The Approach: Data-Driven ICP Discovery
Start with your closed-won deals and work backwards. Analyze win rates and deal sizes across multiple dimensions. Let the data reveal who your best customers actually are, not who you think they are.
Deals
Dimension
Patterns
Your ICP
Iterate
Step 1: Analyze Win Rates by Dimension
For each of these parameters, slice your deals and compare win rates and average deal sizes. The gaps will surprise you.
Industry
Which verticals close at the highest rate? Which have the largest deals?
Company Size
What's the sweet spot? Too small = low value, too large = long cycles?
Job Title / Seniority
Who are the actual buyers? Directors? Engineers? C-suite?
Region
Where do you win most? Geographic patterns often surprise.
Lead Source / Intent
Which sources produce customers, not just leads?
Channel / First Touch
Direct vs. paid vs. organic: what actually converts?
Step 2: Spot the Patterns
When you analyze the data across dimensions, clear patterns emerge:
- Your top 3 industries may account for 70%+ of revenue
- Mid-market companies often convert at 2-3x the rate of enterprise
- Demo requests may close 5x better than whitepaper downloads
- Certain regions consistently outperform, so don't spread budget evenly
- The "ideal buyer" title might not be who you expect
Step 3: Consolidate Into Your ICP
Combine the top-performing segments across all dimensions into a tiered definition:
This tiered ICP becomes the foundation for a lead scoring system you can implement directly in your CRM.
Step 4: Validate & Iterate
Your ICP is a living document. It should get sharper over time.
Test
Score recent deals. Does the ICP predict actual winners?
Verify
Ask your sales team: do the data patterns match their experience?
Refine
Recalibrate quarterly as new deals close and the data grows.
What You Get at the End
Documented ICP
Backed by data, not opinions
Scoring System
Ready to implement in your CRM
Quality Baseline
Measure marketing going forward
Shared Language
"This is a Prime lead" and sales + marketing are aligned