Stories is not Feed, and almost nobody researches it

You spent an hour on a reel. You used the trending audio. You posted it at the time someone said works. It got two thousand views. Then you reposted it to your story. Sixteen people watched it. You look at the numbers side by side. They do not make sense.

So you go looking for what Stories wants. You find guides about Reels. Articles about Feed. Videos about Explore. The advice is all the same. Post at 6pm. Use trending sounds. Get engagement in the first hour. You try it. Your story still dies at a few dozen views while your reels sometimes do thousands.

Here is what nobody says. Stories is not Feed. It is not Reels. It is not Explore. It is its own surface. It has its own ranking system. Its own signals. Its own AI system card that Meta publishes. And almost no one has read it.

We read it. We also read the cards for Feed, Explore and Reels. Then we read what Adam Mosseri wrote about how Stories works. He is the head of Instagram. The cards sit on Meta's Transparency Center. That is a website where Meta publishes pages about how its software decides what to show you. Each page covers one part of Instagram. Meta calls them system cards. Every one of them is linked, dated and quoted below.

Here is what the documents say. Stories ranks by relationship. Feed ranks by relevance. The two are not the same thing. Stories mostly shows you to people who already follow you. Feed and Explore try to find you new people. On those surfaces a post starts cold, and Instagram's own pages describe a new post as an item the system knows nothing about yet. The signals that make a story rank high are about you and the viewer. How often they watch your stories. How often they reply. Whether the software thinks you are family. Those signals do nothing for a reel in Feed.

Every guide treats Instagram like one algorithm. It is not. It is nine surfaces. Each has its own card. Each ranks by different rules. And the surface creators use most is the one almost nobody writes about.

It is nine surfaces, not one algorithm

Start with the mistake everyone makes. Someone says "the algorithm" and you picture one thing. One system. One set of rules. One opinion about your content. That is not how Instagram is built.

Meta publishes a page for each part of the app. Feed has its own page. So do Stories, Explore, Reels, Search, and four others. Ten pages total. Nine of them are Instagram surfaces. The tenth is Threads. Each surface has its own AI system. Each system uses its own signals. Each makes its own guesses about you.

A surface is one part of the app. The word comes from Meta's own writing. Your feed is a surface. The Explore tab is another. So are Stories, the Reels tab, and Search. Each one shows you different things. Each one ranks those things its own way.

Adam Mosseri said it plainest in May 2023:

Instagram doesn't have a singular algorithm that oversees what people do and don't see on the app. We use a variety of algorithms, classifiers, and processes, each with its own purpose. — Adam Mosseri, Instagram Ranking Explained, 31 May 2023

A classifier is software that sorts a post into a category. Like news, or a how-to. An algorithm is a set of steps software follows. Here it means the ranking system. The ranking system scores posts. Then it puts them in order for you.

Two years later Meta's engineers put a number on "a variety". Their post is called Journey to 1000 models. It ran on 21 May 2025. ML in it is short for machine learning. That means software that learns from examples instead of being told every rule. The post says Instagram has scaled to "over 1000 ML models".

So there is no single thing to please. There is no single place where your content wins or loses. A story that sixteen people saw was ranked by one system. A reel that two thousand people saw was ranked by a different system. When someone tells you what "the algorithm" wants, ask which of the nine they mean.

Feed and Explore run a funnel. Stories does not.

Three surfaces publish numbers. Feed, Explore and Reels. The numbers show the shape of the work. Every step costs money to run. So the system starts wide. Then it gets narrow. Cheap steps look at many posts. Expensive steps look at few. Meta's engineers call that a funnel.

Take Feed first. Its system card was updated 29 June 2026. The card describes four steps. The system gathers posts from accounts you follow. It drops anything that breaks the rules. Then a cheap model picks about 700 posts. Then comes the expensive step:

Finally, the system calculates a relevance score for about 500 posts and puts them in order by this score. — Meta Transparency Center, Instagram Feed AI system card, updated 29 June 2026

About 700 posts in. About 500 scored. That is for one person. When you open the app. Then it starts over for the next person.

