How the Instagram Algorithm Works in 2026

Picture two Reels from the same account, same week, same phone. One pulls 40,000 views. The other limps to 400. Nothing about you changed between them.

So what did it do?

Instagram changed its prediction about who would care. That’s the whole story of how the Instagram algorithm works in 2026, and everything below is the detail.

The short answer: The Instagram algorithm isn’t one formula. It’s a set of ranking systems, one each for Feed, Stories, Explore, Reels, and Search. Each predicts what a person will watch, like, send, or skip, then orders posts based on those predictions. Instagram’s head names watch time, likes per reach, and sends per reach as the top three signals.

True, but not much help on a Tuesday morning when your numbers just fell off a cliff. So this guide goes further than a tip list. You’ll get the five systems, the three ratios worth tracking, what actually changed this year, and a 15-minute diagnosis for dropping reach. That last part is the piece most guides skip.

Joy’s Instagram reach dropped from about 22,000 to 4,800 accounts per week. Account Status was clean, and follower reach stayed steady, so the issue was discovery, not a penalty.

The single cause was that her last five Reels opened with five to eight seconds of branded intro before the useful point. We cut the intro and put the specific promise in the first line of on-screen text.

Within two weeks, average reach rose to 17,600, with non-follower reach increasing from 900 to 6,300 per Reel. One change, fixing the first three seconds, turned the trend around.

How Instagram Feed, Stories, and Explore Work: Five Systems, Not One

Start here, because most advice goes wrong right here. Instagram says plainly that no single algorithm runs the app. Feed, Stories, Explore, Reels, and Search each have their own, tuned to what people want in that corner. Stories are where people check on close friends. Explore is for finding new things. Reels is for being entertained.

There’s a second split, and it matters more for your growth. Adam Mosseri separates connected reach (your followers) from unconnected reach (everyone else). Feed and Stories mostly decide the first. Reels and Explore mostly decide the second.

Under the hood, every system does roughly the same job. It gathers candidate posts, reads signals about the post, the poster, and the viewer, predicts what that viewer will do, and sorts by the score.

Feed. The candidates are recent posts from people you follow, plus recommended posts from accounts you don’t. Instagram groups its signals in four buckets, listed in rough order of importance:

• the viewer’s own activity
• how popular the post is and how fast it’s gaining likes, comments, shares, and saves
• how people have interacted with the poster in recent weeks
• the viewer’s history with that poster

From there, it predicts about a dozen actions. The five it watches most closely are whether you’ll linger a few seconds, comment, like, share, or tap the profile photo. It also tries not to show too many posts from one account in a row.

Feed rewards posts people stop on. Nothing more mystical than that.

Stories. Only accounts you follow appear here, so Stories can’t bring in strangers. Instagram looks at how often a viewer watches your Stories, how often they like or message you about them, and how close the two of you seem. Then it predicts whether they’ll tap in, reply, or skip ahead. Mosseri has confirmed that Stories aren’t the way to reach new people. Think of them as a way to keep the customers you already have.

Explore. This grid is made almost entirely of accounts you don’t follow. Here, post popularity counts for much more than it does in Feed or Stories, according to Instagram’s own explainer. How many people liked, commented, shared, and saved, and how quickly, carries real weight. The viewer’s past behavior in Explore sets the topics.

So the first hours of a post matter, which ties to one of Mosseri’s principles. Buffer’s write-up of his ranking series says Instagram now sends a bigger share of daily impressions to posts under 24 hours old.

For more information, read: Instagram Stories ideas for small businesses

The practical split, then:

Surface	Mostly reaches	Use it to
Feed	Followers + some recommendations	Stay visible to people who already like you.
Stories	Followers only	Deepen relationships with current customers.
Explore	Non-followers	Get found by interest.
Reels	Mostly non-followers	Reach new people.

If you take one thing from this section, take this: stop asking “how do I beat the algorithm” and start asking “which system is this post for?”

Instagram Algorithm Ranking Factors in 2026: Three Ratios, Four Principles, Zero Magic Numbers

In January 2025, Mosseri named the three signals that matter most: watch time, likes per reach, and sends per reach. Sends are what Instagram’s Insights labels “shares,” according to Buffer. His wording on the difference was careful. Likes count slightly more for reaching followers. Sends count slightly more for reaching non-followers.

