Lead Scoring Models That Actually Predict Close Probability

Lead scoring is supposed to help sales teams focus on the right opportunities. In practice, most lead scoring models are built on gut feel and arbitrary point values — and they end up being ignored within six months of launch.

Here's how to build a lead scoring model that your team actually trusts — because it's built from data, not opinions.

The Two Types of Lead Scoring

Fit scoring measures how closely a lead matches your ICP: industry, company size, tech stack, geography, seniority. It answers: is this the type of company that buys from us?

Intent scoring measures behavioral signals that suggest active buying intent: website visits, content downloads, email engagement, pricing page views, competitor searches. It answers: is this company looking to buy right now?

The most predictive scoring models combine both. High fit + high intent = your hottest leads. High fit + low intent = long-term nurture. Low fit + high intent = likely not worth pursuing despite their urgency.

Building From Closed-Won Data

The right way to assign point values in a fit scoring model is to work backwards from your closed-won deals. Look at the 50 best deals you've ever closed: What did those companies have in common? Which firmographic attributes were overrepresented compared to your broader prospect pool? Those attributes get high point values. Everything else gets lower scores or zero.

The Dynamic Scoring Advantage

Static lead scoring models go stale. Your ICP shifts as your product evolves and your customer base grows. The best scoring models are dynamic — they continuously recalibrate based on new closed-won data, updating score weights as real-world outcomes confirm or challenge initial assumptions.

RevWave's ICP Intelligence does this automatically. As you close more deals, the model learns which attributes are most predictive and adjusts scores accordingly. Your reps always know which leads are hottest — based on current data, not last year's assumptions.