What Is Engagement Rate, and What Counts as Good?
How to calculate engagement rate correctly, which formula to use when, why saves and shares matter more than likes, and how to build a baseline for your own account.
13 min readBy SocialTools OnlinePublished January 8, 2026Updated September 11, 2026
What engagement rate is trying to measure
Engagement rate expresses how much interaction a piece of content produced relative to the size of the audience that could have interacted with it. It exists because raw interaction counts are not comparable: two hundred comments on an account with a thousand followers means something entirely different from two hundred comments on an account with a million.
Because it is a ratio, both halves matter and both can be defined several ways. That is the source of most confusion about the metric. There is no single official engagement rate, and two tools reporting different numbers for the same post are often both correct according to their own definitions.
This means the useful question is not “what is my engagement rate” but “compared to what.” A number without a consistent basis and a comparison set tells you nothing actionable.
The formulas, and when each is appropriate
Engagement rate by followers takes total interactions on a post, divides by follower count, and multiplies by one hundred. It is the most widely quoted version and the easiest to calculate from public data, which is why brands and agencies use it for comparing accounts. Its weakness is that it penalizes accounts with high reach beyond their following and flatters accounts whose followers are unusually concentrated.
Engagement rate by reach uses the number of unique accounts that actually saw the post as the denominator. This is the more meaningful measure for judging whether a specific post worked, because it separates content quality from distribution. It requires access to your own analytics, so you cannot compute it for someone else’s account.
Engagement rate by impressions uses total views including repeats. It produces the lowest numbers of the three and is mainly useful for paid campaign reporting. Whichever you choose, choose one and label it. Mixing formulas across a report is the fastest way to draw a false conclusion about a trend.
Which interactions should count
Likes and comments are always included. Saves and shares should be, because they are the interactions platforms weigh most heavily and the ones that predict distribution best. Some definitions also include profile visits, link clicks, and sticker taps; if you include them, say so, because it materially raises the number.
The reason saves and shares matter more is that they cost the viewer something. A like is a reflex. A save is a commitment to return, and a share is a claim that the content is worth another person’s time. Both are much stronger evidence that the post delivered value.
Practically, this means a post with fewer likes and many saves is frequently the better post, and reporting that only counts likes will tell you to make more of the wrong thing. If your content is reference material - checklists, size charts, how-tos - saves are your primary success metric and likes are close to noise.
Why published benchmarks should be handled carefully
You will find quoted “good” engagement rates for every platform, and they vary widely between sources because each is computed from a different sample, a different formula, and a different year. Treating any of them as a standard is a mistake: a figure derived from a dataset of large brand accounts says very little about a niche account with two thousand followers.
Two structural effects make cross-account comparison unreliable. Smaller accounts typically show higher rates than large ones, because a tightly connected audience interacts more per person. And platforms differ in what a “view” even means, so TikTok percentages computed against follower count routinely exceed one hundred percent and are not comparable to an Instagram figure at all.
Use external benchmarks, if at all, as a very rough orientation. Use your own history as the actual standard. Whether this month beat last month on a consistent formula is a question you can answer accurately, and it is the one that should drive decisions.
Build a baseline you can actually use
Pull your last twenty posts and record, for each: date, format, topic, reach, likes, comments, saves, shares, and profile visits. Compute engagement rate by reach with a fixed definition. The median of those twenty is your baseline. The median rather than the mean, because one viral post will otherwise distort the number for months.
Now you can ask useful questions. Which format sits above the baseline - carousels, single images, Reels? Which topics? Which caption lengths? Which posting windows? The patterns that appear across twenty posts are real; differences between two posts are noise.
Refresh the baseline quarterly. Audiences change, platforms change, and a baseline computed a year ago will quietly mislead you. Keeping the sheet is unglamorous and it is the difference between strategy and guessing.
Diagnosing a drop
Falling engagement rate has three broad causes, and they need different responses. The content stopped resonating; the distribution changed; or the denominator grew faster than the audience quality. A follower spike from one viral post mechanically lowers your follower-based rate even if nothing about your content got worse.
Check reach before you rewrite your strategy. If reach fell alongside engagement, the problem is distribution - posting when your audience is offline, restricted hashtags, a format the feed is currently favouring less, or simply a quieter period. If reach held steady and engagement fell, the content is the variable.
Also check whether your audience has drifted from your topic. Accounts that grow through one unusually broad post often acquire followers with no interest in the core subject, and every subsequent post is measured against an audience that was never the right one. That is a follower-quality problem, not a content problem, and it resolves as normal posting re-sorts the audience.
Reporting engagement honestly
If you report to a client or a manager, state the formula and the denominator every time. “Engagement rate by reach, counting likes, comments, saves and shares” is one line and it prevents every downstream argument about why the number differs from a tool they checked.
Report alongside a business metric wherever possible. Engagement is a proxy, and a proxy that rises while enquiries fall is not a success. Profile visits, link clicks, saves on the posts that lead somewhere, and actual enquiries make the story defensible.
Be direct about what engagement rate cannot tell you. It does not measure audience quality, purchase intent, or brand perception. It measures whether the people who saw a post reacted to it. That is genuinely useful, and it is not the same as effectiveness.
Tools for this guide
Put these tips into practice with free utilities - no signup required.
Browse all tools →Related guides
About the author
SocialTools Online
Site editorial
Guides and tool instructions are written by the SocialTools Online team. We document how each browser tool works, platform limits we can verify, and practical workflows for creators and small businesses. We do not sell growth services or guarantee search rankings.
Cite or share this guide
Link the canonical URL above. Embeddable counters and citation HTML: Press & embed.