Last updated: September 2026
Key Takeaways
- There is no single Instagram algorithm. Feed, Stories, Explore and Reels each run a separate ranking system, and each one weighs signals differently, which is why the same post can do well in one place and die in another.
- Instagram’s own documentation names four inputs on every surface: your activity, information about the post, information about the person who posted, and your history of interacting with them.
- For video, Adam Mosseri has pointed at watch time, likes relative to reach, and sends relative to reach. Likes tend to move reach among your followers. Sends move reach among people who don’t follow you yet.
- Shadowbanning isn’t something Instagram admits doing, but recommendation ineligibility is real and checkable. Settings, then Account Status, tells you whether your content can be recommended at all.
- Resetting your suggested content changes what you see. It does nothing to how your own posts get ranked. Two different systems, and mixing them up wastes weeks.
- Since December 2025 you get five hashtags per post. Anything after the fifth is ignored, so hashtag volume is no longer a lever you can pull.
Instagram crossed three billion monthly users in September 2025, up from two billion three years earlier.
In the same stretch, average engagement per post went the other way. Socialinsider’s 2026 benchmark study, built from 35 million posts across 447,613 pages, puts the platform average at 0.48 percent, down 24 percent year on year.
More people, less reach each. That gap is why everyone suddenly wants to understand the algorithm.
Here’s the thing though: most of what gets written about it is guesswork dressed up as insight. Instagram publishes a lot about how ranking works, and Mosseri answers questions about it almost weekly. The information is there.
So this guide sticks to what Instagram has actually said, what the numbers actually show, and where the popular advice is flatly wrong.
What Is the Instagram Algorithm?
The Instagram algorithm is a set of separate ranking systems, one for each part of the app, that predict how likely you are to engage with a piece of content and order what you see accordingly. There isn’t one algorithm. There are several, and they disagree.
That distinction matters more than it sounds. Instagram’s own ranking explainer walks through Feed, Stories, Explore and Reels one at a time, because each has a different job.
Feed and Stories mostly show you people you already chose. Explore and Reels mostly show you people you didn’t. A system built to rank your friends cannot use the same weights as a system built to find you strangers.
Which is why “the algorithm changed” is usually the wrong diagnosis. More often, your content shifted from one of these systems to another.
What Signals Does Instagram Actually Rank On?
Four inputs show up on every surface. What changes is the order.
Look at what’s missing from the Stories column. There’s no popularity signal in there at all.
That single absence explains a complaint we hear constantly: someone’s Instagram Stories reach keeps sliding and they assume they’ve been throttled. They haven’t. Story reach is capped by how many of your followers actively watch you, and nothing about a better story fixes a weak viewing relationship this week.
Watch time, likes and sends
For video specifically, Mosseri has been unusually direct. The three signals he keeps naming are watch time, likes measured against reach, and sends measured against reach.
The split between those last two is the useful part. Likes tend to move your reach among people who already follow you. Sends move reach among people who don’t.
Notice that both are ratios. A post seen by 800 people and liked by 80 is doing something a post seen by 40,000 and liked by 400 is not, even though the second number looks better in a screenshot.
This is also why chasing raw follower count stopped working. If you can read your Instagram analytics properly, you already know the ratio is the thing being scored.
Why Did My Instagram Reach Suddenly Drop?
A sudden drop is almost never a mysterious penalty. In order of how often it turns out to be the cause:
- One unusually good post has skewed your baseline. A reel that went wide three weeks ago is still sitting in your average. Everything after it looks like a collapse because you’re comparing to an outlier, not a norm.
- You changed format. Moving from carousels to reels, or reels to stills, moves you into a different ranking system with different weights. Same account, different competition.
- You’ve lost recommendation eligibility. This is the real version of what people call a shadowban, and unlike the rest of this list you can check it directly.
- Platform-wide drift. Socialinsider’s 24 percent year-on-year decline isn’t about you. Some of your drop is just the tide going out.
Before you rebuild your whole content plan, rule out the first two. They cost nothing to check and they’re the answer most of the time.
Is Shadowbanning Real, and How Do You Check?
Instagram’s position is that it doesn’t deliberately suppress accounts. Its ranking explainer says plainly that when it ranks content, hiding you “is not our intention.”