Explore is wider. Its card was updated 22 June 2026. It names three steps. The first is retrieval. That means gathering posts worth a look. The card says "up to 1500 media are fetched". Media is the card's word for posts. That is a ceiling. Then a cheap model scores them and passes the best 100 forward. Only those 100 get the expensive model.

Up to 1500 in. 100 out. For one person. Every time the tab loads.

Reels Chaining is the narrowest. It picks about 100 reels. About 100 get scored. That card was updated 11 November 2025. Chaining is Meta's word for the run of reels that keeps arriving once you open one.

Stories does none of that. Its card was updated 18 September 2025. It describes five steps. None of them narrow a big pool down to a small one. No 1500. No 700. No 100. The card gives no funnel counts at all.

Here is why. Stories only ranks content from accounts you already follow. The pool is not billions of posts. It is however many of your follows posted a story in the last 24 hours. That might be ten people. It might be fifty. It is never 1500. So there is no expensive narrowing to do. The system just scores them and puts them in order.

That one difference changes everything. Feed is trying to pick 500 posts out of billions for you. Stories is trying to order the twenty people you follow who posted today. Different problem. Different system.

Stories guesses your relationship. Feed guesses your interests.

Every ranking system makes guesses. Meta prints what those guesses are. The guess list is different for every surface. Here are four of them. Read them one after the other. Four different bundles. No two are the same. Source for all four: Meta Transparency Center system cards for Feed, Stories, Explore and Reels Chaining, retrieved 17 August 2026.

• Feed predicts:
  • How likely you are to skip the post.
  • How much time you will spend on the author's profile after seeing the post.
  • How likely you are to share the post.
  • How much time you will spend on the post.
  • How likely you are to swipe through a carousel.
  • And others.
• Stories predicts:
  • How likely you are to click on a story at the top of your home page.
  • How likely you are to interact with responses to a question sticker.
  • How likely it is that you and the story's author are family or good friends.
  • How likely you are to reply to the story.
  • How likely you are to send a like.
  • How likely you are to tap through to the author's profile.
  • How likely you are to save a Reels effect.
  • How likely you are to swipe to the next story.
  • How likely you are to exit the current story.
  • How likely you are to interact with a poll or a slider.
• Explore predicts:
  • How likely you are to follow the author.
  • How likely you are to spend five seconds or more viewing the post.
  • How likely you are to watch 95% or more of the video.
  • How likely you are to click "Not Interested".
  • How likely you are to comment.
  • How likely you are to like the post.
  • How likely you are to reshare it.
  • How likely you are to click through and engage.
  • How likely you are to save it.
  • How likely you are to open it full screen.
• Reels Chaining predicts:
  • How likely you are to reuse the audio.
  • How likely you are to dismiss the reel in under three seconds.
  • How likely you are to click "Interested".
  • How likely you are to comment.
  • How likely your watch time lands above the 95th percentile.
  • How likely you are to click from Explore into Reels.
  • How likely you are to reshare it.
  • How likely you are to share it outside Instagram.
  • How likely you are to follow the author.

Four separate AI systems. Each one guessing something different about you.

Look at the Stories list. Three guesses stand out. One is about the viewer. How likely they are to click on your story at the top of their home page. Another is about action. How likely they are to reply, or send a like, or swipe to the next story. The third is about relationship. The card's exact words:

How likely it is that you and the story's author are family or good friends — Meta Transparency Center, Instagram Stories AI system card, updated 18 September 2025

That guess exists nowhere else. Not in Feed. Not in Explore. Not in Reels. Only Stories asks whether you are family.

Now look at Feed. Its guesses are about relevance. Will you skip the post. How much time will you spend on it. Will you share it. Feed is trying to guess whether this post is interesting to you right now. Stories is trying to guess whether you care about this person.

