ICP Scoring Model

An ICP scoring model gives every account a number based on how closely it matches your Ideal Customer Profile. Instead of treating all leads equally, sales and marketing can focus on the accounts most likely to close, close quickly, and grow over time.

How it works

Your ICP comes from your best existing customers: the ones who actually buy, stay, and expand, not the ones you wish you had. The model turns their shared traits (industry, company size, tech stack, growth stage, funding) into weighted attributes and adds them into one score. That score then drives account priority, spend, and territory planning.

How to do it

Study your best customers. Pull your top accounts by revenue, retention, and expansion. List the traits they share: industry, employee count, revenue, technology used, location, and buying triggers.
Pick attributes and weights. Choose 8-12 attributes. Weight them by how strongly each one relates to won deals. Example: industry match 25 points, 200-2,000 employees 20, uses a complementary tool 15, Series B or later 10, strong growth 10, matching job titles 10, geographic fit 10.
Score your market. Apply the model to your account database and any third-party data (for example ZoomInfo or 6sense). Score each account 0-100 and group into tiers: A (80-100), B (60-79), C (40-59), D (below 40).
Back-test it. Score closed-won and closed-lost deals from the past 12-24 months. Check whether won deals consistently score higher than lost deals. If not, adjust the weights.
Put it in the CRM and review quarterly. Show the score on every account record. Direct spend toward A and B accounts. Each quarter, compare scores with actual conversion and recalibrate.

What to measure

Revenue share by tier: closed-won revenue from A-tier accounts, divided by total closed-won revenue.
Sales cycle by tier: average days from opportunity to close, grouped by tier.
Back-test accuracy: share of historical deals where the score correctly ranked won above lost.

Best practices

Weight attributes by data, not gut feel. Check how each attribute relates to closed-won outcomes before assigning points.
Add negative scores. Deduct points for traits that predict long or failed deals, such as a known multi-year competitor contract.
Combine fit with intent. Add behavioral signals (site visits, content downloads, review-site research) to capture timing as well as fit.
Put the score where reps work: dashboards, alerts, and lead routing. A hidden CRM field does nothing.

Common pitfalls

Scoring for aspirational customers. If you have never sold to enterprises, ranking enterprise accounts as A-tier sets sales up to fail.
Using too many attributes. A 20-factor model is hard to maintain and hard to explain. Stay with 8-12 strong signals.
Never updating the model. Your ICP shifts as the product and market change. An old model may be scoring the wrong profile.