But there’s a real mechanism sitting right next to the myth, and it has a name: recommendation eligibility.
Instagram’s Recommendation Guidelines let content stay on the platform while making it ineligible to be pushed to strangers in Explore, Reels and suggested posts. Your followers still see you. New people stop arriving. From the inside, that feels exactly like being shadowbanned.
The difference is that this one is visible. Go to Settings, then Account, then Account Status. It tells you whether your content can be recommended, shows you which posts are causing the problem, and gives you an appeal route.
Check that screen before you pay a single rupee to a shadowban testing tool. Those tools are guessing from the outside at something Instagram will tell you for free. We’ve covered more of this territory in our piece on Instagram myths and algorithm rumours.
Can You Reset the Instagram Algorithm?
You can reset what Instagram recommends to you. You cannot reset how Instagram ranks you. These get conflated constantly, and the confusion is expensive.
The real feature exists. Go to your profile, tap the menu, then Content preferences, then Reset suggested content. It clears your recommended content across Explore, Reels and Feed, and it offers to walk you through the accounts you follow and your ad topics at the same time.
Two costs worth knowing before you tap it. You can’t undo a reset. And you’re throwing away years of accumulated signal about what you like, so your feed gets noticeably worse for a while before it gets better.
None of that touches your own distribution. There is no creator-side reset button, and anyone selling you one is selling you nothing.
What Actually Improves Your Ranking?
1. Make things people forward to one person
Sends are the signal that reaches strangers. So the question to hold over a draft isn’t “will people like this”, it’s “would someone send this to exactly one friend with no caption”.
That’s a much higher bar than a like, and it changes what you make. Useful, funny, or oddly specific gets forwarded. Polished and general does not.
2. Stop treating hashtags as a growth lever
In December 2025 Instagram capped posts at five hashtags. Add more and the extras are simply ignored.
Mosseri’s framing was that a few specific tags beat a long list of generic ones, and that hashtags help with search rather than reach. Which is a polite way of saying the tactic you built in 2019 is finished.
3. Post things you actually made
Since 30 April 2026, accounts that post someone else’s content without materially changing it 10 or more times in a rolling 30 days lose recommendation eligibility. Eligibility comes back 30 days after the last unoriginal post.
Crediting the creator doesn’t exempt you. Neither does a watermark or a crop. The threshold is meaningful transformation: commentary, a remix, an actual point of view.
4. Test on strangers before you test on your followers
Trial Reels show a reel to non-followers only. You get metrics after roughly 24 hours, and if it performs in the first 72 hours Instagram can push it to your followers automatically.
It’s the closest thing Instagram gives you to a clean experiment, because it isolates the cold audience. You need a professional account to use it.
5. Pick the format on evidence, not vibes
Everyone will tell you reels are all that matters. Socialinsider’s 2026 numbers say carousels edge them out on engagement rate: carousels at 0.55 percent, reels at 0.52, single images at 0.37.
Reels still win on raw reach, because that’s where Instagram spends its distribution. But if the thing you’re optimising is response rather than impressions, carousel posts are being quietly underrated.
6. Skip the hacks
In April 2026 Mosseri put a stake through one of the most repeated ones. Resharing your own feed post to your Stories doesn’t meaningfully lift its reach, because “Feed generally gets more reach than Stories anyway” and a repost of your own post doesn’t change its eligibility.
His broader line is worth taping to your monitor: those tactics “sometimes work, usually don’t, and when we find them, we usually shut them down.”
While we’re here, putting “link in bio” in your caption does not restrict your reach either. That one has survived years past its expiry date.
Does Posting Time Still Matter?
Less than it used to, and not in the way most advice implies.
Recency is a Feed signal, so timing affects how high you place among the accounts someone already follows. It does almost nothing on Explore or the Reels tab, where a post can surface days later.
So if most of your reach comes from recommendations, your posting time is a rounding error. If most of it comes from followers, it’s worth getting right. Our breakdown of the best time to post and the companion piece on how often you should post both go deeper.
FAQs
How do you reset the Instagram algorithm?
Go to your profile, tap the menu, then Content preferences, then Reset suggested content. That wipes your recommendations across Explore, Reels and Feed. It changes what Instagram shows you, and nothing about how your own posts are ranked.
