Algorithm updates

Sends per reach: the Instagram signal behind non-follower reach

Mosseri says send rates matter more for people who don't follow you. What sends per reach is, how to read it in Insights, and how to cut clips for it.

Updated October 4, 20269 min read12 sources

In short

Sends. Instagram head Adam Mosseri said on May 19, 2024 that sends per reach correlate with overall reach more than anything else, named watch time, likes and sends as the top three ranking signals on January 22, 2025, and said on May 26, 2026 that send rates matter more for people who don't follow you. A send is a share of your post to another person, and "per reach" means you divide by the unique accounts that saw the post, not by views. Clip pages get most views from non-followers, so a clip people forward to a friend usually does more for reach than one they only like.

On this page

What Instagram has said about sends, in order
Likes serve followers, sends serve strangers
What Meta's Reels system card says the ranking predicts
Per reach, not per view: get the denominator right
Reposts carry a clip to someone else's followers
How to cut a clip that people want to send
Claims about sends that have no source
What to do now: a sends routine for clip pages
Questions
Sources

What Instagram has said about sends, in order

Instagram has never published ranking weights. What it has published is a series of statements that point the same way. Here they are by date:

Date	Who and where	What was said
May 31, 2023	Instagram, "Instagram Ranking Explained"	The most important Reels predictions include how likely you are to reshare a reel, watch it to the end, like it and go to the audio page
May 19, 2024	Mosseri, reported by Social Media Today	Rates matter more than counts; "sends per reach correlate more, in my experience, with overall reach than anything else"
August 7, 2024	Mosseri, reported by Tubefilter	Views became the main metric across formats; he pointed creators to sends-to-reach as a metric to keep watching
January 22, 2025	Mosseri, reported by Social Media Today	Named "watch time, likes and sends" as the top three ranking signals; sends are slightly more important for unconnected content
November 11, 2025	Meta Transparency Center, Instagram Reels Chaining system card (last update)	Significant predictions include how likely you are to reshare a reel and to share it off Instagram
May 26, 2026	Mosseri, reported by Social Media Today	Focus on engagement rates, "not how much reach your post got"; send rates matter more for people who don't follow you

The Mosseri statements reached us through the outlets named in the table. We did not open his original Instagram or Threads posts. The 2023 article and the system card are Instagram's and Meta's own pages.

Likes serve followers, sends serve strangers

Mosseri splits reach in two. Connected reach is people who follow you. Unconnected reach is people who do not. In January 2025 he said likes are slightly more important for connected content, and sends are "slightly more important for unconnected content." In May 2026 he repeated the split: like rates matter more for followers, and "send rates matter more for those who don't follow you."

He also gave a reason in 2024. Instagram wants to help people find content "they want to connect with friends over", and he called sends a good proxy for that. Our reading of why: a like says "I enjoyed this", while a send says "someone I know should see this". The second is a stronger hint that a stranger will enjoy the post too.

A clip page is almost all unconnected reach. Its followers are few, and its clips are found in Reels and Explore by people who never followed it. On Mosseri's own split, sends are the rate that matters most for that kind of account. This is our application of his words; he did not speak about clip pages.

Note the word "slightly". Mosseri has never said sends outweigh likes by a fixed factor. Watch time remains one of the three named signals, and the Reels card shows that how long people watch a reel feeds the prediction of whether you will reshare it.

What Meta's Reels system card says the ranking predicts

Meta publishes "system cards" for its ranking systems in the Transparency Center. The Instagram Reels Chaining card, last updated November 11, 2025, describes the ranking for the Reels feed. A light model picks about 100 candidate reels, then models predict how you will react to each, and the reels are sorted by a relevance score. The card lists these significant predictions:

How likely you are to reshare a reel.
How likely you are to share a reel off Instagram.
How likely you are to watch more of a reel than 95% of users who watched reels of the same length.
How likely you are to watch less than three seconds of a reel.
How likely you are to comment on a reel.
How likely you are to follow the author.
How likely you are to use the reel's audio in a reel of your own.
How likely you are to tap Interested when the Interested and Not interested options appear.

Three points matter for clippers. First, two of the predictions are about sharing, inside and outside Instagram. Second, watch time is judged against reels of the same length, so a 40-second clip competes with other 40-second reels, not with 8-second loops. Third, the early-skip prediction uses how many people watched at least 3 seconds and how often the reel was skipped soon after opening. This matches the skip rate Instagram added to Reels insights on August 24, 2025: the share of views where people skipped in the first 3 seconds.

A like prediction does not appear in this card's list, although Mosseri names likes as a top signal and Meta's Explore card lists one. Meta says the models and signals "change frequently", so read the card as a snapshot, not a formula.

Per reach, not per view: get the denominator right

Instagram's insights use two counts that are easy to confuse. The Help Center defines them:

Views: "The number of times your reel starts to play or replay." One person who watches three loops adds three views.
Viewers: the number of unique accounts that saw your reel on screen at least once, whether or not it played. Instagram notes that this metric is estimated and in development.

"Reach" in Mosseri's words is the number of people, so Viewers is the closest match in today's insights. Sends per reach is then shares divided by Viewers. This mapping is our reading; Instagram has not published a sends-per-reach figure in the Help Center, and vendor claims about a new share-rate metric in April 2026 have no primary source.

