Most B2B SaaS teams know they need an ICP. Few have actually defined one with enough precision to act on it.
The common failure: defining an ICP as "mid-market B2B SaaS companies" and calling it done. That is not an ICP. That is a market. An ICP tells you which companies within that market are most likely to buy, adopt, and renew — and why.
This framework shows you how to build one that actually shapes decisions.
What is an Ideal Customer Profile (ICP)?
An ICP is a precise description of the company type — defined by firmographic, behavioural, and situational attributes — that is most likely to buy your product, realise its value quickly, and become a long-term customer. It describes a pattern, not a single company.
Why ICP Definition Fails
Three failure modes kill most ICP exercises before they start.
Failure 1: Built in a room, not from evidence. The team assembles for a workshop, somebody writes on a whiteboard, and everyone leaves with a persona document that reflects internal opinions rather than customer reality.
Failure 2: Targeting the dream logo, not the current traction. Companies define the ICP as the customer they aspire to have, not the customers who are already winning with the product. The result is a strategy built on hope.
Failure 3: Never revisiting it. The ICP document is written, shared in Slack, praised at the all-hands, and forgotten. Markets shift. Products evolve. The ICP that was right at Series A is wrong at Series B.
It is better to be the thing for someone than something for everyone. Narrow focus creates depth. Depth creates reputation. Reputation creates inbound.
The ICP Prioritisation Framework: Five Steps
This framework is called the PMF-Propensity Matrix. It scores potential customer segments on two axes — how well your product fits them, and how likely they are to buy — then prioritises the segments that score highest on both.
Step 1: List Potential Segments (Day 1)
Start by generating 3-5 hypotheses about who your best customers are. Do not debate them yet. Write them down.
Sources for hypotheses:
- Your 10 happiest current customers. What do they have in common?
- Your 5 fastest-closing deals. What characterised those buyers?
- Your 5 most churned accounts. What did they have in common?
- Sales team intuition — what patterns do they see?
Example hypotheses for a hiring platform:
- Segment A: Mid-market HR tech companies, 200-500 employees, hiring 50+ people per year
- Segment B: Enterprise manufacturing companies, 1,000+ employees, high-volume hourly hiring
- Segment C: Fast-growth SaaS startups, Series A-B, scaling engineering teams
Step 2: Size Each Segment (Day 2-3)
Use a bottom-up approach to estimate how many companies fit each segment.
Tools: LinkedIn Sales Navigator, Apollo, Clay, Clearbit.
Filter by: industry + headcount + geography + tech stack indicators.
You are not looking for the biggest segment. You are looking for the segment with enough volume to build a business, but tight enough to win with focus. For most B2B SaaS at Series A, that is 500-5,000 addressable accounts.
Record for each segment: total addressable accounts, estimated deal count per quarter at current close rate, and any concentration risk (e.g. over-indexed on one geography or vertical).
Step 3: Score Product-Market Fit (Days 3-5)
For each segment, score your current PMF on a scale of 1-5 across four criteria:
PMF Scoring Criteria
- Pain severity (1-5): How acutely does this segment feel the problem your product solves? Score 5 if it is mission-critical, 1 if it is a nice-to-have.
- Product alignment (1-5): How well does your product solve their specific version of the problem, right now? Score 5 if it maps directly, 1 if significant customisation is needed.
- Customer evidence (1-5): How many customers from this segment have already bought, adopted, and renewed? Score 5 if you have 5+ case studies, 1 if you have none.
- Use-case clarity (1-5): How clearly can your sales team articulate the use case for this segment? Score 5 if reps can demo it in 20 minutes, 1 if it requires custom discovery.
Average the four scores to get a PMF score per segment.
Step 4: Score Propensity to Buy (Days 3-5)
For each segment, score propensity to buy on a scale of 1-5 across three criteria:
- Accessibility (1-5): How easily can you reach the decision-maker? Score 5 if they are on LinkedIn with clear titles and you have relevant content, 1 if they are behind procurement gatekeepers.
- Sales velocity (1-5): Based on current or comparable data, how fast does this segment move from first contact to close? Score 5 if under 30 days, 1 if over 6 months.
- Openness to change (1-5): Is this segment early adopters or conservative buyers? Score 5 if they are actively evaluating new solutions, 1 if they are deeply committed to existing vendors.
Average the three scores to get a propensity score per segment.
Step 5: Plot and Prioritise (Day 5)
Plot each segment on the 2x2 PMF-Propensity Matrix:
| Low Propensity to Buy | High Propensity to Buy | |
|---|---|---|
| High PMF | Invest to unlock. Great fit but slow buyers. Build more proof, reduce friction in sales process. | Primary ICP. Focus here first. These are your best-fit, most-likely buyers. |
| Low PMF | Deprioritise. Wrong fit, wrong timing. Not worth investing in. | Approach with caution. Easy to sell but high churn risk. Consider whether you want this revenue. |
Your primary ICP is the top-right quadrant. That is where you invest your GTM resources first.
