A model that scores your audience by their likelihood to convert. The inputs are behavioral signals, purchase history, demographic attributes, and whatever first-party data you have. The output is a ranked list: high propensity, medium propensity, low propensity. You target accordingly. The model is only as good as the training data and the conversion definition. If your conversion events are poorly tagged or your data is thin, the model finds patterns in noise.
Ad tech glossary
Propensity Modeling
Technical definition
Statistical analysis that predicts the likelihood of specific audience segments to take desired actions or convert.
The ADLINGO translation
Get the full Mega Glossary as a PDF
Nearly 200 ad tech terms, translated. Plus the ADLINGO newsletter to your inbox. Free.
Where this shows up
Could Bill Gates daughter face prison time for cookie stuffing?
Phoebe Gates, daughter of Bill and Melinda Gates, is in the press lately for an alleged "Cookie stuffing" operation. If charged and convicted, penalties could include actual jail time. Do you know how it works?
But the ads served, right?
Brands have a massive gap in understanding what their technology vendors are actually solving on their behalf and what the definition of "paid for" actually means.
News: Brand's $4 Billion Blindspot
Despite the myths, news environments prove to be high-return and brand-safe. With performance upwards of 8.5x higher, avoiding news media might be the most misguided marketing myth yet.