We describe an integrated platform that aggregates consumption data from multiple sources towards building a unified model for collaborative filtering for more accurate user and content representations. The goal is to provide a framework that combines various signals spanning explicit ratings, implicit information of watching behaviors and metacontent information in a single model that potentially goes beyond the usual goal of maximizing consumption and incorporates metrics that capture “likeness” and “discovery”. We also feed the usage data back into meta-content to determine more accurate content representations that aid in targeting content-based recommendations more effectively.