What's a good Instagram engagement rate in 2026? (Why every benchmark you found disagrees)

You searched for a benchmark and got a mess. One 2026 report says the average Instagram engagement rate is around 0.5%. Another says micro-creators average 3.5%. A third says Reels land between 4% and 7%. Then you open your own insights and see a number that matches none of them.

The short version: those numbers aren't measuring the same thing. Public benchmark reports almost always divide by followers, because followers are the only number they can see from the outside. Your own insights divide by reach or views. Same account, same week: the follower-based figure will typically be a fraction of the reach-based one. Before you judge your rate, find out which denominator produced it.

The short version:

• Every public benchmark divides by followers. Socialinsider puts Instagram at 0.48% for 2025. Rival IQ's 2026 report puts the median at 0.30%. Both are follower-based, and both are right about different samples.
• Your insights divide by reach, so your number will be several times larger. That is not you outperforming the benchmark. It is a different sum.
• There is no published benchmark for engagement on reach, at any follower size. Reach is private data, so nobody outside your account can measure it. Anyone showing you a reach-based band made it up.
• Carousels, not Reels, top the engagement-rate tables in the largest datasets, while Reels win on interactions and discovery. Which one is "best" depends on the metric.
• The only rigorous comparison is your account against itself, on one denominator, split by format.

Once you know that, the question "is 3% good?" becomes answerable. Here's how.

Three formulas, three completely different numbers

Every engagement rate is interactions ÷ something. The numerator is fairly settled (likes, comments, saves, shares, sometimes replies). The denominator is where the chaos lives.

ER by followers = interactions ÷ follower count. This is what third-party benchmark tools, agency reports, and most "industry average" tables use. Not because it's the best metric, but because it's the only one computable without access to the account. Reach is private data. An outside tool can count the likes under your post and read your follower number off your profile; it cannot see how many people the post actually reached.

ER by reach = interactions ÷ unique accounts reached. This is the honest one for creators, and it's what Instagram's own Professional dashboard leans on. It answers: of the people who actually saw this, how many did something?

ER by views = interactions ÷ total views (what used to be split across impressions and plays before Instagram consolidated its video metrics). Views count repeats, so this denominator is the largest of the three and the resulting percentage is the smallest.

The gap between these isn't small. If a post reaches 20,000 accounts on a 40,000-follower profile, the reach-based rate is exactly double the follower-based one, from identical engagement. If a Reel goes wide to non-followers, reaching 120,000 on that same 40,000-follower account, the follower-based number inflates to something that looks incredible and means very little.

This is the whole reason the published tables contradict each other. They're not wrong. They're answering different questions and labelling both "engagement rate."

So what's actually good? 📊

Here is every platform-wide figure I could open at its source, with the denominator and the sample stated. They disagree, and the disagreement is the useful part.

| Source | Instagram engagement rate | Basis | Sample | |---|---|---|---| | Socialinsider, full year 2025 | 0.48%, "registering a 24% decrease YoY" | followers | 35M posts, 447,613 pages | | Socialinsider, quarterly | 0.45% in Q1 and Q2 2026, "dipping from 0.52% in Q1 2025" | followers | 70M posts, Jan 2024 to Dec 2025 | | Rival IQ / Quid, 2026 report | 0.30% median, "falling from 0.36% to 0.30%" | followers | 150 companies from each of 18 industries |

Three studies, three answers, and a factor of 1.6 between the top and the bottom. All three divide by followers, because that is the only denominator an outside tool can see. So a 0.30% median and a 0.48% average are not a contradiction to be resolved. They are different panels of accounts.

One thing to note before you carry the 0.48% anywhere: Socialinsider says of its own study that it "contains 2025 values, presented as 2026, because, at the time of publication, there wasn't enough data for analysis, with the year just starting." A benchmark labelled 2026 is frequently last year's number.

The band you actually want does not exist

There is no published benchmark for engagement on reach, at any follower size. Not from Socialinsider, not from Rival IQ, not from Sprout, not from Metricool. Anyone showing you one built it out of nothing.

The reason is the same one that opened this post. Reach is private. An outside tool can count the likes under your post and read your follower number off your profile, and that is all. So a reach-based band is not a number somebody has failed to publish yet. It is a number nobody outside your account is in a position to calculate.

