Topics · Signal
Recency
Contents
Recency is how recently a post was shared, which Instagram considers alongside other information when deciding what to show someone12. Instagram’s June 2021 explanation describes publication time as one of thousands of details used to decide the order of posts, rather than a standalone rule for visibility1.
How it has changed
- 2016
March
Instagram announced that Feed would move to an order based on likely interest, the viewer’s relationship with the person posting and the post’s timeliness, while retaining all posts2.
- 2018
March
Instagram announced an adjustment making newer posts more likely to appear first and a separate New Posts button test that let viewers refresh to new posts or stay where they were3.
Recency in Feed, Stories and recommended videos
Meta research and patents. These may not describe how Instagram works today.
Instagram’s June 2021 explanation says that selecting recent posts comes before deciding their order in Feed and Stories: it starts with posts from accounts the viewer follows, with exceptions such as ads1. It then gathers information about the posts, their authors and the viewer’s preferences, calling these pieces of information “signals”1.
Meta’s September 2026 research involving Instagram Reels and Facebook Video examines how recency can be affected when a recommendation system chooses a whole sequence of videos together rather than assessing each video separately4. Meta’s research found that the resulting lists changed how much content from different creators appeared, alongside changes in topics, video lengths and familiarity4. Meta’s researchers reported that the share of very fresh videos decreased and treated that decrease as a worsening of freshness—not a quality improvement4.
Availability of information about new posts
Meta research and patents. These may not describe how Instagram works today.
Meta’s July 2026 Reels research identifies a separate obstacle for very recent posts: missing stored information can leave an item out of the system’s estimates of how viewers might respond to it5. Meta’s research found that its comparison system could find that information for fewer than 20 percent of posts under 15 minutes old5. The proposed system maintained 100 percent coverage across post ages, making estimates available for every item, including newly created Reels5.