You open the insights under a post: “Views 4,812”, “Reach 3,105”, “Interactions 214”. Which of these is good? Why are views higher than reach if it’s one post? What counts as an “interaction” when likes are 150 and comments are 9? Instagram shows you the numbers but never ships the dictionary.

This article is that dictionary. We walk through Insights tab by tab and explain how each metric is calculated, how it differs from the one next to it, and what range counts as normal. You won’t find “the 7 metrics that matter” here; that’s covered in the Instagram analytics guide. This is the reference: so that any number you see, you know what’s behind it.

One caveat up front: Insights only exists on a professional account (Business or Creator). Switching is free, reversible and lives in Settings. Data starts accumulating from the moment you switch; Instagram won’t backfill history.

Quick glossary: the terms people look up most

If you came for one specific word, here is the short version. Each metric gets a full explanation below, tab by tab.

Term in InsightsWhat it means
Viewers (accounts reached)How many unique accounts saw the post at least once. Same thing as reach: one person counts once, however many times they open it.
ViewsHow many times the post was shown on a screen. One person can generate several views, so views are always higher than viewers.
ReplaysThe gap between views and viewers: the same person watched again. On Reels they count toward views, and a high replay share means the clip gets rewatched.
Profile activityWhat people did after tapping through to your profile: visits, link-in-bio taps, Call, Email and Get directions buttons, follows. Shows the content created interest in you, not just in the post.
Skip rateReels only: the share of viewers who swiped away in the first seconds. Lower is better; read it together with views longer than 3 seconds.
InteractionsLikes, comments, shares and saves added up. Instagram weights them differently: shares and saves matter more than likes.
ForwardStories: a tap to the next slide inside your own sequence. Normal reading behaviour.
Next storyStories: a swipe to another account, so the viewer left your Stories. The slide got boring.
ExitedStories: the viewer closed Stories altogether. Worse than Next story: the slide pushed them away.
Professional dashboardThe screen in a Business or Creator profile that opens Insights, tools and the ads manager.

The map: what lives where

Account-level Insights (profile → Professional dashboard → Insights) has four blocks and a period switcher: 7, 14, 30, 90 days, last month or a custom range. The app won’t look further back than 90 days. Older history either goes into a spreadsheet once a month or lives in a third-party tool.

SectionWhat’s insideThe question it answers
ViewsTotal views, by format, from followers and non-followersHow many times your content was shown
InteractionsLikes, comments, shares, saves by formatHow many times someone did something with it
FollowersGrowth, unfollows, cities, age, gender, active hoursWho watches you and when
Content you sharedEvery post, Reel and Story in the period, sortable by any metricWhat worked best

Separately, each post has its own insights right under it, more granular than the account view.

Views, reach, impressions: the difference

In 2024 Instagram collapsed three older metrics (impressions, plays, video views) into one: Views. It’s shared across every format: photos, carousels, Reels, Stories, Lives.

  • Views is how many times the content appeared on a screen. One person opening a post three times gives three views. For Reels, replays count too.
  • Reach is how many unique accounts saw the content at least once. That same person gives one reach.
  • Impressions is the old metric and no longer exists in the app. If a third-party tool still shows “impressions”, it’s either views under the old name or cached data.

So views are always ≥ reach. The views-to-reach ratio tells you whether people rewatch. For photos the norm is around 1.1–1.3; for Reels, 1.5 and up means the clip is being replayed. Below 1.1 on a Reel means one look and a scroll.

The account-level Views section splits followers vs non-followers. This is the most underrated number in all of Insights: if 90% of views come from followers, the algorithm isn’t carrying you beyond your own base. For an account that wants to grow, a healthy non-follower share starts at 30–40%, and a Reel that lands can reach 80–90%.

Interactions: four actions, four different weights

Interactions is the sum of likes, comments, shares and saves. The parts aren’t equal, and Instagram has said as much: for ranking, shares and saves carry noticeably more weight than likes.

  • Likes are the cheapest action. A mood indicator, nothing more.
  • Comments count everything, including your replies. If you answer every comment, honestly halve the figure.
  • Shares (Sends) are forwards to DMs and to Stories. That’s a person recommending you to another person, and the algorithm reads it as the main “show this to similar people” signal.
  • Saves mean “useful later”: prices, how-tos, lists, templates.