The signals show the same split. A signal is one fact about a post or a viewer that the software is allowed to read. Stories reads signals Feed does not. Adam Mosseri listed the most important ones for Stories in May 2023. They are in rough order of importance:

1. Viewing history. How often you view an account's stories.
2. Engagement history. How often you engage with that account's stories. Like sending a like or a message.
3. Closeness. Your relationship with the author overall. How likely you are to be connected as friends or family.

The Stories system card adds more. Number of times you have viewed stories from this account. Number of messages exchanged with the author. Whether you are connected on Facebook. Total stories in the author's collection. How many of their stories you have not seen yet.

Feed reads different signals. Its card lists these among others. How recently the post was created, within 90 days. Whether the post is news, a product review, a tutorial or a how-to. How many people liked, commented on, shared or saved the post. Information about the post, like whether it is a photo, video or carousel.

Feed wants to know what the post is and whether other people liked it. Stories wants to know who made it and how often you watch them. The two systems are solving different problems.

Stories is the only surface that mostly reaches your followers

Here is the line that changes how you read every guide. Stories shows content from people you follow. That is its entire pool. Mosseri's 2023 post says it plainly. Stories displays content "from people you've chosen to follow, as well as ads".

Feed is different. Feed carries two kinds of post. Ones from accounts you follow. And ones from accounts you do not. Meta splits those into two cards. The Feed card covers posts from accounts you follow. The Feed Recommendations card covers posts from accounts you do not. Both cards were updated 29 June 2026.

Explore is discovery. Its whole purpose is to show you posts from accounts you do not follow yet. The Explore card says it plainest. Retrieval pulls candidates from "accounts the user does not follow".

Reels Chaining is mixed. It gathers reels from accounts you follow. It also gathers "reels that are similar to those you've recently engaged with". Those can come from anywhere.

Stories is the outlier. It is the one surface built to reach people who already follow you. Not to find you new people. Not to test your content on strangers. To surface your story to the people who chose to follow you in the first place.

That makes Stories the opposite of what most advice assumes. Most advice is about reach. How to get seen by more people. How to grow. How to break out. Stories is not built for that. It is built to reach the people who are already there.

Here is where your content can go, surface by surface. Source: Meta Transparency Center system cards, retrieved 17 August 2026.

• Shown mostly to people who follow you:
  • Stories.
  • Feed, the posts that come from accounts you follow.
  • Reels Chaining, in part. It gathers reels from accounts you follow and from accounts you do not.
• Shown mostly to people who do not follow you:
  • Feed Recommendations.
  • Explore.
  • Reels Chaining, in part, for the same reason.

Stories and the from-followers half of Feed are the only two surfaces built to reach people who already follow you.

If your story reached sixteen people and your reel reached two thousand, that is not a mystery. They went to different places. Your reel went into Explore and Reels Chaining. It competed for strangers. Your story went into the stories tray. It competed for your followers. And if only sixteen of your followers opened their stories tray while your story was live, sixteen is what you got.

The signals that matter for Stories do nothing for Reels

Every surface has its own signal list. The lists do not overlap much. What makes a story rank high does almost nothing for a reel. What makes a reel rank high does almost nothing for a story.

Take the three most important signals for Stories. Viewing history. Engagement history. Closeness. All three are about one viewer and one creator. How often this viewer watches this creator's stories. How often this viewer replies to this creator. Whether the software thinks they are family.

None of those signals appear on the Reels Chaining card. Not one. Reels Chaining reads different things. How many times users watched at least three seconds of the reel. How many times the reel has been skipped within two seconds. How much time viewers from your country spent on this reel. Those are counts of many people. Aggregate numbers. Not one viewer's relationship with you.

The Feed card is the same. It lists how recently the post was created, within 90 days. Whether it is informative content. How many people liked, commented, shared or saved the post. No relationship signals. No viewing-history signals. No closeness guess.

Explore adds one relationship signal. The system predicts how likely you are to follow the author. But that is about whether you will follow them after seeing the post. Not whether you already do. Everything else on the Explore list is aggregate or content-based. How many people have seen the post on Explore. How many watched more than 95% of the video. Whether the post has text overlay.