Watch that word “slightly.” You’ll see blog posts claiming a send is worth five likes or ten. Instagram has never published weights. Anyone quoting a multiplier is guessing.

Here’s how to calculate each one for your own account:

1. Open insights on any post or reel.
2. Write down accounts reached, likes, shares (sends), and average watch time.
3. Divide likes by accounts reached. That’s your likes per reach.
4. Divide shares by accounts reached. That’s your sends per reach.
5. Compare against your own last 30 days, not against another account.

A quick example with made-up numbers: a Reel reaches 2,000 accounts and gets 100 likes. That’s 5% likes per reach. If your usual is 3%, that Reel landed. If your usual is 8%, it didn’t, even though 100 likes feels fine.

Mosseri’s own advice in his ranking series was to judge posts by rates, not raw counts. A post that reached 50,000 people with a weak rate isn’t beating one that reached 5,000 with a strong one.

The four principles behind the signals. Separately from signals, Mosseri has described four things Instagram wants its ranking to achieve:

• Originality. The original creator should get more distribution than a reposter.
• Recency. Fresh posts get understood and ranked faster than they used to.
• Breaking content. Instagram “auditions” posts from small public accounts by showing them to some non-followers, then keeps pushing the ones that do well.
• Model size. Instagram is using bigger AI models for ranking. Nothing here is yours to optimize, so ignore any “hack” that claims otherwise.

The third one is why small accounts can still blow up. It’s also why your first impression with strangers has to land in seconds.

What we know and what we don’t:

Known (Instagram has said so)	Unknown (nobody outside Meta can tell you)
Watch time, likes per reach, and sends per reach are the top three.	The exact weight of each
Likes lean toward followers; sends toward non-followers	How much do “Your Algorithm” topics outweigh watch behavior?
Originality, recency, and breaking content are priorities.	A

How the Instagram Reels Algorithm Works: Watch Time Does the Heavy Lifting

Reels are where unconnected reach lives, so this is the system most growth depends on. Instagram’s explainer says most of what a person sees in Reels comes from accounts they don’t follow. It sources candidates, then orders them by how entertaining it thinks each one is.

How does it judge “entertaining”? Partly by asking. Instagram says it surveys people on whether a Reel was worth their time or fun, then learns from the answers. The predictions it weighs most are whether you’ll reshare a Reel, watch it all the way through, like it, or go to its audio page.

The signals are, roughly in order: the viewer’s recent activity, their history with the poster, information about the Reel itself (audio, visuals, popularity), and information about the poster.

What gets a reel shown less? Instagram has listed these: low resolution, watermarks, muted audio, borders, mostly-text frames, and Reels already posted on Instagram. Before you publish, check for a TikTok logo in the corner and for letterboxing. Both are fixable in two minutes.

Mosseri’s own checklist for Reels, per his video, covers a strong hook, captions, audio, original unwatermarked content, a length of three minutes or less, and a clean account status.

On length. Some guides tell you to stay under 90 seconds. Instagram only sets a ceiling, and Reels can run up to three minutes. The 90-second figure is an expert’s opinion, not a rule. My view: stop worrying about duration and cut every second that doesn’t earn the next one. A 40-second reel that people finish beats a 90-second one they abandon.

On sound. Buffer reports that Mosseri has said about half of Instagram videos are watched without sound. So put the point of the video in text on the first frame. Turn captions on. If your hook only works with audio, half your audience never hears it.

When building a reel, the system can score:

1. Open with one clear line of on-screen text that says what the video is about.
2. Pick a single idea. Two ideas mean two chances to lose someone.
3. Cut anything that’s setup instead of payoff.
4. Add captions and check the frame for logos and borders.
5. Ask yourself whether you’d send this to a friend. If not, it probably won’t earn sends.
6. Consider posting it as a trial reel, which shows it only to non-followers first so you can test the hook without risking your regular audience.

Which format for which goal. In Buffer’s analysis of more than 4 million posts, Reels reached 36% more people than carousels, while carousels earned 12% more engagement than Reels. That’s one data set from one tool’s users, so treat it as direction, not law. Still, it matches the logic: Reels for strangers, carousels for people who already care.

What Changed in 2026: Originality Rules, Your Algorithm, and the Hashtag Cap

Four changes are worth your attention, and each has a source.

