ICP Definition: The 4-Layer Framework That Gets Your Outbound Working in 2026

The static attributes of the company.

The mistake: being too broad. "B2B SaaS, 10-1000 employees" includes 50,000 companies. Your outbound cannot personalise at that scale.

• Industry: Specific verticals (not "B2B SaaS" - "B2B SaaS, 50-200 employees, US/UK, Series A-C")
• Size: Employee count (50-200), revenue ($5M-$50M), or both
• Geography: US, UK, EU, or specific
• Stage: Startup, growth, mature
• Funding: Bootstrapped, Series A-C, PE-backed

The technology stack of the company.

Example: "B2B SaaS, 50-200 employees, US/UK, using HubSpot Marketing Hub Pro or higher, NOT currently using Clay or Apollo for enrichment."

The technographic layer is the highest-intent filter. If they use a competitor, hard to convert. If they use a complementary tool, easy to convert.

• Current tools: What CRM, marketing automation, sales tools, data tools they use
• Recent changes: New tool adopted, tool replaced, integration added
• Tool maturity: Are they on a basic plan or enterprise?

The events that indicate the company is in-market now.

Signal-based outreach converts 3-5x higher than generic. The signal is the personalisation. The signal is the timing.

• Funding event: Just raised Series A (60-180 days ago is peak)
• Leadership change: New VP of Sales, CRO, Head of Ops
• Hiring surge: 3+ sales or ops roles posted in 30 days
• Expansion: New office, new region, new vertical
• Negative signal: Layoffs, missed earnings, customer complaints
• Tech adoption: New CRM, new tool, new integration

The specific humans at the company who will buy.

The person layer is where most teams stop. They define a role and call it an ICP. But the role is only 25% of the person filter.

• Role: VP of Sales, Head of RevOps, Director of CX, CTO, CFO
• Seniority: IC, manager, director, VP, C-suite
• Tenure: In role for 6+ months (not a fresh hire)
• Background: Previously at a similar company, previous founder, etc.
• Engagement: Has engaged with your content, attended your webinar, etc.

A 50 to 200-person B2B SaaS in the US or UK that:

And the buyer:

That is a falsifiable ICP. You can build a list against it. You can personalise against it. You can measure conversion against it.

• Uses HubSpot Marketing Hub Pro or higher
• Has 3+ open sales or ops roles on LinkedIn
• Recently raised Series A or B (last 6 months) OR
• Recently hired a new VP of Sales or Head of RevOps
• VP of Sales, Head of RevOps, or Director of Sales Ops
• In role 6+ months
• Previously scaled a similar team
• Has engaged with content on outbound, AI workflow, or sales ops topics in the last 30 days

1. Too broad. "B2B SaaS" or "mid-market" is not specific enough.
2. Only firmographic. No technographic, no signal, no person filter.
3. No falsifiability. If you cannot say "this company is not in my ICP because X," the ICP is too soft.

Pull your last 20 closed-won and last 20 closed-lost deals. For each, score against your proposed ICP layers.

If 80%+ of closed-won deals match and 80%+ of closed-lost do not, the ICP is validated.

If not, refine. The pattern in your closed-won deals is the actual ICP. The pattern in your closed-lost deals is what you should be filtering out.

• Week 1: Pull closed-won and closed-lost data. Identify common patterns.
• Week 2: Define all 4 layers. Write the ICP on one page.
• Week 3: Build a list against the ICP using Clay, Apollo, or LinkedIn Sales Navigator.
• Week 4: Test outbound against the ICP. Measure reply and meeting rate. Refine.