Accounts engaged is a separate metric; don’t confuse it with interactions. It’s the number of unique accounts that took at least one action. Accounts engaged divided by reach is the honest engagement rate by Instagram’s own definition. How to calculate it by hand and what’s normal is in the engagement rate article; a 2026 reference for feed posts is 3–6% of reach, lower on Reels, higher on carousels.

Post insights: what every line means

Under a specific post, the metrics run in this order (labels vary slightly by app version):

  1. Views, split into followers and non-followers.
  2. Reach, the same two groups but unique accounts.
  3. Views from: Home, Profile, Explore, Hashtags, Other. Home is your followers’ feed. Explore and Other (Reels recommendations, shares) are what bring in new people. If 95% of views come from Home, the post lives on followers alone.
  4. Interactions, the four actions broken out.
  5. Profile activity: profile visits, follows, link taps in bio, taps on the Call and Email buttons. Follows counted here are only those that came from this post. Benchmark: 2–4% of reach visits the profile, and 10–15% of those follow. If the first number is fine and the second isn’t, the problem is your bio, not the post.

Carousels add views per card: you see how far people swipe. The first drop of more than 40% shows where the carousel should have ended.

Reels: retention over views

Reels have their own lines:

  • Watch time is the total, in minutes. Useless on its own, but it’s the number the algorithm compares between clips.
  • Average watch time is watch time ÷ views. Compare it with the clip’s length: 9 seconds on a 30-second Reel means two thirds of viewers left.
  • Retention graph shows the share of viewers second by second. A cliff in the first 3 seconds is a bad opening frame. A steady slope means the clip is longer than its idea. A bump in the middle means people are scrubbing back to a specific moment: make a separate clip about it.
  • Views from non-followers has a different norm for Reels than for the feed: 60–90% non-followers on a good clip.

Some accounts already see a Skip rate line: the share of viewers who scrolled past in the first seconds. If you have it, it’s the most direct answer to “does my first frame work”.

Stories: navigation only makes sense in pairs

Stories insights look bloated because there are five navigation numbers per slide. They make sense only when read in pairs:

MetricWhat it meansBad sign
Views and reachSame as postsReach on the last slide < 25% of the first
ForwardTap “next” within your storyNormal, that’s just reading
BackReturned to the previous slideGood: something was worth re-reading
Next storyLeft for another accountThe slide bored them but didn’t repel
ExitedClosed Stories entirelyWorse than Next: the slide pushed them away
Replies and sticker tapsReactions, replies, poll and link tapsZero on a slide that asks a question

A normal retention for an 8–10 slide series is 30–40% from first to last. One slide with a spike in exits is worth a week of hypotheses: almost always it’s a wall of text or a sudden jump to selling.

Followers: demographics and hours

The section unlocks at 100 followers; before that it’s empty, and that’s not a bug. Inside:

  • Growth and unfollows by day. Look at unfollow spikes and match them to specific posts, not at total growth.
  • Cities and countries. Essential for local business: if 40% of a Denver bakery’s followers live elsewhere, ads targeted at followers are wasted money.
  • Age and gender. Instagram infers these from behaviour, not from profiles, so accuracy is moderate. Decide on large groups, not single percentage points.
  • Most active times shows when followers are online, not when they respond to you. It’s a hypothesis to test, not a schedule. How to test slots on your own data is in the best time to post guide.

Three traps when reading the numbers

  • Comparing boosted and organic posts. As soon as you “boost” a post, organic and paid reach blend into the same figures. Flag those posts in your spreadsheet.
  • Conclusions in the first 48 hours. Reels and carousels keep collecting views for weeks. Score a post after 7 days.
  • Percentages on small numbers. At a reach of 300, one friend’s share is “+60%”. Below 500–1,000 followers, look at topics and comments, not trends.

If you run several accounts and are tired of copying these numbers from the app into a spreadsheet, Repost21 pulls Instagram, YouTube and Threads analytics into one view, and its content calendar suggests slots from the same data. Start for free; one connected account is enough to see whether you need a layer on top of built-in Insights.