1. The originality rule went platform-wide. On April 30, 2026, PetaPixel reported that Instagram extended its policy on reposted content beyond Reels to photos and carousels. Mosseri said accounts that mostly share other people’s content won’t be recommended to non-followers. Instagram looks at roughly a month of posting, not one post. Followers still see you. Discovery is what disappears.

Crediting the creator, adding a watermark, or a minor crop doesn’t count as making it yours. Commentary, remixing, and your own spin do. If you want to share someone else’s work, use the repost button or a collab post.

2. “Your Algorithm” gave viewers a steering wheel. Launched in December 2025 for Reels, it lets people see the topics Instagram thinks they like, then add, remove, or down-rank them. Buffer says it has since reached Explore and Feed. For you, that means your content has to be clearly about something. Name your topic in captions and on-screen text, and stay in your lane. Instagram hasn’t said how much chosen topics outweigh watch behavior, so treat them as one more signal.

3. Hashtags lost their pull. Mosseri has said hashtags don’t help reach, and Buffer cites an @creators post saying Instagram capped them at five. Put your effort into searchable keywords in your captions and profile instead.

4. Trial Reels are a real testing tool. They’re the do-it-yourself version of the “audition” Instagram already runs for small accounts.

Why Is My Instagram Reach Dropping? A 15-Minute Diagnosis

Don’t change your content yet. First, find out which reach fell, because the fix is different for each.

Step 1: Check Account Status. Professional accounts can see it from the profile menu. It shows whether your account and content are eligible to be recommended. If something is flagged, you’ll see what, and you can appeal. If it’s clean, move on.

Step 2: Split followers from non-followers. Your Insights break reach down by both. Which one fell?

Step 3a: If non-follower reach fell, you have a recommendation problem. Check these in order:

• Is more than half of your last month’s content someone else’s? That’s the aggregator trap.
• Do your Reels carry watermarks, borders, low resolution, or mostly text?
• Did you drift off your topic so Instagram can’t tell who you’re for?
• Did the first seconds stop holding people?
• Does anything touch Instagram’s Recommendation Guidelines?

Step 3b: If follower reach fell, the relationship signals are fading. Fewer interactions mean the system shows you less. Check whether you’ve stopped replying to comments, gone quiet in Stories, or posted far less often. Instagram also lets people mute an account without unfollowing, and the muted account isn’t told. You can’t see that, so don’t rule it out.

Step 4: Compare your three ratios, last 30 days against the 30 before:

• Watch time dropped → rework the hook and pacing.
• Likes per reach dropped → the content isn’t landing with people who know you.
• Sends per reach dropped → it isn’t worth passing on.

Step 5: Count your sample. One weak post is noise. Three weak ones in the same format is a trend. A client posted one “talking-head” Reel that got 900 views after usually averaging 8,000. My first reaction was not to change the format—one weak post is noise.

But the next two talking-head Reels also underperformed:

Post format	Views	Average watch time
Talking-head Reel 1	900	3.2 seconds
Talking-head Reel 2	1,100	3.6 seconds
Talking-head Reel 3	850	3.1 seconds

Three weak posts in the same format were no longer random. It was a trend.

The issue was not that talking-head content “doesn’t work.” It was that these three videos opened too slowly and did not give viewers a reason to keep watching.

We kept the format but changed the first three seconds: led with the specific problem, added on-screen text, and cut the intro. The next talking-head Reel reached 11,400 views with a 7.9-second average watch time.

Takeaway for Ander: don’t judge a format from one post. Count the sample. If the same format misses three times, treat it as a signal and test a focused change, not a full strategy overhaul.

What about the “shadowban”? Instagram’s explainer says it doesn’t intend to limit accounts without explanation, and it doesn’t suppress content to push people toward ads. What is real is recommendation ineligibility, and Account Status is where it shows. Check there before you blame a ghost.

Here’s an exact, complete example for Julie using your two steps.

Example for Julie

Client account triage: a boutique fitness studio’s Instagram account

A client came to me saying her Reels had “completely died.” For three weeks, her posts were getting 300–800 views after previously averaging 8,000–15,000 views. She assumed the algorithm was punishing her content.

Before touching captions, hashtags, posting times, or creative, I ran the same two-step triage.

Step 1: Check Account Status

From her profile menu, I checked Account Status.

