Probabilistic Modeling | AppsFlyer

Probabilistic Modeling

We first introduced our probabilistic modeling solution back in 2018, with the rollout of AppsFlyer’s adaptive lookback attribution. Probabilistic modeling leverages our scale and machine learning to measure campaign performance without compromising on privacy.  

Probabilistic modeling is completely anonymized to ensure the highest possible level of privacy. This form of attribution is based on probabilities, not ID matching, therefore lookback windows cannot be defined. The actual attribution window may be anywhere from zero to 24 hours.

Probabilistic modeling  has proven to work on all iOS versions in the past, and has already been tested and validated for iOS 14 —Extensive internal research, based on historical data and analysis between deterministic attribution and AppsFlyer’s probabilistic modeling, the Probabilistic Modeling aggregated accuracy rate is 92% with coverage rate of 89%. AppsFlyer manages to reach the highest level of campaign impact accuracy, leveraging a unique adaptive lookback window thanks to its machine learning technology and immense scale.

« Back to Glossary Index