seven figures over two years, and refers three other accounts. The other never converts despite eighteen months of outbound, three demos, and a custom proposal that took two weeks to prepare.
The firmographic profile told you nothing about that difference. The accounts looked identical and behaved like opposites. What separated them was invisible to the firmographic filter — but not invisible to the data, if you knew where to look.
The accounts that look identical in your CRM are the ones that behave most differently in your pipeline. The variance is in the data your firmographic filter is not capturing.
What firmographics miss — the four invisible variables
The reason two firmographically-identical accounts behave nothing alike is that buying behavior is governed by dimensions firmographic data does not capture. Four of them, specifically.
How they buy
Some accounts make decisions top-down: a single executive evaluates, decides, and the rest of the org rolls. Others run by consensus: five-to-seven stakeholders need to agree before anyone signs. Some accounts buy quickly when a problem becomes acute. Others run a 12-month evaluation regardless of urgency. Some have a defined procurement function that gates every deal. Others have no procurement at all and the budget owner just expenses it.
None of this is in firmographic data. All of it determines whether your sales motion fits. If your motion is built for a six-week procurement-light cycle and the account runs an eighteen-month committee process, you will lose — and the firmographic filter will not warn you.
What's in their stack
Two companies in the same industry can have radically different technology infrastructures. One runs on a modern API-driven stack that integrates with anything in 48 hours. The other runs on a 15-year-old legacy ERP that requires custom middleware and a six-month implementation for any new tool. If your product depends on integration, the second account is not a real prospect, regardless of what the firmographic filter says.
Technographic data exists — vendors will sell you tech-stack signals, and you can sometimes infer them from public sources. Most teams either don't buy this data or don't operationalize it inside the ICP filter. The result is a target list that includes accounts where your product is technically infeasible to deploy.
What's happening inside them right now
Buying happens when something inside the company shifts. A new VP arrives with a mandate. A board sets a strategic priority. A competitor wins a major customer. A funding round closes. A regulatory change forces a new initiative. A previous tool's contract is up for renewal. Each of these is a buying trigger. None of them are visible in the firmographic profile.
The accounts that close fastest are the ones where one of these triggers fired in the last 90 days. The accounts that never close are usually the ones in steady-state operations with no internal pressure to change anything. Firmographics cannot distinguish them. Situational signals can.
Whether they can actually deploy
Some companies buy software and never use it. They have no internal owner for the implementation, no budget for change management, no political alignment behind the purchase. Six months after signing, the tool is unused, the contract gets cancelled at renewal, and the account becomes a churn statistic. This was not a good win — it was a bad fit pretending to be a good fit.
Organizational maturity — the ability to actually adopt and operationalize a new tool — is a real dimension of an ICP. It is not in the firmographic data. Some accounts that look ideal on paper cannot deploy what they buy. Catching this before the deal closes saves you from a renewal cliff a year later.
The cost of running on a firmographic-only filter
Three specific costs compound over time when your ICP is just firmographics.
False confidence in pipeline. The list looks big. The targeting looks precise. The forecast multiplies a respectable conversion rate against an account count that includes hundreds of accounts that will never convert. When the actual pipeline materializes at half the projected number, nobody can tell whether the targeting was wrong or the execution was wrong. It is almost always the targeting.
Burned budget on the wrong accounts. ABM tier-one programs typically spend $5K-$25K per account per year. If twenty percent of your tier-one list is firmographically-correct but situationally-wrong, that's six figures of budget produced no pipeline. Multiply across tiers and the number gets large quickly.
Sales-marketing misalignment. Marketing routes leads from "ICP-fit" accounts. Sales rejects them because the rep can tell within five minutes that the account is not ready, not technically able, or not buying anytime soon. Marketing argues the leads are good. Sales argues they are not. Both are right. The disconnect is the ICP filter underneath them, which marketing trusts and sales does not.
What a real ICP looks like
A real ICP layers four dimensions on top of each other. Firmographics narrows the universe. Behavioral signals predict how an account will buy. Technographic signals predict whether your product fits. Situational signals predict whether they are buying now. Together, the four dimensions produce a filter that survives scrutiny — distinguishing the accounts that can actually convert from the ones that just look right.
That four-dimension model is what I call The 4D ICP Framework. I'll be publishing the complete walkthrough on Monday — every dimension, how to gather data for each, how to layer them together, how to operationalize the filter inside your CRM. If you want to be notified when it goes live, subscribe to The Operator.
For now, the immediate test: pull your current ICP definition. Count the dimensions it actually defines. If the answer is one — firmographic — you are running on the wrong filter. The accounts you are targeting include too many that look right but cannot buy. The fix is not more accounts. It is a deeper filter.
— Erik R. Miller