Here is the split, sorted by the two kinds of signal. Source for all four surfaces: Meta Transparency Center system cards, retrieved 17 August 2026.

• Relationship signals, meaning one viewer's history with one creator:
  • Stories: how many times you have viewed stories from this account. How many messages you have exchanged with the author. How likely it is that you and the author are family or good friends.
  • Feed: none listed on the card.
  • Explore: how likely you are to follow the author. That is a guess about a follow that has not happened yet, not a record of one you already made.
  • Reels Chaining: none listed on the card.
• Aggregate signals, meaning what many people did:
  • Stories: none listed on the card.
  • Feed: how many people liked, commented on, shared or saved the post. How recently it was created. Whether it is informative content.
  • Explore: how many people have seen the post on Explore. How many watched more than 95% of the video. Its click-through rate.
  • Reels Chaining: how many times users watched at least three seconds. How many times it was skipped within two seconds. How much time viewers from your country spent on it.

The reading is plain. Stories is the only surface where relationship signals dominate. Every other surface leans on what many people did with the content. Or what the content is. Stories asks what this one viewer does with this one creator.

That is why advice written for Reels does not work for Stories. The software is reading different things. Trending audio matters for Reels. The Reels card lists audio reuse as a prediction. Whether the viewer will click "use audio" after watching. That signal does not exist for Stories. The Stories card never mentions audio. Not once.

Posting time matters for Feed. Maybe. The Feed card says the system reads "How recently the post was created (within 90 days)". Ninety days is the only window Meta puts on a post's own age. That holds across all five documents we read. But Stories does not rank by post age the same way. Stories ranks by relationship. If a viewer always watches your stories, your story will rank high for them no matter when you post it. If a viewer never watches your stories, your story will rank low for them no matter when you post it.

The advice you followed was written for the wrong system.

Early engagement does not move a relationship signal

Most engagement advice rests on one mechanism. Engagement is a signal. The ranking systems read it. A post with some engagement beats a post with none. So engagement in the first minutes helps a post rank.

That mechanism is plausible for Feed, Explore and Reels. All three read aggregate signals. Likes. Comments. Shares. Saves. How many people watched. How many skipped. Those signals exist in the first hour of a post's life. Feeding them early might help. Might, because no published experiment proves it. The claim is plausible. It is not proven.

For Stories the mechanism is weaker. Stories ranks by relationship signals. Those signals are about history. How often you have viewed this account's stories before. How often you have engaged with this account's stories before. How many messages you have exchanged with this account. Whether the system thinks you are family.

None of those signals change because ten people engaged with your story in the first five minutes. The viewer's history with you does not update that fast. The family guess does not flip because your story got early likes. The signals that dominate Stories ranking are slow signals. They are built from weeks or months of behavior. Not minutes.

That makes Stories the one surface where early engagement probably matters least. Feed and Explore read fresh signals. What people did with this post, now. Stories reads old signals. What this viewer did with you, before. You cannot shortcut a viewer's history. That history already happened.

There is one narrow exception. The Stories card lists one aggregate signal. It is not in Mosseri's top three. The card lists "How likely you are to click on a story at the top of your home page" as a prediction. To make that guess the system might read how many other people clicked on your story. That is an aggregate signal. Early engagement could feed it. But that is one prediction out of ten the card lists. And it is not in Mosseri's top three. The three that matter most are all relationship signals.

So whatever you add to a story on the day you post it has less to work with here than on any other surface. Stories is scored on what came before.

Recency matters, but not the way you think

One signal turns up everywhere. Time. But the cards describe it differently for every surface.

Feed says "How recently the post was created (within 90 days)". Ninety days is a long window. It tells you a post does not die in an hour. It says nothing about whether the first hour matters more than the second.

The Stories card says nothing about your story's age at all. Not in the signals list. Not in the predictions list. Nowhere on the page. The closest it comes is this. The system predicts "How likely you are to swipe to next story" and "How likely you are to exit the current story". Those are about the viewer's behavior. Not the story's age.