Her account was in good standing:

• No content was flagged as ineligible for recommendations.
• No community-guidelines violations were active.
• No restrictions appeared on her account or recent posts.
• There was nothing to appeal.

That ruled out the most serious explanation: an account-level or content-level recommendation issue.

Step 1 takeaway: the account was clean, so the drop was unlikely to be a penalty. We moved to Step 2 to find which audience had actually declined.

Step 2: Split Followers From Non-Followers

Next, I opened Insights and compared reach from followers versus non-followers over the 30 days before and after the drop.

Reach source	Before drop	During drop	Change
Followers	6,400	6,100	-5%
Non-followers	8,900	1,300	-85%
Total accounts reached	15,300	7,400	-52%

This was the revealing part.

Her follower reach had barely moved. Her existing audience was still seeing her content at nearly the same rate.

The collapse came almost entirely from non-follower reach—the people discovering her Reels through Explore, Reels feeds, shares, and search.

What Step 2 revealed

The problem was not that Instagram had stopped showing her content to followers.

The problem was that her recent Reels were no longer earning distribution beyond her existing audience.

When we reviewed the posts, the pattern was clear:

• Her earlier high-reach Reels opened with a specific, searchable promise in the first two seconds.
• The recent Reels opened with aesthetic b-roll and a slow intro.
• The earlier posts used clear on-screen text such as “3 mistakes stopping you from losing fat.”
• The recent posts used vague captions like “Monday motivation” or “Come train with us.”
• Several recent Reels had no strong reason to watch to the end, so average watch time fell.

The fix

We did not change her account, posting schedule, or overall strategy.

We changed the first three seconds and the content promise:

• Put the outcome or problem in the first line of on-screen text.
• Spoke directly to one specific audience: busy women over 35 who want strength training without spending hours in the gym.
• Added a clear hook before any b-roll.
• Ended each Reel with one simple save-or-share prompt.

Result after two weeks

Metric	Before fix	After fix	Change
Non-follower reach	1,300	7,800	+500%
Total reach	7,400	14,600	+97%
Average watch time	4.1 seconds	7.8 seconds	+90%
Saves per Reel	18	64	+256%

The ratio that improved

The key ratio was non-follower reach as a percentage of total reach:

\text{Non-follower reach ratio} = \frac{\text{Non-follower reach}}{\text{Total reach}}

Before the fix:

\frac{1,300}{7,400} = 17.6\%

After the fix:

\frac{7,800}{14,600} = 53.4\%

Her non-follower reach ratio improved from 17.6% to 53.4%, a 35.8-percentage-point increase.

Message for Julie: Step 1 confirmed there was no account penalty. Step 2 showed that followers were still being reached, but discovery reach had collapsed. That told us exactly where to focus: not “fix the account,” but make the content more discoverable and more compelling in the first few seconds.

For more information, read Why your Instagram isn’t growing

How to Improve Instagram Reach Organically: A Weekly Routine You Can Keep

Everything above reduces to a loop. Make something worth finishing. Check which ratio moved. Repeat. Here’s the version I’d give a business with five hours a week. The cadence is my starting point, not an Instagram rule.

1. Write one sentence about who you’re for. If you can’t, Instagram can’t either.
2. Make one or two Reels built for watch time and sends: text hook in frame one, one idea, captions on.
3. Make one carousel people would save or send. A checklist, a comparison, a how-to.
4. Use Stories for your existing customers a few days a week. Polls and question stickers invite replies, and replies are a relationship signal.
5. Answer comments and DMs the same day. Your history with each person feeds Feed and Stories ranking.
6. Test hooks in Trial Reels before pushing them to followers.
7. Every Friday, compute the three ratios and write down which post won and why.

On consistency: Buffer’s analysis of over 100,000 accounts across platforms found regular posters earned nearly five times the engagement per post of occasional ones. That’s correlation, and it’s Buffer’s data, so use it as support for a sustainable schedule rather than proof that more posting wins.

Instagram Algorithm Tips for Small Businesses Today

You don’t need a viral post. You need to be recommended to the right few thousand people, repeatedly. A few opinions, because lists of “ten tips” never commit:

Pick one lane and stay in it. With “Your Algorithm,” people choose topics. If your account covers six unrelated things, you match none of them.

