This study investigates the process of predicting a mobile gaming app installation from the point of view of a particular demand-side platform (DSP) while paying attention to user privacy and exploring the trade-off between privacy preservation and model performance.
The data for this study was collected from real-mobile gaming advertising campaigns run through Bigabid’s DSP platform. Bigabid is a leading DSP that specializes in the gaming app market. Bigabid operates at high-speed bidding on over one million opportunities per second, employing advanced machine learning techniques to train its business models on a vast dataset of over 4 billion unique devices.
This study will demonstrate that while considering user privacy in the gaming app installation prediction model, though the performance of the model is less accurate than when user privacy is not taken into account, our user privacy-aware models still managed to exhibit high performance in predicting app installations.
We conclude that privacy-aware models can still preserve significant capabilities, enabling companies to make better decisions while protecting their user’s data.