That second sentence is the one people need. Search volume for “how to reset instagram algorithm” is split between two completely different problems: viewers whose feed has drifted somewhere unpleasant, and creators whose reach has fallen. The feature only solves the first. There is no creator-side reset anywhere in the app.
If you are the viewer: the reset can’t be undone, there’s no cap on how often you do it, and Instagram will offer to walk you through unfollowing accounts and editing ad topics in the same flow. Take that offer. A reset with your follow list untouched drifts straight back, because your follows were feeding the recommendations in the first place.
The cost is real. You’re deleting years of learned preference, so expect a genuinely worse feed for one to two weeks while it re-learns. Plenty of people reset, hate the blandness, and conclude they broke something.
Decision rule: reset if your feed has been hijacked by a topic you hate and muting hasn’t fixed it. If your problem is your own reach, close that screen and open Account Status instead.
Is shadowbanning real, and how do I know if I’ve been shadowbanned?
Instagram says it doesn’t deliberately hide accounts, but losing recommendation eligibility is real, and Settings, then Account, then Account Status will tell you whether yours is intact.
What people mean by shadowban is usually this: followers still see you, strangers have vanished. That’s the exact shape of recommendation ineligibility. Instagram’s Recommendation Guidelines allow content to stay up while barring it from Explore, the Reels tab and suggested posts, which cuts off new audience without removing anything.
Account Status shows you a sample of the posts or profile elements causing it, lets you edit or delete them, and lets you request a human review. Set against that, third-party shadowban checkers are inferring from outside what Instagram will show you directly, and several charge for the privilege.
The common failure is diagnosing a shadowban from a two-week reach dip. Normal variance between your best and worst post in a month is routinely wider than that, so most “shadowbans” are just a quiet fortnight.
Decision rule: open Account Status first. Green light means your problem is the content, not a penalty, and no amount of appealing will change it.
Does Instagram punish you for deleting a post and reposting it?
There’s no confirmed penalty for deleting a post, but reposting the same thing rarely gets you a better result, and if reposting other people’s work is your pattern there is now a documented penalty.
Separate the two cases, because they get blended into one anxiety. Deleting your own underperformer and putting it back up is a strategy question. Reposting content you didn’t make is a policy question.
On the strategy side, Mosseri has said a repost of your own content doesn’t change its eligibility, since it was already posted. You’re spending a slot in your own schedule to show the same thing to a similar audience. Occasionally that catches a different moment. Usually it doesn’t.
On the policy side the numbers are specific. Since 30 April 2026, 10 or more posts in a rolling 30 days of content you didn’t create or materially enhance makes your account ineligible for recommendations, and you stay ineligible until 30 days after the last one. Credit lines, watermarks and small crops don’t count as enhancement.
Decision rule: delete a post only if it’s wrong, embarrassing or off-brand. If it merely flopped, leave it and make the next one better, because the repost costs you a slot and buys you almost nothing.
Does the Instagram algorithm favour Reels over carousels and photos?
Reels get more distribution, but carousels get more engagement per post, so the honest answer depends on which one you’re actually trying to buy.
The question behind this is usually a resource question. Reels cost several times what a carousel costs to make, and people want to know whether that’s a tax they have to pay.
Socialinsider’s 2026 study of 35 million posts puts carousels at 0.55 percent engagement, reels at 0.52 and single images at 0.37. That’s close enough between the first two that “reels win” is not supportable as a general claim. Where reels do win is raw reach, because the Reels tab is a recommendation surface pointed at strangers and Feed largely isn’t.
The failure mode is predictable. A team with a carousel-shaped audience switches to reels for algorithmic reasons, watches impressions climb and saves collapse, and can’t work out why the numbers look better while the business doesn’t.
Decision rule: reels when you need new people, carousels when you need the people you have to act. If you can only maintain one format properly, pick the one you can sustain weekly, because consistency beats format every time.
Why did my Instagram reach suddenly drop?
Most sudden drops are one of four things: an outlier post skewing your baseline, a change of format, lost recommendation eligibility, or the platform-wide decline that’s hitting everyone.
Start with the baseline, because it’s the most common and the least dramatic. One reel that travelled three weeks ago is still inside your 30-day average. Every normal post after it reads as a crash when you’re really just comparing against your best day.
