Resource Hub>GTM Dictionary>Propensity model

GTM Dictionary

Propensity model

also: propensity score, fit model

A propensity model is a machine-learned estimate of the probability that an account becomes a customer, trained on your closed-won and closed-lost history using only pre-contact attributes.

Two rules make or break it. Train only on data available before first contact, or you leak the outcome into the prediction. And hold out data to validate, because a model that is not back-tested is a point system with better marketing.

Propensity tells you who will buy. It says nothing about who will stay, which is why a durable score pairs it with lifetime value.

Watch: Adam explains it (badly)

ReelsCHILI PIPER
♥💬➤
“It's a crystal ball, except it actually read your last four thousand deals first.”
AAdam
Comedian-in-residence · 0:15

Fifteen seconds of propensity model, as explained by Adam, our comedian-in-residence. If the definition did not land, maybe the joke will.

Reel placeholder. Real cut coming soon.

See propensity model in context in the full methodology.Read the guide