Engagement rate: how to count it honestly in 2026
A client asks: “what’s our engagement rate?” You open three analytics tools and get 1.8%, 3.4% and 11.5% — same account, same month. Nobody made a mistake. All three just divided by something different.
Engagement rate is the one metric in social media that everyone calculates differently and then compares anyway. That’s where arguments like “our ER is 2%, is that bad?” come from — a question nobody can answer until you say what was in the denominator.
Below: the three formulas and when each one is honest, a worked example with real numbers, benchmarks by account size and by network, ways to lift engagement without faking it, and the cases where ER is worth ignoring entirely.
Why everyone gets a different ER
The numerator is nearly always the same — total interactions with a post. The disagreement is about what you divide by.
- By followers — “how much of my base reacted.”
- By reach — “how many of the people who saw it reacted.”
- By views — the same thing, but counting replays.
On a single post, the gap between the first and second option is routinely three or four times. So rule one: an ER number without its denominator isn’t a metric, it’s a number. In a report, never write “ER 3.4%” — write “ER by reach 11.5%, ER by followers 3.4%.”
Three formulas on one example
Real numbers: a specialty coffee brand with 8,200 followers. A post about a new seasonal menu gets 2,400 reach, 3,100 views, 180 likes, 22 comments, 45 saves, 30 sends.
Total interactions: 180 + 22 + 45 + 30 = 277.
| Formula | Math | Result | When it’s honest |
|---|---|---|---|
| ER by followers | 277 / 8,200 | 3.4% | Comparing you to your past self |
| ER by reach (ERR) | 277 / 2,400 | 11.5% | Judging one specific post |
| ER by views | 277 / 3,100 | 8.9% | Video and Reels, where replays matter |
The point: ER by reach grades your content, ER by followers grades your base. If ER by followers is falling while ER by reach holds steady, the content is fine — you’re simply not being shown to your followers anymore. Two different problems, two different fixes, and one blended number hides both.
There’s a fourth number agencies like: account-level average ER — add up the ER of every post in a period and divide by the count. Use the last 10–15 posts and take the median, not the mean. One viral post lifts a mean so far that the report stops meaning anything.
What counts as an interaction in 2026
The classic “likes + comments” formula is a fossil from a time when there were no other buttons. On Instagram today you should count four actions, and they are not equal:
- Sends — the strongest signal there is: someone vouched for your content to a specific person.
- Saves — “I’ll need this,” pure practical value.
- Comments — they cost effort, especially the substantive ones.
- Likes — half a second and almost zero cost.
Many analytics tools still add up only likes and comments, because saves and sends aren’t exposed publicly for every account. That’s the eternal gap between a third-party dashboard and what you see inside the app. For your own ER, count all four — the data is in per-post insights, and where to find it is covered in our guide to Instagram analytics.
One thing to keep out: link clicks and profile visits are conversion, not engagement. Track them, but don’t fold them into ER, or the metric stops being comparable to its own history.
What counts as a good ER
A caveat first: what follows isn’t “according to research.” These are the orders of magnitude you see on live accounts. Treat them as a traffic light — green, amber, red — not as a standard.
Instagram, ER by reach (likes + comments + saves + sends over reach):
- under 2% — the content isn’t landing with the people it reached;
- 3–6% — the working normal for most niches;
- 8% and up — a strong post. Take it apart and do it again.
Instagram, ER by followers depends heavily on size, and that’s arithmetic rather than quality:
| Followers | Typical range |
|---|---|
| under 1,000 | 4–8% |
| 1,000 – 10,000 | 2–4% |
| 10,000 – 100,000 | 1.5–3% |
| 100,000+ | around 1% and below |
ER dropping as the account grows is normal. Your first thousand followers are people who are personally interested in you; your tenth thousand followed after one lucky Reel and forgot. So “we used to have 6% ER, now it’s 2%” usually describes growth, not decline.
Niche shifts the bar too: local businesses, experts and small communities run higher; broad-reach news and entertainment accounts run lower. Comparing a neighbourhood studio to a media brand tells you nothing.
Other networks count it differently
Telegram. Here ER means views over subscribers: a 4,000-subscriber channel, a post with 1,600 views → 40% ERR. Live channels sit around 20–40%, and small niche ones go higher. Consistently under 15% almost always means an aging base or bought subscribers. Count reactions and forwards separately: forwards are Telegram’s main organic growth channel, as we covered in growing a Telegram channel.
X. The platform shows its own engagement rate in post analytics: all interactions — including clicks and expands — over impressions. Clicks inflate it relative to Instagram, so never compare the two head-to-head.
Threads. There’s no public ER metric, so calculate it by hand: replies + reposts + quotes over views. Replies weigh more than likes here — the feed is visibly friendlier to threads with actual discussion.
YouTube and Shorts. ER is secondary; retention and thumbnail CTR decide everything. Likes and comments over views are useful only as a signal about the topic.
The general rule: compare ER within one network and one format only. Putting Reels and carousels in the same report column is already an error.
How to lift ER honestly
- Ask a question that’s easy to answer. “What do you think?” gets nothing. “Do you spend more or less than $50 a month on delivery fees?” gets comments, because a one-word answer is enough.
- Make things people save. Prices, sizes, checklists, packing lists, comparison tables. Saves are the most underrated part of the numerator.
- Ask for sends, not likes. A direct line at the end of a post reliably outperforms a vague “show some love.”
- Reply within the first 30–60 minutes. Your replies are comments too, and they keep the discussion alive while the post is still being distributed.
- Cut the weak posts. ER averages across everything you publish: three filler posts a week drag the average down harder than one strong post lifts it. Two good posts beat five “because we had to” — which is easier to sustain with a content plan.
And how ER gets faked: giveaways with “tag three friends,” bought comments, engagement pods. The number in the report goes up; the reach of your next posts doesn’t, because the new audience came for a prize and never reacts again. A month later ER is below where it started: the base grew, interest didn’t. If a contractor shows you 12% ER on a 50,000-follower account, ask for a screenshot of reach — that usually ends the conversation.
When ER is useless
- Under 500 followers. One share from a friend moves the percentage by half. Watch absolute comment and send counts instead.
- Posts with ad spend. Paid reach is colder than organic, so ER drops mechanically. Report boosted posts separately.
- Long sales cycles. For an industrial supplier, 40 comments can be worth less than two — from two factories. ER doesn’t know the size of the deal.
- When the goal is traffic. For a link post, clicks matter and saves don’t. Your best-performing post in revenue terms often has the month’s worst ER.
Tracking ER is easiest when the numbers collect themselves: in Repost21, Instagram, YouTube and Threads analytics sit next to the publishing calendar, so you see not just “how much” but “after which post.” Try it free — and remember that the benchmark that matters isn’t the market, it’s you three months ago.