Netflix Recommends

Recommendation engines on streaming platforms, which suggest films and shows based on what a person has already watched, have sparked two opposing reactions. Some hope they will deliver tailored, carefully chosen viewing, while others fear they will trap audiences in narrow bubbles and flatten the range of what gets seen. Both camps tend to assume that these tools are new, work well, and are relied on by most viewers.

This study tests that assumption through a close look at Netflix, the streaming service with the largest subscriber base worldwide. Mattias Frey examines everyday viewers, the company's promotional messaging, its technical workings and business strategies, and earlier precedents for automated suggestions. He finds that these systems are not as revolutionary or threatening as either side claims, and that they are neither as trusted nor as widely used as commonly believed.

The book maps the mix of sources people actually draw on when deciding what to watch in the digital era. It also argues that, although many worry that automated systems are pushing aside human cultural expertise, audiences' appetite for filters, curators, and critics has grown rather than shrunk.