The quiet tally. The core mechanic is simple: the same interaction data that already feeds the algorithm gets a second job. Every like, save, or lingering scroll nudges the recommendation engine the way it normally would, but it also quietly builds up a tally against categories like food, music, pets, travel. Nothing new is being collected here. I'm just giving a second life to data the app was gathering anyway, and making it work for the person generating it, not just the platform.