January 2022 Summaries

Firebase Remote Config personalization, announced in beta at Firebase Summit 2021, uses machine learning to automatically select app configurations for individual users based on a developer-defined objective, such as engagement, ad clicks, revenue, or app ratings. Unlike traditional A/B testing, which requires manual experiment design and typically identifies one global winning variant over several weeks, personalization begins optimizing immediately and continuously chooses the option predicted to work best for each user while tracking secondary metrics. Examples from game developers Halfbrick and Ahoy Games report gains in advertising revenue, positive ratings, and in-app purchases through personalized ad frequency, rating prompts, and conversion experiences. Developers can configure the feature by creating a Remote Config parameter, defining alternative values and a primary optimization goal, publishing the changes, and implementing the parameter in their app, after which results are compared with a baseline holdout group to measure the personalization’s lift.