YSocial study: popularity algorithms concentrate visibility, collaborative filtering distributes it

A team of researchers (Morini, Pansanella, Pappalardo, Pedreschi, Rossetti) used agent-based simulations on YSocial—a virtual twin of social media—to study how recommendation logic and network structure jointly affect content and creator visibility. They tested 7 recommendation strategies (including popularity, collaborative filtering, and hybrid) across 2 network topologies (random and scale-free). The results reveal that the recommender logic sets the overall visibility regime: popularity-based algorithms create a self-reinforcing loop where early engagement boosts later exposure, concentrating visibility on a tiny fraction of content and limiting creator visibility to those whose content enters that loop. In contrast, collaborative filtering spreads visibility more evenly across the active catalogue and user base.

When the follower graph shapes candidate selection (as in many real platforms), network structure changes the direction of inequality. Under popularity ranking, creator-level concentration becomes comparable to global popularity, but visibility is systematically redirected toward creators who already have high social status (follower count). Network topology modulates the magnitude of these effects without altering their qualitative pattern. The study demonstrates that visibility allocation must be evaluated across multiple dimensions—content, creators, network position, and temporal feedback—and that controlled simulations like YSocial can help predict how feed design will distribute visibility before deployment.