ICP Definition & Segmentation
Most ICP documents are a description of who the team hopes to sell to. A useful ICP is a testable claim derived from who actually bought, stayed and expanded, and it has to end up inside your CRM, not inside a slide.
Key Facts
- Focus
- ICP definition and market segmentation
- Category
- GTM Services
- Defined outputs
- 5 deliverables
- Regions served
- India · United States · United Kingdom · UAE · Singapore
- Last reviewed
- 2026-09-10
Your ICP Was Written in a Workshop and Has Never Been Tested Since.
The typical ICP is assembled from a founder's intuition and a competitor's website, written as a paragraph about company size and industry, and then never validated against outcomes. Meanwhile your closed-won data contains a far sharper pattern, often a specific team structure, a technology already in place, or a recent trigger event, and your churn data contains the mirror image, which nobody looks at at all.
Segments are defined by firmographics alone. Industry and headcount are weak predictors compared to what a company does, uses or has recently changed.
The negative ICP is undefined. Knowing precisely who not to sell to protects far more margin than another loosely qualified segment does.
It never reaches the systems. A definition that is not a scored field on the account record cannot route, prioritise or suppress anything.
Deriving an ICP From Evidence and Shipping It Into Production
Model Closed-Won Against Closed-Lost and Churn
We look at all three populations together, because the attributes that predict winning and the attributes that predict staying are frequently different, and the accounts that close fast then churn hard are the most expensive thing a GTM motion can produce. The output is a set of ranked, weighted attributes rather than a paragraph.
Find the Non-Obvious Discriminators
The strongest signals we find are usually operational rather than firmographic: a specific tool in the stack, a role that exists on the org chart, a compliance regime, a recent funding or leadership change. These are also the ones that are actually observable at scale through enrichment, which makes them usable.
Size the Segments Honestly
Each segment gets a defensible count of addressable accounts, built from real source data rather than a top-down market figure. This is where enthusiastic segmentation usually collapses. A beautifully specific segment with four hundred companies in it cannot support a sales team, and it is better to know that before you build the motion.
Write It Into the Stack
The model becomes a scored field on every account, refreshed as enrichment updates, driving routing priority, sequence selection, ad audiences and suppression. It also becomes a monitored dashboard, so when the pattern shifts you find out from the data rather than from a bad quarter.
Deliverables
- A weighted, evidence-derived ICP model built from won, lost and churned accounts
- An explicit negative ICP with automatic suppression rules
- Defensible addressable-account sizing for each segment
- A live ICP score written onto every account and refreshed with enrichment
- Drift monitoring that flags when the winning pattern changes
Is This You?
Strong fit
- You have at least fifty closed-won accounts and some churn history to learn from.
- You sell more than one segment and suspect the resource split between them is wrong.
- You can enrich accounts against the attributes the model identifies.
Not a fit yet
- You have fewer than about twenty customers. Talk to all of them yourself; the sample is too small for this.
- You want validation for a segment you have already committed to. The data may disagree, and we will report what it says.
Your Closed-Won List Already Knows
Bring an export of won, lost and churned accounts to the call. We will walk through what separates them in your data. The discriminator is usually something nobody in the room had written down.
Book a 30-Min Strategy CallSend a Request
We'll be in touch!
Expect a call within 1 business day.
Common Questions
How many customers do we need for this to be meaningful?
Fifty closed-won accounts is a workable floor for finding directional patterns; a hundred and fifty or more supports proper weighting and segment-level confidence. Below fifty we would rather run structured customer interviews with you than produce statistics that imply more certainty than the data supports.
How often should the ICP be revisited?
Formally once a year, but the drift monitoring should be continuous. Product changes, pricing changes and new competitors all move the pattern, and the failure mode is discovering the shift two quarters after a well-executed motion stopped converting.
Can you do this if we sell to several very different segments?
Yes, and it is often where the most value is. Multi-segment businesses usually allocate resources by revenue contribution rather than by unit economics, and modelling each segment separately frequently shows one of them is being subsidised by the others.
What if the data says our best segment is not the one we want to serve?
That happens, and it is a legitimate strategic choice to pursue a segment for reasons the historical data cannot see: a product roadmap, a market moving, a deliberate move upmarket. Our job is to make the trade-off explicit and quantified rather than to make the decision for you.
Related GTM Systems
Lead Enrichment Automation
Multi-source enrichment waterfalls with coverage tracking, cost control and freshness SLAs, so every record your reps and agents touch is trustworthy.
Building an ICP From Closed-Won Data
A practical method for deriving your ideal customer profile from evidence, including the churn analysis most teams skip and the sizing step that matters.
Sales Messaging Systems
Positioning turned into a reusable message architecture: battlecards, sequences and enablement that stay consistent as your team grows.