What is a propensity model?
A propensity model is a predictive model that scores how likely each person is to take a specific action, such as buying, subscribing, or canceling. Marketers use the scores to decide who to target and how.
Last reviewed .
How a propensity model works
You pick the action to predict and gather past data on people who did and didn't take it. The model learns which traits and behaviors go with the action, then scores current customers or prospects.
Common types include propensity to buy, to churn, to respond to an offer, and to upgrade.
Putting scores to work
High scorers might get a faster sales follow-up or a lookalike audience built from them. Low scorers might get lower-cost nurturing.
Remember that a high propensity to buy can mean the person would buy anyway. To learn whether your outreach changes the outcome, test it against a holdout.
Common questions
What data goes into a propensity model?
Typical inputs include purchase history, site and email activity, demographics, and lead source. First-party data tends to be the most useful and the most reliable.
What is the difference between a propensity model and lead scoring?
Lead scoring is often a set of manual rules and points. A propensity model learns its weights from past outcomes with statistics or machine learning.
How do you know if a propensity model is working?
Check that higher-scoring groups convert more often than lower-scoring ones on data the model hasn't seen. Keep monitoring because accuracy can fall over time.
Can a propensity model replace testing?
No. It ranks people by likelihood but doesn't show whether your marketing changed their behavior, which takes a controlled test.
Want help putting this to work?
Tell us where your marketing stands today, and we'll show you what to do first.