Filled Example: Hiring Platform
Scoring three hypothetical segments for a B2B hiring platform:
| Segment | PMF Score | Propensity Score | Quadrant | Decision |
|---|---|---|---|---|
| Mid-market HR tech (200-500 employees, 50+ hires/year) | 4.3 | 4.0 | Top-right | Primary ICP. Focus all GTM here. |
| Enterprise manufacturing (1,000+ employees, high-volume hourly) | 3.5 | 1.8 | Top-left | Invest to unlock. Build enterprise features before pursuing heavily. |
| Series A-B SaaS startups, scaling engineering | 2.2 | 3.8 | Bottom-right | Caution. Easy to sell, but product does not fully serve their needs yet. High churn risk. |
ICP Attributes Reference
Once you have identified your primary segment, document the ICP in full. These are the attributes that define a good-fit company:
| Attribute | What to Define | Example |
|---|---|---|
| Industry | Specific verticals, not broad categories | HR tech, professional services, financial services |
| Company size | Headcount range or ARR range | 200-1,000 employees, $10M-$100M ARR |
| Growth rate | Headcount or revenue growth signals | Growing headcount 20%+ YoY, Series A-C funded |
| Geography | Markets where you can sell and support | UK, DACH, Nordics — English-language operations required |
| Culture/mindset | How they make decisions and adopt new tools | Data-driven HR team, willing to replace legacy ATS |
| Tech stack signals | Tools that indicate fit or readiness | Uses Greenhouse or Lever (not Workday), HubSpot CRM |
| Trigger events | Moments that create urgency to buy | Recently raised funding, new CHRO hired, DEI initiative announced |
| Goals | What the company is trying to achieve | Reduce time-to-hire, improve quality of hire, hit diversity targets |
The Validation Loop
Your ICP is a hypothesis, not a permanent truth. Revisit it every 6-12 months using three signals:
Signal 1: Win rate by segment. If your win rate in the primary ICP segment is declining, something has changed — either the market, the product, or the competition. Investigate before assuming the ICP is still right.
Signal 2: Churn patterns. High churn in a segment you are actively selling into is a clear indicator of PMF problems. Either the segment is wrong, or the product is not delivering what was promised.
Signal 3: Where expansion revenue comes from. Your best customers — the ones who expand, refer, and renew — tell you more about your real ICP than any research exercise. If expansion revenue is concentrating in a segment you have not prioritised, pay attention.
Run a mini win/loss analysis annually. Interview 5 recent wins and 5 recent losses within your primary ICP. The patterns that emerge will tell you whether to stay focused, tighten the definition, or expand.
Common Mistakes
Mistake 1: Confusing ICP with TAM. The total addressable market is not your ICP. Your ICP is the specific slice of TAM where you win consistently. Start narrow.
Mistake 2: Letting enterprise aspirations override startup reality. Selling to enterprise before you have enterprise-grade product, support, and sales motion is how you create churned reference accounts. Win the mid-market first. Use those wins to build the case for enterprise.
Mistake 3: Building the ICP in isolation. Product, Sales, and Marketing all need to agree on the ICP. If they each have a different mental model of the buyer, your GTM will fracture. Run the scoring exercise as a cross-functional session.
Mistake 4: Treating all ICP attributes as equal. Some attributes are hard filters (the company must have X) and some are soft signals (it would be better if they have Y). Be explicit about which is which, or Sales will waste time on near-miss accounts.
How to Use the ICP in Practice
An ICP is only useful if it changes decisions. Here is where it should show up:
- Outbound prospecting: Sales filters all outbound lists against ICP attributes. Any account that does not fit is deprioritised.
- Lead scoring: Marketing scores inbound leads against ICP attributes. Companies that fit score higher and get faster follow-up.
- Deal qualification: Sales uses ICP attributes as qualification criteria. Deals that fall outside the ICP are flagged early and either deprioritised or escalated to the ICP working group.
- Product roadmap: Product prioritises features that solve problems for the primary ICP. Feature requests from non-ICP customers are noted but not automatically actioned.
- Content strategy: Marketing creates content that speaks to the ICP's specific pains, goals, and triggers — not to a generic B2B audience.
Frequently Asked Questions
How narrow is too narrow?
If your ICP has fewer than 500 addressable accounts globally, it is probably too narrow to build a scalable business on. But most early-stage companies define their ICP too broadly, not too narrowly. If you are not sure, go narrower first. You can always expand.
What if two segments score equally on the matrix?
Pick the one where you have more existing customers, stronger case studies, and a clearer sales motion. Execution quality matters more than theoretical fit. Tie-break on evidence.
When should we expand the ICP?
Expand when you have: 10+ reference customers in the primary ICP, a repeatable sales motion with a win rate above 25%, and either product-led signals from adjacent segments (they are signing up without being sold to) or consistent inbound from a new segment. Do not expand to chase revenue. Expand to follow traction.
Who should own the ICP definition?
Product Marketing owns the documentation and the process. But the ICP must be agreed by Sales, Product, and Marketing leadership. If the VP of Sales is targeting a different segment than the one in your ICP document, the document is wrong — or the VP of Sales is wrong. That conversation needs to happen.
ICP prioritisation: scoring model for focus
Most ICP documents are descriptive. Useful ICP frameworks are selective. They force trade-offs so sales, marketing, and product stop chasing contradictory opportunities. Build a scoring model that ranks segments by urgency, fit, and expansion potential.
Core scoring dimensions
- Pain urgency: how expensive the problem is right now.
- Solution fit: how quickly value can be realised.
- Buying feasibility: stakeholder access and budget reality.
- Retention risk: likelihood of sustained adoption.
- Expansion path: potential for multi-team or multi-product growth.
Score each segment 1-5 on every dimension. Weight pain urgency and solution fit highest in early stage. Re-weight toward expansion and retention as you scale.
Turn scores into action
Define three tiers: Tier 1 (primary focus), Tier 2 (opportunistic), Tier 3 (deprioritised). Then align channel budget, SDR time, and roadmap requests to that tiering. If your calendar and spend do not match the scorecard, your ICP process is theatre.
Re-run quarterly using win/loss data and churn patterns. ICP quality compounds when it is maintained as an operating habit, not a static document.