Probabilistic modeling is a statistical technique used to take into account the impact of random events or actions in predicting the potential occurrence of future outcomes.
What is probabilistic modeling
Based on the fact that randomness or uncertainty plays a role in predicting outcomes, predictive modeling is used in a wide variety of fields and disciplines, from predicting the weather to potential nuclear fallout.
In the realm of marketing these types of models are most often used to explore consumer behavior and, more specifically, in the mobile ecosystem, in the pursuit of a more holistic view of campaign performance
Probabilistic modeling at AppsFlyer
Probabilistic modeling at AppsFlyer leverages scale and machine learning to measure campaign performance without compromising on privacy.
This form of attribution is based on probabilities, not ID matching with probabilistic modeling parameters collected initially on the click or ad view (if enabled) and again when a given app is launched.
Extensive internal research, based on historical data and analysis between deterministic attribution and AppsFlyer’s probabilistic modeling, and continued development of the algorithms behind this model combine for an aggregated accuracy rate of 92% and coverage rate of 89%.