Then format. Moving between stills, carousels and reels moves you between ranking systems with different weights and different competition. Same account, different race.
Then eligibility, via Account Status. Then accept that some of it isn’t yours to fix: platform-average engagement fell 24 percent year on year in Socialinsider’s 2026 data, so a slow slide with no other explanation is probably the tide.
The expensive mistake is responding by posting more. Volume at lower quality pushes your ratios down, and ratios are what’s being scored, so the fix makes the problem worse.
Decision rule: compare your median post to last quarter’s median, never your average to your best. If the median is flat, nothing is wrong.
How many times a day should you post on Instagram?
Once a day is plenty for almost everyone, and most accounts that post more than twice a day are diluting their own ratios rather than buying extra reach.
The real worry underneath this question is usually the opposite one: people are afraid that posting too little gets them punished. It doesn’t. There’s no penalty for a quiet week, and no consistency bonus you can claim by filling slots.
What the observed data shows is moderate. Socialinsider found accounts averaging roughly 8 reels, 7 images and 5 carousels a month across 2025. That’s around two-thirds of a post a day in total, from a set of 447,613 pages.
The mechanism that punishes volume is the ratio. Because likes and sends are scored against reach rather than counted raw, a fourth mediocre post doesn’t add reach so much as it lowers your average performance. Do that for a month and your whole account looks weaker to the ranking system.
Decision rule: pick the highest frequency you can hold at full quality for three months, then stop there. If you can’t tell which of today’s two posts is the better one, you only had one post.
Does Instagram punish you for scheduling posts?
No penalty for scheduling has ever been confirmed by Instagram, and the platform ships its own scheduling tools, including scheduling for Trial Reels.
This fear is a holdover from the early 2010s, when third-party tools posted by automating a logged-in session and Instagram genuinely did fight them. That era ended when the official publishing API arrived. Scheduled posts now go through Instagram’s own pipes.
It’s worth being precise about what’s still true. Instagram does act against automation that fakes engagement, follows and unfollows at scale, or operates accounts in bulk. Publishing a post at a chosen time is not in that category.
The genuine cost of scheduling isn’t algorithmic, it’s editorial. Scheduled content is written a fortnight out, so it can’t react, and reactive content is what gets forwarded. Accounts that schedule everything tend to drift into a tone that’s technically fine and completely inert.
Decision rule: schedule your baseline so you never go dark, and keep at least one slot a week open for something you make that day. If your calendar is 100 percent pre-built, your sends will show it.
Does the Instagram algorithm favour new or small accounts?
Small accounts aren’t favoured, but they’re not disadvantaged the way people assume either, because follower count sits near the bottom of Instagram’s stated signals on every surface.
Look back at the Reels column in the diagram above. Information about the person who posted is fourth, behind your activity, your interaction history and the reel itself. On a recommendation surface, Instagram is mostly asking whether this specific piece of content works, not who made it.
That’s genuinely better for small accounts than the old follower-graph model, and it cuts the other way too. A large account posting something mediocre now gets a mediocre result, which is why so many established brands feel like their reach broke.
Where small accounts really do struggle is the ratio trap. With 300 followers your reach is small enough that a handful of loyal likes produces a flattering rate, and then a slightly wider distribution tanks it. Early numbers are noisy, and reading strategy into them will send you in circles.
Decision rule: under about 1,000 followers, ignore engagement rate entirely and judge yourself on sends and saves in absolute numbers. Ratios only start telling the truth once your reach is stable enough to compare.
Conclusion
Here’s the uncomfortable summary. Almost every algorithm tactic that worked five years ago is either capped, ignored or actively shut down, and what replaced it is harder: make something a person would forward to one friend.
That’s not a hack you can implement on a Tuesday. It’s a content problem.
The good news is that the system is more legible than it’s ever been. Instagram publishes its signals, Account Status tells you your standing, and Trial Reels lets you test on strangers before you risk your followers. That’s more transparency than the people still hunting for shadowbans are using.
If you’d rather not spend the next six months working this out by trial and error, that’s what we do. Talk to AB Media about your Instagram strategy, and read up on where Instagram is heading next and what growing a following actually takes while you’re deciding.