But recency does matter for Stories. In a different way. Stories disappear after 24 hours. That is not a ranking signal. That is a product rule. If you post a story at 6pm on Monday, it is gone by 6pm on Tuesday. If a viewer does not open Instagram in those 24 hours, they do not see it. Ever.

That makes posting time matter for a different reason. Not because the software ranks fresh stories higher. Because your followers have to open Instagram while your story is live. If most of your followers open Instagram in the morning, and you post at night, fewer people will see your story. Not because it ranked low. Because they were not there.

The viewing window is the constraint. Not the ranking system. Stories rank by relationship. But they only rank for viewers who are there. A viewer who is not there in the 24 hours is not ranked at all. They are just gone.

This is clearest when you compare surfaces. A reel can surface in Explore six months after you post it. The Feed card says posts are read within 90 days. Both of those are long windows. Stories has 24 hours. Then it is gone. Not deprioritized. Gone.

So posting time matters. But the mechanism is not "fresh stories rank higher". The mechanism is "your followers have to be there in the 24-hour window". That is a different problem. And it has a different answer. You do not optimize for ranking. You optimize for when your followers open the app.

What Meta does not publish

Three things would make this more useful. Meta does not publish any of them. So we say what is missing and stop.

The first is weights. Every surface's card lists predictions and signals. None of them give the weight on any prediction. So we know Stories tries to guess whether you are family. We do not know how much that guess moves the score. Mosseri's post says the three relationship signals are "the most important" for Stories. It also says they are "in rough order of importance". Rough is not a number. So every claim about which signal dominates is unproven. Ours included.

The second is absolute reach. The cards describe ranking. They do not describe reach. Feed scores about 500 posts for one person. That does not tell you how many people see your post. Explore fetches up to 1500 candidates and narrows to 100. That tells you nothing about whether your post was in anyone's 1500. The Stories card gives no funnel numbers at all. So we know Stories ranks by relationship. We do not know how many of your followers it shows your story to. Or whether it shows every story to every follower, just in a different order.

The third is how interaction changes the signals. Say a viewer never watches your stories. Then one of your stories picks up engagement from ten other people. Does that story now rank higher for the viewer who never watches? Or do the relationship signals stay fixed, because that viewer's own history with you has not changed? The cards do not say. The signals listed are histories. Whether one story getting outside engagement updates those histories is not documented.

So we know the system reads relationship signals. We do not know the weights. We do not know how much reach a high score gives you. And we do not know how fast the signals update. That is the honest state of it.

Where this leaves you

You reposted a reel to your story. The reel did well. The story did not. That is not a mystery anymore. They went to different systems. The reel went into Reels Chaining and maybe Explore. It was ranked by aggregate signals. How many people watched. How many skipped. How many saved. It competed with other reels for strangers.

Your story went into the stories tray. It was ranked by relationship signals. How often each of your followers watches your stories. How often they reply. Whether the system thinks you are family. It competed with other stories from accounts your followers already follow. If you have a hundred followers and ten of them opened Instagram in the 24 hours your story was live, ten is the ceiling. No amount of trending audio changes that.

The advice written for Reels will not fix Stories. The surfaces rank by different rules. Reels reads what many people did. Stories reads what one person does with you, over time. You cannot shortcut that. You build it by posting stories your followers reply to. By replying when they message you. By being someone they choose to watch.

Most guides never say this because most guides are copying each other. And the thing they are copying was written about Feed. Not Stories. Feed and Stories are both Instagram. But they are not the same system. One ranks by relevance. One ranks by relationship. The difference is in Meta's own documents. You just have to go read them.

So someone tells you to post at 6pm or use trending audio to fix your Stories reach. Ask them which system card that advice comes from. Then ask if they have read the Stories card. Probably they have not. Now you have.

What we could not determine

Four things we looked for and did not find.

Whether Stories ranks every story for every follower, or only some. The cards for Feed, Explore and Reels give funnel numbers. Feed scores about 500 posts out of the pool. Explore narrows from up to 1500 to 100. The Stories card gives no such numbers. So we know it ranks stories by relationship. We do not know if it shows every story to every follower in a ranked order. Or if it only surfaces some stories to some followers based on the relationship signals. The card does not say.

