Watch Time

Watch time is the total number of seconds an account spent watching your Reel, summed across the audience. Instagram reports two flavours in the Reels Insights panel: total watch time (the cumulative figure used for ranking) and average watch time per play (your AVD). Mosseri publicly listed watch time as the single most important Reels ranking signal in January 2025, with sends per reach and likes per reach trailing.

The mechanism is simpler than creators assume: Instagram's model predicts, for any given viewer, how long they will watch a given Reel. Reels that beat their predicted watch time get pushed to more non-followers; Reels that fall short are throttled. The first three seconds carry disproportionate weight — Instagram's 2026 "What impacts your views" disclosure named skip rate as the top-weighted factor in Reels distribution. A skip in second one is treated as a much stronger negative signal than a partial view.

What to actually target

A reasonable internal benchmark is AVD at 70-80% of your Reel's total length. For a 30-second Reel, you want average view duration above 21 seconds. Loops matter too: replays count toward total watch time but not toward first-play AVD, so a tight 8-second loop can outperform a sprawling 60-second cut on raw watch time even if AVD looks lower.

Common mistakes

• Front-loading the payoff. If the hook gives away the answer, viewers leave. Tease the conclusion; deliver it past the halfway mark.
• Long for the sake of long. 90-second Reels with weak retention destroy AVD faster than 20-second Reels with the same content. Shorter Reels with high completion rate compound better.
• Ignoring loop architecture. A Reel that ends on a question or visual cliffhanger pulls viewers back into a second loop, which inflates total watch time without requiring better content.

Watch time + sends per reach together explain almost all Reels distribution variance — see what changed in the 2026 algorithm for the full hierarchy.