Anti-Fragile GTM

FAQ · Concepts

Why do you use Laplace smoothing in scoring?

Short answer

To stabilize models trained on small samples of closed-won data.

When you have fewer than 50 closed-won deals, raw scoring produces extreme estimates. Laplace smoothing pulls scores toward the overall conversion rate, producing usable rankings until the sample grows.

It is a statistical humility technique, not a permanent crutch.