The example below is invented to show the arithmetic. Two clips each get 50,000 views. Clip A reaches 30,000 viewers and gets 600 shares: 600 / 30,000 = 2.0% sends per reach. Clip B reaches 45,000 viewers and gets 450 shares: 1.0%. Clip A had fewer people but each person was twice as likely to forward it. On Mosseri's account, A is the stronger clip.

Campaigns pay on views, and views include replays. Ranking, on Mosseri's account, looks at rates. So the number that pays you and the number that predicts your next clip's reach are different. For the general engagement-rate formula and campaign minimums, see engagement rate.

Reposts carry a clip to someone else's followers

On August 6, 2025, Instagram added Reposts for public reels and feed posts. Meta said reposts are "credited to the original poster", and a reposted post "may be recommended to that person's followers, even if those people don't follow you." The person who reposts can add a short note.

A repost works like a send to a whole audience. If a viewer reposts your clip, it can reach their followers through recommendations, with your account credited. This is the organic version of what clippers try to fake with shoutout accounts.

Do not arrange reposts through groups. Clipster's campaign rules ban sharing content in engagement groups as artificial view inflation. Instagram's Recommendations Guidelines also name "coordinated comment networks" as low-quality. Reposts from real viewers are the only kind worth having.

How to cut a clip that people want to send

Instagram does not say what makes people send a reel. The points below are our reading of the signals above, applied to clipping. Test them on your own account with Trial Reels before you trust them.

Make it stand alone. A viewer forwards a clip when the friend will understand it without context. Start where the idea starts, not mid-sentence, and end at the payoff.
Put the context on screen. Name who is speaking and why it matters. A friend who receives the clip needs that line, and Instagram's 2026 originality rules count context text as an enhancement. See originality rules.
Pick moments that point at a person. A claim people want to argue about, a line that describes a type of person, or a useful number gives the viewer a reason to send it to someone specific.
Win the first 3 seconds. The early-skip prediction and the skip-rate metric both measure this window. See hooks for clips.
Keep length honest. Watch time is compared with reels of the same length. Padding a 25-second moment to 60 seconds lowers the comparison.
Do not ask for the send in words. Instagram has said it will not recommend content that "explicitly asks for engagement through shares, comments, tags, or other actions". See caption rules.

The last point is a real tension. Sends are the most valuable action, and asking for them is named engagement bait. The clip itself has to create the urge to share.

Claims about sends that have no source

Vendor blogs fill the gap left by Instagram's silence on weights. We found no Instagram or Meta source for any of these:

"Sends weigh 3 to 5 times as much as likes." Mosseri said "slightly more important" for unconnected content. No multiplier has been published.
"Watch time is 35%, saves 25%, DM shares 20%, comments 15%, likes 5%." These percentages circulate in 2026 posts without a primary source. Instagram has never published weights.
"A skip rate above 50% means demotion." Instagram added skip rate as an insight. It has not published a threshold.
"Instagram added a share-rate metric to Insights in April 2026." Claimed by vendor blogs. We found no Instagram page that confirms it.

Treat any exact weight as a rumour until Instagram or Meta publishes it.

What to do now: a sends routine for clip pages

Open insights for your last 20 reels. Note Views, Viewers and Shares for each.
Compute shares per viewer. Divide Shares by Viewers. Rank the reels by this rate, not by views.
Find the pattern in your top five. Look at topic, opening line, length and the on-screen context. Write down what they share.
Check the skip side. Open the retention chart and the skip rate. A clip with a high share rate but a high skip rate has a strong middle and a weak opening; fix the first 3 seconds.
Cut the next batch for sends. Self-contained moments, context on screen, payoff at the end, no padding.
Test before campaign posts. Post two cuts of the same idea as Trial Reels and compare shares per viewer on the trials.
Keep captions clean. No "send this to" or "tag a friend" lines.
Repeat monthly. Meta says its models change often, so recheck your own numbers rather than relying on last year's advice.

Questions

What counts as a send on Instagram?

Meta's Reels card describes sharing a reel through Stories, direct message, text or other apps. Its ranking predicts both resharing and sharing off Instagram. Mosseri's comments focus on sending to friends.

Are sends more important than watch time?

Mosseri named watch time, likes and sends together as the top three signals in January 2025. In 2024 he said sends per reach correlated with reach more than anything else in his experience. Instagram has not ranked them with numbers.

Do saves matter?

Mosseri counts saves among engagement actions, but he has not named them among the top three signals. Meta's Explore card lists a save prediction for posts. For clips, treat saves as useful but secondary to sends.

Why do my views and my viewers differ so much?

Views count every play and replay. Viewers count unique accounts that saw the reel. A short looping clip can have far more views than viewers.

Should I ask viewers to share my clip?

No. Instagram has said it will not recommend content that explicitly asks for shares, comments or tags. Let the clip create the reason to share.

Does a repost of my clip count as a send?

Instagram has not said how reposts enter ranking. It has said a repost may be recommended to the reposter's followers with credit to you, so it extends your reach either way.

Sources

We checked every claim on this page against these sources on the date above. Platform rules and rates change often. Found an error? Tell us.

In short

Questions

Sources

Related guides