Skip to content
Social Rocket

Instagram Reels: Nearly 45 percent of recommendations are less than a day old

Meta's Q3 figures show more fresh Reels in recommendations. Why the 45 percent figure is not a 24-hour rule, and what social teams should actually measure.

From Social Rocket editorial team

This text was created with AI assistance and reviewed editorially before publication. All sources are linked below.

Stock image: Instagram app on a smartphone; no capture of the new features function
Photo: Brett Jordan / Unsplash

Just under 45 percent of Reels recommended on Instagram are now less than 24 hours old. Meta reported this as part of its quarterly results on 21 October 2026. Compared with the second quarter, the share rose by more than eleven percentage points. For creators and social teams, this points to fresher recommendations, but it does not confirm a fixed 24-hour threshold for reach.

What Meta measured – and what it didn't

In the official earnings call, Meta CFO Susan Li explained that changes to Reels ranking had increased Reels watch time by more than four percent. Part of the improvement comes from a model architecture that evaluates sequences of videos rather than scoring each Reel in isolation, and from additional signals about playback time.

The two figures describe different things: 45 percent is the share of particularly new content among recommended Reels, while the effect of more than four percent refers to Reels watch time resulting from specific ranking changes. Neither figure supports a guaranteed reach increase for new accounts or a general performance advantage for every freshly posted video.

Meta also reports that time spent per Instagram user worldwide reached a high in the third quarter and rose by double digits year on year. These are platform statements, not a publicly verifiable forecast for individual marketing accounts.

Facebook and Threads also grow in time spent

On Facebook, Reels watch time also reached highs according to Meta, with growth in the double-digit percentage range compared with the previous year. A new system for finding potentially relevant content is now used for roughly half of Facebook Reels views. Meta puts the resulting gain in Reels playback time at one percent.

On Threads, time spent per person grew by more than 20 percent year on year. This is not a statement that the number of active users rose by 20 percent. The earlier debate about Threads and X concerns a different metric and should be read separately.

Implications for content teams

If you test Reels, you can deliberately evaluate the first 24 hours: Which content generates playback time, reactions and profile visits? Comparing after seven days is just as important, so that strong evergreen videos are not written off too early. Meta's figures make freshness in distribution visible – they do not prove that posting more often alone improves the algorithm.

Our articles on controlling the Facebook Reels algorithm and on Instants and Reels series on Instagram cover further factors in the organic environment. For collaborations, Meta's Creator Marketing Hub is also becoming more important: according to the Q3 call, creator search covers more than ten million profiles.

The finding should not be equated with advertising costs. How Meta's ad prices and delivered impressions have changed is explained in the separate Q3 analysis for advertisers.

Sources: Meta, earnings call transcript from 21.10.2026; Meta, quarterly results from 21.10.2026; Social Media Today, 22.10.2026.

Illustrative image: Brett Jordan / Unsplash. Not a screenshot of the ranking system.