Instagram Algorithm 2026: How the Views-First System Actually Works
A View on Instagram is no longer a vague cousin of impressions. Sprout Social’s 2026 guide says it is now the primary metric across Reels, Stories, photos, and carousels, and that it counts every time a post hits a screen — including repeats. Ranking still is not one algorithm. Operators who treat one viral clip as a network-wide asset, or who fill accounts with other people’s posts, are reading the wrong dashboard.
Key Takeaways
- Sprout Social (January 28, 2026) says Instagram moved to Views as the primary metric across all formats, and that a View counts every screen appearance, including repeat views.
- Hootsuite (July 15, 2026), citing Mosseri, names watch time, likes, and shares (or sends) as the signals that matter for both connected and unconnected reach. Sprout, also citing Mosseri, calls DM sends arguably the strongest value signal. Those are not the same sentence.
- Instagram still does not run one ranker. Mosseri’s ranking explainer says Feed, Stories, Explore, Reels, and Search each use their own algorithm.
- On April 30, 2026, PetaPixel quoted Mosseri: if most of what you post is someone else’s content, the account is no longer recommendable to non-followers — now including photos and carousels, not only Reels.
- Copying one clip onto a stack of logins does not pool ranking credit. Each account is scored alone. Shared devices and shared IPs are how one suppressed login becomes a portfolio problem — dedicated real phones and isolated environments (own device, own IP per account) are the ceiling of what we will say about that.
What Does Instagram Count as a View in 2026?
A View is the number Instagram wants you to compare across formats. Sprout Social’s January 28, 2026 algorithm guide says Instagram “has officially transitioned to ‘Views’ as the primary metric across all formats (Reels, Stories, Photos, Carousels), unifying how performance is calculated.” On the same page: “A ‘View’ is counted for every instance a post appears on a screen, including repeat views.”
Help’s sentences for Views, watch time, viewers, average watch time, and follows are on the watch-time vs views page.
That second sentence is the one most dashboards still get wrong. Reach is unique accounts. Views include the person who watched twice, the Story they opened again, and the Reel that looped. If Views climb and Reach does not, you did not find new people — you got replays. That is useful. It is not growth.
| Number on the tile | What the cited pages say it counts | How to misread it |
|---|---|---|
| Views | Sprout Social: every instance a post appears on a screen, including repeat views. Now the primary metric across Reels, Stories, photos, and carousels. | Treating a doubled View count as doubled audience. Loops, replays, and a second Story open all add Views without adding people. |
| Reach | Unique accounts. Sprout does not redefine Reach in the 2026 Views paragraph; the contrast is implicit in “including repeat views.” | Optimising only for Reach and ignoring whether anyone sat with the post. Feed ranking still cares whether you spend time on it. |
| Likes / comments | Still in Mosseri’s Feed five: comment, like, share, profile tap, and spending a few seconds. Hootsuite still lists likes next to watch time and sends. | Chasing like count after Sprout called DM sends the stronger value signal. Likes are not gone. They are not the discovery lever. |
| Sends per reach | Hootsuite (Mosseri): shares or sends sit in the top three for both follower and non-follower reach. Sprout: private DMs are arguably the strongest value signal. | Forcing “comment this and I’ll DM you,” which trains comments. A send is someone putting the post in a friend’s inbox without being asked. |
Instagram monthly active users (Zuckerberg, Sep 24 2025)
CNBC, Sep 24 2025
Primary metric across Reels, Stories, photos, carousels
Sprout Social, Jan 28 2026
Photos or videos now allowed in one carousel post
Sprout Social, Jan 28 2026
Explore prediction: finish the video (watch 95% or more)
Hootsuite, Jul 15 2026
The 3 billion figure is not a 2026 ranking change. CNBC, published September 24, 2025, reported that Meta CEO Mark Zuckerberg said Instagram now has 3 billion monthly active users. The same article says Meta last disclosed Instagram’s user figures in October 2022, when Zuckerberg said the app had crossed 2 billion monthly users, and that in April 2024 Meta stopped disclosing monthly and daily active user numbers for Facebook and sibling apps on a quarterly basis. Older academy copy on this URL still said 2 billion. That was the 2022 figure CNBC names — not a 2026 MAU print.
Views is a reporting change, not a promise of more distribution
Which Ranking Signals Actually Matter Now?
Stop treating “the algorithm” as a single ranking. Adam Mosseri’s Instagram Ranking Explained post says Instagram “doesn’t have a singular algorithm that oversees what people do and don’t see on the app.” Each part of the app — Feed, Stories, Explore, Reels, Search and more — uses its own algorithm. Hootsuite’s 2026 rewrite of that point is the same idea in operator language: Instagram doesn’t run on one algorithm.
