A unified recommendation model for features summarization

Personalization is a key requirement for most customer experiences in a music streaming service, such as landing page, station song sequencing, or search. A common approach is to develop dedicated personalization ML models, one for each experience, that directly integrate with all the personalization signals alongside experience-specific signals. However, this is not scalable as it is costly for each product team to replicate improvements and experiment with new personalization features. It can result in inconsistencies across the various experiences and slow down adoption of new features in all the models. We propose an approach to make it easier to consume many personalization features in multiple experiences at a low experimental cost. Specifically, we train a customer relevance score that incorporates all those personalization signals, and vend that score so that product teams can integrate it in their models instead of having to integrate with all the personalization signals directly. We validate the approach on the personalized home page of our service across different use-cases.