The weight on the family guess. The Stories card says the system predicts "How likely it is that you and the story's author are family or good friends". That is the only family signal across all five ranking documents we read. It is striking. It is also unweighted. Mosseri's post lists closeness third among the most important signals. It does not say how much more the first signal matters than the third. So we know the guess exists. We do not know its size.

How interaction from outside your followers updates the ranking. Say you have fifty followers. Ten of them watch your stories regularly. The other forty do not. Then twenty people from outside your fifty followers view one story and engage with it. Does that story now rank higher for the forty who usually skip you? Or do their personal relationship signals with you stay unchanged? The cards list histories as signals. They do not say how one story's outside engagement feeds into those histories.

How long a relationship signal takes to update. Stories ranks by viewing history and engagement history. Say a follower has never watched your stories before. They watch one today. Does that update their viewing history signal immediately? Does your next story rank higher for them? Or do the signals move slowly, over weeks? The cards list the signals. They do not describe the update speed.

These four would make the findings more useful. Meta does not publish them. So we name what we could not find and stop there.

How we did this

Scope: this post reads published documents. No data collection. No experiment of our own.

How we read them: each source was opened directly on 17 August 2026. Every quoted sentence was copied off the page as a reader sees it. Five of the six system cards are the Instagram Stories card, the Instagram Feed card, the Instagram Explore card, the Instagram Reels Chaining card, and the Instagram Feed Recommendations card. All five sit on Meta's Transparency Center and all five are built by code in the browser. So downloading the address on its own returns an empty page. They were read in a browser. The sixth card is the Transparency Center's index page listing all Instagram ranking pages. Mosseri's ranking post and the two Engineering at Meta posts arrive as finished pages from their servers. One typographic change is applied to every quotation. Curly apostrophes and quote marks in the sources are rendered as straight ones here, for consistency across the post. No other character is altered.

Limits: nine of eleven sources are first-party statements from Meta or Instagram. A system card is a description Meta chose to publish. Not an audited account. Meta rewrites these cards without saying when. So every quotation here carries the card's "updated" stamp, the one it showed on the day.

Sources

1. Instagram Stories AI system card, stamped UPDATED SEP 18, 2025. Meta Transparency Center. Read in a browser. Retrieved 17 August 2026.

1. Instagram Feed AI system card, stamped UPDATED JUN 29, 2026. Meta Transparency Center. Read in a browser. Retrieved 17 August 2026.

1. Instagram Explore AI system card, stamped UPDATED JUN 22, 2026. Meta Transparency Center. Read in a browser. Retrieved 17 August 2026.

1. Instagram Reels Chaining AI system card, stamped UPDATED NOV 11, 2025. Meta Transparency Center. Read in a browser. Retrieved 17 August 2026.

1. Instagram Feed Recommendations AI system card, stamped UPDATED JUN 29, 2026. Meta Transparency Center. Read in a browser. Retrieved 17 August 2026.

1. Our approach to explaining ranking, stamped UPDATED DEC 31, 2023. Meta Transparency Center. The index page listing all Instagram ranking system cards. Read in a browser. Retrieved 17 August 2026.

1. Adam Mosseri, Instagram Ranking Explained, 31 May 2023. About Instagram. Source of the "singular algorithm" quote and the three most important Stories signals listed in rough order of importance. Retrieved 17 August 2026.

1. Vladislav Vorotilov and Ilnur Shugaepov, Scaling the Instagram Explore recommendations system, 9 August 2023. Engineering at Meta. Retrieved 17 August 2026.

1. Luke Levis, Sing Sing Ma and Eduardo Nava, Journey to 1000 models: Scaling Instagram's recommendation system, 21 May 2025. Engineering at Meta. Source of the "over 1000 ML models" statement. Retrieved 17 August 2026.

Changelog

• 2026-08-18: First version drafted.