The signals people argue about are the ones Mosseri keeps naming in 2026 recaps. Hootsuite (July 15, 2026) says that for both connected reach (people who already follow you) and unconnected reach (people who don’t), “the same signals matter most: watch time, likes, and shares (or sends).” It then tells you to watch engagement rate, average watch time, likes per reach, and sends per reach in Insights. The URL that used to call watch time a solo #1 is updated on the watch-time page.
Sprout Social, linking a Mosseri Reel, lists a different emphasis on the same theme: “Sends Per Reach: Private sharing (DMs) is now arguably the strongest signal of value.” That page also says Instagram “now rewards content that is shared privately (Sends) more than content that is simply Liked.” Hootsuite still puts likes in the top three. Sprout ranks sends above likes. This article cites both. It does not silently pick Sprout’s cleaner headline. The Instagram algorithm tips 2026 page lays the two publisher lists side by side with the Feed/Reels/Stories/Explore surfaces.
Negative signals are in the recap pages too. Hootsuite says Instagram also looks at how often users skip a post, tap “Not interested,” or bounce after a few seconds of viewing, and that the Feed algorithm “limits back-to-back posts from the same creator and blends content from followed and recommended accounts.” Mosseri’s explainer makes the same anti-stacking point: Instagram tries to avoid showing too many posts from the same person in a row, or too many suggested posts back to back. If you are firing five logins into the same viewer’s first screen, you are asking the ranker to do the thing it says it avoids.
Users can wipe the recommendation history. Sprout says Instagram rolled out “Reset Suggested Content,” and that people can go to Settings → Content Preferences to wipe their algorithmic history. For a brand, that is not a growth tactic. It is a reminder that last year’s interest graph is not a contract. You re-earn Explore.
Reels still have an engagement-rate number on Sprout’s page that is easy to steal out of context. The sentence is: Sprout Social’s 2025 Content Benchmarks Report “shows that Reels already achieve a 2.46% engagement rate,” and that Instagram is now rewarding longer storytelling by recommending those Reels in Explore. That 2.46% is a 2025 benchmark on a 2026 algorithm page — not a 2026 ranking weight, and not a promise that your Reel will clear 2.46%.
| Surface | What the cited 2026 pages actually say | What that means for a post |
|---|---|---|
| Feed | Mosseri’s explainer: the five Feed interactions watched most closely are spending a few seconds on a post, commenting, liking, sharing, and tapping the profile photo. Hootsuite adds a more specific Feed prediction: spend more than 10 seconds on the first post. | A skipped first tile is a negative signal. A save-and-share carousel can outrank a like-heavy photo that nobody sits with. |
| Reels | Hootsuite: Reels ranking looks at how long people watch, whether they rewatch or reshare, how quickly they interact, and whether they follow. Sprout lists sends and shares, retention past 3 seconds, and watermark / unoriginal-content scans. | The hook is a ranking input, not a vibe. Sprout’s aggregator-penalty example is reposting TikToks with watermarks. |
| Stories | Mosseri: Stories ranking starts from accounts you follow. Signals include viewing history, engagement history, and closeness. Hootsuite, citing Mosseri, says the top factors are how likely you are to tap into a Story, like it, or reply to it. | Stories do not grow you the way Reels do. They defend the relationship graph the Feed ranker also reads. |
| Explore | Hootsuite: Explore predicts whether you will follow the creator, view the post for at least five seconds, and finish a video (watching 95% or more). Mosseri’s explainer says likes, saves, and shares are the actions Explore predicts most closely, and that popularity signals matter more here than in Feed or Stories. | Explore is a non-follower surface. Originality and completion matter more than your follower count. |
Hootsuite disagrees with itself on Reel length
Carousels are how a still format earns View time. Sprout says carousel limits increased to 20 slides, “up to 20 photos or videos in a single post (double the previous limit),” because every swipe is dwell time. That is the same 20-slide cap our photo and carousel bonuses guide treats as the format Instagram is now willing to pay for — Views are how those bonuses get counted.
What Happens If Most of Your Posts Aren't Original?
Originality is no longer a Reels-only scold. PetaPixel, dated April 30, 2026, reports that Instagram is expanding a policy that already applied to Reels so it covers photos and carousel posts as well. The quoted Mosseri line on that page is specific: “If most of what you post to Instagram is someone else's content, your account is no longer going to be recommendable. That means that we're no longer gonna show your posts to people who don't follow your account proactively.”
