Most early-stage SaaS founders think their ICP problem is a targeting problem. They are not reaching enough of the right companies. The actual problem is usually deeper. The ICP itself is not specific enough to guide a targeting decision.
An ICP built on firmographics is a starting point, not a definition. Company size, industry, headcount, revenue range: these describe who might be a good customer. They do not say who will be.
The gap between might-be and will-be is where early-stage SaaS companies lose time. Sales, marketing, and product are all following the same ICP. All three are working from a definition that is one layer too shallow.
The Three-Layer ICP Model builds depth progressively. This article is for the founder whose team knows the target segment but cannot agree on which leads are actually worth pursuing.
Why Firmographic ICPs Fail Product and Sales Decisions
The obvious failure mode: the ICP is too broad. A description that fits hundreds of companies in the target market provides no basis for distinguishing which ones will succeed from which ones will struggle.
The less visible failure is that firmographic ICPs are lagging indicators. They describe companies that have already become customers. They say nothing about why those companies succeeded or what they did that churned customers did not.
The deepest failure is a behavioral blind spot. The customers who drive retention, expansion, and referrals almost always share a specific behavioral pattern during onboarding and early use. That pattern is almost never visible in a firmographic profile.
It lives in usage data, call recordings, and success team notes. It is the most predictive information the company has.
The team cost shows up in alignment gaps. Sales has one model of the ideal customer. Customer success has a different one based on who is easy to retain.
Product has a third based on who asks for the features they want to build. Three teams, three partial definitions, no shared foundation.
The Three-Layer ICP Model
The Three-Layer ICP Model builds depth progressively. Layer 1 defines the addressable market. Layer 2 identifies the segment ready to buy today. Layer 3 predicts which customers in that segment will succeed.
Layer 1 establishes the firmographic baseline
Layer 1 is the market definition. It answers one question: what type of company is this product built for?
Layer 1 specifies:
- Company size range
- Industry vertical or use-case category
- Geographic focus, where applicable
- Funding stage or revenue band
These criteria are necessary but not sufficient. They define the pond. Layers 2 and 3 identify which companies in the pond are worth pursuing now.
The most common Layer 1 mistake is too many criteria. A definition with eight requirements often excludes most of the company’s current customers when checked against actual data. Start with two or three non-negotiable criteria and add constraints as Layer 2 data accumulates.
Layer 2 adds workflow specificity and problem depth
Layer 2 narrows from the market to the segment that has the specific problem the product solves. It answers the question a sales rep can ask in minute five of a discovery call: what have you already tried to fix this?
Layer 2 specifies:
- The specific workflow or context where the problem is acute
- What the customer has already tried to solve it
- Why those attempts fell short
The question that surfaces Layer 2 specificity is simple: “What did you try before?” Three consistent answers across ten customer conversations is a Layer 2 pattern worth writing down.
Layer 2 distinguishes “could buy” from “ready to buy.” A company that fits Layer 1 but hasn’t experienced the Layer 2 situation may become a customer later. They are not a good fit today.
Layer 3 identifies the behavioral success signature
Layer 3 is the hardest layer to build and the most valuable to have. It answers one question: what do the most successful customers do in their first 30 days that churned customers did not?
The behavioral success signature is specific and operational. Not “they used the platform more.” Something observable: completed a specific integration in week one, made a targeted query by day ten, added a second team member within 30 days. These events can be built into onboarding and measured.
Layer 3 converts the ICP from a demographic profile into an activation roadmap. Once the behavioral signature is documented, product knows where to focus onboarding. Customer success knows what to check in week two. Marketing knows which usage stories resonate with the right customers.
How the Living Library Updates the Behavioral ICP
At the close of each month, the founder opens the ICP definition and sees it already reflects the latest cohort. No one compiled a report. Layer 3 now carries a fresh pattern: customers who set up a Collection in week one show 40% higher retention at month three.
The Living Library is the working layer of Kiluma. It reads customer success patterns, usage records, and sales conversations as they arrive, then turns them into maintained work. The behavioral ICP is one such artifact, refined each time a new cohort adds evidence. What used to sit in scattered spreadsheets compounds into a definition that keeps pace with the customer base.
So when a new prospect appears, the founder turns to the Conductor. It is Kiluma’s AI that reasons over the company’s own records rather than a generic model. Asked whether a new account’s onboarding behavior matches the Layer 3 success signature, it answers from real cohort history. The founder decides whether to lean in or deprioritize before sales time goes in, on what this business has actually lived.
Build Layer 3 Before You Touch Layers 1 and 2
Start with Layer 3. Pull the last ten customers who have been active for six or more months. Write down the three things they all did in their first 30 days that your churned customers did not.
Then pull the last ten customers who churned before month three. Compare their first 30-day behavior to the first group. The behavioral divergence is Layer 3.
The positioning statement in How to Write a Positioning Statement That Actually Helps Your Team Make Decisions, Article 02, depends on Layer 1 existing. But the Layer 1 criteria that actually exclude the wrong customers are grounded in Layer 3. Build the behavioral signature first. Then let it sharpen the firmographic boundaries.
The ICP That Predicts Success Is Built From Behavior
Most targeting problems resolve themselves when Layer 3 exists. Customers who match the firmographic profile and the behavioral signature convert and stay. Customers who match the firmographic profile but not the behavioral signature churn early and attribute it to product fit.
The Three-Layer ICP Model is the difference between a profile that describes customers and one that predicts them. Try Kiluma free for 14 days at kiluma.ai.