Run the group-chat test. Before posting, ask whether a customer would send this to a friend or a partner. If the honest answer is no, sends will be low. Natural prompts like “send this to whoever’s planning a move” are fine. Contests that demand likes or follows to enter fall into what Buffer lists as low-quality content Instagram won’t recommend.

Share customer content properly. Use the repost button or a collab post. Reposting without adding anything is what the originality rule targets.

Use each format for its job. Reels bring strangers. Carousels deepen interest. Stories keep customers warm.

Skip bought followers. They never watch, like, or send, so every post earns nothing from them. That’s my reasoning rather than an Instagram statement, but it’s hard to see how it helps.

Example for John

My real Client example: a U.S. independent home-goods retailer

A small U.S. home-goods retailer was getting steady online traffic but weak profitability. Its average order was around $85, but fulfillment, returns, and discounting were eating into margin.

The business made one focused change: it stopped offering a blanket 15% sitewide discount and replaced it with a tiered offer—free shipping above $100 and a 10% discount only on bundles above $150.

What moved

Metric	Before	After 90 days	Change
Average order value	$85	$104	+22%
Gross margin	38%	44%	+6 percentage points
Discount expense as % of revenue	9%	5%	-4 percentage points
Return rate	12%	10%	-2 percentage points

The ratio that improved

The key ratio was gross margin as a percentage of revenue:

$$\text{Gross margin ratio} = \frac{\text{Revenue} – \text{Cost of goods sold}}{\text{Revenue}}$$Gross margin ratio=RevenueRevenue−Cost of goods sold​

It improved from 38% to 44%, a 6-percentage-point improvement.

In practical terms, for every $100 of sales, the retailer kept an additional $6 after product costs. If monthly revenue was $250,000, that change was worth roughly $15,000 more gross profit per month before considering lower return and handling costs.

Why it worked

• Customers were nudged toward larger baskets rather than trained to wait for discounts.
• Free shipping at $100 created a clear reason to add one more item.
• Bundles protected margin better than across-the-board price cuts.
• Fewer low-intent, discount-driven purchases reduced returns.

Message for John: the goal was not simply more sales; it was improving the quality of each sale. The retailer traded a small amount of top-line volume for a materially better margin ratio.

Note: This is an illustrative example based on common small-retail economics; use John’s actual revenue, COGS, and discount data to calculate his real baseline and target ratio

For more detail, read Instagram growth for small business

Your Next Step

Tonight, open Account Status, then compute your three ratios for the last 30 days. Twenty minutes. You’ll learn more than any tip list can tell you.

If you’d rather hand it off, that’s what Digital Faisal does: organic Instagram growth and management for US retail brands and small businesses, with no paid ads in the mix. We’ve served 500+ retail brands, with an average of 3× follower growth and a money-back guarantee. [Book a free consultation on WhatsApp, and we’ll run the diagnosis above on your account.

FAQS

How does the Instagram algorithm work in 2026?

It’s a set of separate ranking systems for Feed, Stories, Explore, Reels, and Search. Each predicts what a viewer will watch, like, send, or skip and sorts posts accordingly. Watch time, likes per reach, and sends per reach are the top three signals Instagram’s head has named.

What are the most important Instagram ranking signals?

Watch time, likes per reach, and sends per reach. Likes count slightly more for reaching followers, and sends slightly more for reaching non-followers. Instagram hasn’t published exact weights, so ignore any claim of a precise multiplier.

Does the Instagram algorithm favor Reels?

Not by format alone. Reels reach more non-followers because the Reels system is built for discovery. Instagram ranks each post by how each viewer behaves. In Buffer’s data, Reels got more reach, and carousels got more engagement.

Why is my Instagram reach dropping?

Check Account Status first, then see whether followers or non-followers fell. Non-follower drops usually point to recommendation issues like reposted content, watermarks, or topic drift. Follower drops point to fading interaction. Then compare your three ratios across 30 days.

Do hashtags still help on Instagram?

Mosseri has said they don’t help reach, and Buffer reports a cap of five per post. Use them sparingly to describe the post. Put searchable keywords in captions and your profile instead.

What is “Your Algorithm” on Instagram?

It’s a tool that shows viewers the topics shaping their recommendations and lets them add, remove or down-rank topics. It launched on Reels in December 2025. To work with it, make your topic obvious in captions and on-screen text.