That is a recommendability cut for non-followers, not a follower-graph shutdown. PetaPixel says existing followers can still see the account. Discovery — recommendation surfaces like Explore — is what drops. The same article says crediting the original creator, adding a watermark, or making a minor crop is not “original,” and that Mosseri pointed aggregators at three paths: remix the content, use the official repost button, or publish a collab post.
Sprout Social describes the same idea as an “Aggregator Penalty”: Instagram “now actively penalizes ‘aggregator’ accounts that repost content without adding significant value,” and “will often replace your post with the original creator’s version in recommendations.” We are not attaching a made-up “2x original-content boost” to that. The pages we fetched do not print that multiplier.
Your own clip, ten logins, is a different problem
How Should Multi-Account Operators Adapt?
Ranking credit does not pool. Sends-per-reach on account A does not lift account B. Views on a meme page you do not actually create for will not rescue a brand login that never earns a DM. If you are running more than one Instagram account — brands, niches, or promo logins — treat each one as its own ranker with its own originality mix and its own Insights.
- 1
Read Views against Reach, per account
Pull both numbers for the last 14 days. If Views are up and Reach is flat, you are looping existing viewers. That can be a content win. It is not a distribution win. Sprout’s definition (every screen appearance, including repeats) is why those two tiles can diverge. - 2
Track sends per reach, not like count
Hootsuite’s Mosseri recap puts shares/sends in the top three signals for both follower and non-follower reach. Design the post so someone would actually put it in a DM. “Comment X and I’ll send the link” trains comments, not sends. - 3
Audit the last 30 days for other people’s content
Mosseri’s April 2026 line is about what most of the account posts, not one bad reshare. If a login is mostly scraped carousels, it is the account PetaPixel is describing. Remix, collab, or stop asking Explore to save it. - 4
Do not clone one Reel across the stack
Use Trial Reels to test a cut with non-followers on the account that will publish it. Then ship a real variant — different hook, different crop, different caption — not a file copy. - 5
Keep the logins from sharing a device
A suppressed or non-recommendable login on a shared phone and a shared IP is how one bad originality mix becomes five. Dedicated real phones and isolated environments (own device and own IP per account) are the approved description of how we keep that from happening. We will not add a third sentence about infrastructure.
If reach fell off a cliff and you are trying to tell a ranking change from a restriction, start with Instagram’s Account Status tool — Mosseri’s ranking explainer points creators there when content is less visible — and run a diagnosis before you rewrite the content calendar. The free Instagram shadowban checker is the in-body check. If the account is actually suppressed, the 14-Day Recovery System is the recovery path, and the shadowban avoidance guide is the longer version of what not to do next. Ranking advice on a restricted account is noise.
For the rest of the growth system around these signals — pillars, hashtag cap, Reels structure — use the Instagram growth strategy. For why a single login is a fragile place to put the whole business, use the multi-account strategy.
Frequently Asked Questions
What does Instagram count as a View in 2026?
Sprout Social’s January 28, 2026 algorithm guide says Instagram transitioned to Views as the primary metric across Reels, Stories, photos, and carousels, and that a View is counted for every instance a post appears on a screen, including repeat views. Reach still counts unique accounts. A rising View count with flat reach means the same people are seeing the post again, not that you found a new audience.
Do likes still matter, or is it all sends now?
Likes still sit in the ranking mix. Hootsuite’s July 15, 2026 guide, citing Adam Mosseri, says the same three signals matter most for both follower reach and non-follower reach: watch time, likes, and shares (or sends). Sprout Social, also citing Mosseri, calls private DM sends “arguably the strongest signal of value.” We do not pick a winner between those two sentences — they sit next to each other in this article.
What happens if most of what I post is someone else’s content?
PetaPixel published Adam Mosseri’s line on April 30, 2026: if most of what you post to Instagram is someone else’s content, your account is no longer going to be recommendable, meaning Instagram will no longer show those posts proactively to people who do not follow you. The same piece says the policy now covers photos and carousel posts, not only Reels. Mosseri points aggregators at remixing, the official repost button, or a collab post.
How long should a Reel be for the 2026 algorithm?
Two sentences on Hootsuite’s own July 15, 2026 page disagree. A key takeaway says keep Reels short (under 90 seconds). Later, in Mosseri’s Reels basics list, the same page says keep Reels 3 minutes or less. Sprout Social writes that you can post 3-minute tutorials without being penalized if people keep watching. Length is not a ranking cheat code; completion is.
Stop reading one account's Views as a network result
Each Instagram login is its own ranker. Dedicated real phones, isolated environments (own device and own IP per account), years of running accounts at scale, and proprietary sequences are the whole list of how we keep those logins from collapsing into one detection event. If you are still posting the same file from five phones, the Views tile will not save you.