User satisfaction

Instagram's goal for satisfying content discovery is to help people find content they are excited about, Mosseri explained in August 20262. Meta's Transparency Center describes how suggested Feed posts are ordered according to their predicted value to the individual viewer, so recommendations more closely match that person's preferences8. To make these predictions, the system considers information such as the viewer's interests and how they have interacted with similar content8.

Instagram's work on a more positive experience also includes research into how people feel using the app, according to Meta's Newsroom1. In May 2021, Meta's Newsroom described collaboration with outside experts and an Instagram Guide developed with The Jed Foundation and creators to offer advice on managing pressure online1. Addressing comfort with time spent, Mosseri explained in June 2025 that daily limits and reminders can notify people when they reach their chosen amount of use3.

Meta patents describe ways to collect direct judgments about feed content that could inform assessments of its quality, without establishing that Instagram uses these methods to measure satisfaction79. One Meta patent describes asking people to choose between two feed items, then using those choices to build or adjust a system that scores content quality7. Another Meta patent describes checking whether human raters are consistent and representative of the intended audience, then using their ratings to compare and refine systems that order feed content9.

Meta's engineering research explores ways to support lasting satisfaction through recommendations, although the work described here concerns Facebook Reels and recommendation systems more broadly, not established Instagram practices56. Meta's engineering team describes using Facebook Reels survey responses about whether videos match people's interests to adjust recommendations, including recommendations of niche content5. Meta's research finds that matching interests involves audio, production style, mood and motivation, not just a video's topic5. In broader recommendation research, Meta's engineering team recommends varying creators, formats and topics while considering longer-term interests, warning that optimizing only for measures such as likes can eventually produce boredom and disengagement6.