Which means the reach-based rate in your own insights has exactly one honest comparison, and it is your own trailing average. Put the number on your media kit with the denominator named, and put your own trend next to it rather than an industry table.

By format, and the ranking may surprise you. In Socialinsider's data, carousels come out on top, not Reels: for full-year 2025, carousels 0.55%, Reels 0.52%, images 0.37%. For Q2 2026, carousels 0.50%, Reels 0.48%, images 0.33%. Metricool's study of 24,364,803 posts from 375,118 accounts points the same way on saves, finding carousels generate "nine times more saves than single-image posts", while Reels generate "more than four times the interactions of single-image posts". So Reels win on interactions and discovery, carousels win on engagement rate and saves, and single images lose on everything: Metricool has their reach down 21.96% and engagement down 45.98% year over year. Never average your formats together. A single account rate that is really two very different format rates is two stories, and only one of them should decide next week.

By niche. Rival IQ's 2026 report covers 18 industries and finds "Higher Education and Sports once again led the pack", with "Retail and Beauty remained near the bottom", against an Instagram median of 0.30%. Two warnings about niche tables generally. The categories are industries, not creator niches, so photography, travel and fitness do not appear in any of them. And the pets-and-sports-do-well finding that circulates every year is, in this report, a YouTube result, not an Instagram one. A rate in a low-engagement category may well beat a higher rate in a high-engagement one. Compare to your own niche or don't compare at all.

The comparison that actually matters

Here's the uncomfortable part: none of the above tells you whether your account is doing well. Benchmarks are averages of strangers with different niches, audiences, geographies, and posting cadence.

The only rigorous comparison is your account against itself over time. Your engagement rate last month versus this month, on the same denominator, split by format. That's a controlled comparison: same creator, same audience, one variable changed.

Which runs into a practical problem: Instagram's insights only look back so far. Meta's own documentation puts account-level metrics at 90 days, and what that 90-day window quietly deletes is exactly the baseline this comparison needs. If you want to know whether your saves-per-reach on carousels is better now than in March, you need to have been recording it in March. Most creators weren't, so they end up comparing themselves to a stranger's benchmark table instead.

That's the specific gap we built Junibloom to close. We snapshot your numbers daily from the official Instagram API (read-only, no scraping, no password), so your own history is always there to compare against. The engagement panel shows the reach-based rate, split by format, against your own trailing baseline rather than a generic industry average.

A five-minute audit you can run today

1. Pull your last 30 days: total reach, and likes + comments + saves + shares.
2. Divide interactions by reach. That's your real rate. Write it down with today's date. You're starting the history you'll want in three months.
3. Do it again for Reels only, and for carousels only. Note which is carrying the account.
4. Check the outliers. Your best post by engagement rate is not usually your best post by reach. The wide-reach one went to strangers, the high-ER one went to your actual audience. Both are useful; they're just useful to different sponsors.
5. When you quote a rate on a media kit or in a pitch, state the denominator: "4.1% engagement on reach." Any brand manager who knows the space will trust you more for it, and the ones who don't will have learned something.

Common questions ❓

What is a good Instagram engagement rate benchmark? On a follower denominator, which is what every public benchmark uses, the published figures for 2025 and 2026 run between 0.30% and 0.48% depending on whose panel you read. Anything you see quoted above about 1% is either a much smaller sample, a different denominator, or an invention.

Is 3% a good engagement rate on Instagram? Against a follower-based benchmark, 3% is many times the platform average and worth checking. Against your own reach-based number from the app, 3% is unremarkable, because that sum has a much smaller denominator. The figure is meaningless until you say which one produced it.

Why is my engagement rate higher than every benchmark I find? Almost always because your insights divide by reach and the benchmark divides by followers. If a post reaches half your followers, your reach-based rate is roughly double the follower-based one from identical engagement.

How do I calculate my Instagram engagement rate? Pick the denominator first, then keep it. Socialinsider's follower version is "the sum of likes and comments" divided by followers, times 100. The reach version is total interactions, including saves and shares, divided by accounts reached. Do not mix them across months and call it a trend.

Does engagement rate drop as you get more followers? On a follower denominator, yes, and mostly for arithmetic reasons rather than because your content got worse. Your follower count grows faster than the share of followers any single post reaches. The same effect makes a falling reach number look worse than it is.

The one-line takeaway

A good engagement rate is one you can define, reproduce, and beat next month. Everything else is a stranger's spreadsheet.