How to calculate engagement rate so the number survives review

Three sources, three formulas, three different numbers for the same post. One divides by followers, one by reach, one by impressions, and the agency deck uses whichever is highest. Then a client compares your figure to one their other agency produced, and somebody has to explain why the same post scored 2.1% and 6.8%.

This is how to calculate engagement rate so that the number survives that conversation. There is no single correct formula, there is a correct formula for who you are measuring, and the rest is writing it down.

TL;DR

Step one: decide what counts as engagement

The denominator gets all the attention and the numerator causes as many arguments. Interactions is not a defined term, and two people counting it differently will disagree before they even reach the division.

• Likes and comments. Public everywhere, on any account. This is the honest default.
• Shares and reposts. Public on most networks, and a much stronger signal than a like.
• Saves. Owner only on most networks. Including them makes your own accounts look better than anybody you compare to.
• Clicks. Owner only. Same problem, larger.

The rule that keeps you out of trouble: the numerator must be available for every account in the comparison. The moment you add a term only your own accounts have, you have built a formula that flatters you by construction.

Any formula using a number only you can see will make you look better than every competitor, forever, and for no real reason.

Step two: pick a denominator you can defend

This is the whole decision. Reach is the most meaningful denominator and it is given to the account owner, so it exists for your accounts and for nobody else's.

That leaves a simple rule. Measuring only accounts you own, use reach and get a genuinely better number. Comparing to anybody else, use followers, because it is the only denominator both sides publish. The difference between the two metrics people confuse here is in reach vs impressions.

TikTok and YouTube are the exception worth knowing: both publish a view count on every video, to everybody. On those two networks you can use views as a denominator for a competitor, which is a better comparison than followers and is not available anywhere else.

Step three: the calculation, worked through

A post with 84 likes and 12 comments, on an account with 3,200 followers. Interactions are 96. Divided by followers that is 0.03, which is 3%.

The same post, if you own the account and it reached 1,900 people, is 96 divided by 1,900, which is 5.05%. Same post, same interactions, two numbers that differ by two thirds. Neither is wrong and putting them in the same column is.

1. Per post, then averaged. Compute the rate for each post and average the rates. This is the usual method.
2. Or totals over a period. All interactions divided by followers, across a month. Smoother, and it hides a single viral post.
3. Never mix the two. A month computed one way next to a month computed the other is a trend line that means nothing.

Step four: handle a growing follower count

A subtle one that quietly corrupts a year of data. If the account is growing, the follower count you divide by changes every week, so an unchanged post performance shows as a declining rate.

That decline is real in one sense and misleading in another: reaching the same number of people out of a larger audience is genuinely a smaller share, and it is not a drop in output quality. State which you are reporting, and keep the absolute interaction count in the same table so the two can be read together.

The same problem in its purest form, and how to compute a growth figure that survives it, is in follower growth rate.

Step five: be careful what you compare it to

Published benchmarks are the most quoted and least comparable numbers in this market, because almost none of them state their formula, their numerator or their sample.

The safest comparison is against yourself over time, computed identically. The second safest is a small set of named rivals you compute yourself, which is set out in how to benchmark social media performance.

Step six: write the definition in the report

One line, in the footer, every month. It costs nothing and it is the difference between a number somebody trusts and a number somebody queries.

Something like: engagement rate is likes plus comments divided by followers at the end of the period, computed per post and averaged. That sentence makes the figure checkable by anyone, which is the only property that matters when a client compares your deck to somebody else's.

And if you ever change the definition, say so in the report before anybody spots the jump. A silent change makes month thirteen look like a step change that never happened, and that conversation is much worse than the one you avoided.

What changes network by network

The formula is the same everywhere and what is available to put in it is not, which is why a single cross-network engagement rate is usually an average of unlike things.

Two consequences worth acting on. On TikTok and YouTube you can use a view denominator for a competitor, which is a genuinely better comparison than followers and is unavailable elsewhere. On Instagram an account can hide its like count, which removes the numerator entirely, and a tool that keeps drawing a line there is decorating rather than reporting.

So compute and report per network, and resist the combined figure. If somebody insists on one number, make it posts published, which is the only thing that means the same on all four.

What the rate does not tell you

It is the most requested metric in this market and it is a weaker signal than its popularity suggests. Worth knowing before it becomes a target.

• It falls as an account grows. Reaching the same people out of a larger audience is a smaller share and not a decline in quality.
• It rewards small accounts. A profile with 300 followers will beat one with 30,000 on this metric almost every time.
• It counts a like and a considered comment identically. Unless you weight them, and then it is your metric rather than a comparable one.
• It says nothing about who engaged. Ten replies from target buyers beat four hundred likes from peers, and the rate cannot see the difference.

Keep it because it is comparable and cheap, and pair it with one qualitative line about who is actually replying. That pairing is what stops a team optimising toward a number that flatters them and sells nothing.

The arithmetic is the easy half. What the word covers, which reactions are worth anything, and why chasing the number directly makes accounts worse, is in social media engagement.

How Groowth handles this

It records the inputs rather than imposing a formula: what was published and when, the follower count on the day, and the public likes and comments on each post. Because the readings are taken daily, the follower count used for a given period is the one that was true at the time rather than today's.

It never computes a rate against reach or impressions for an account it does not own, because those numbers are not available to it and inventing a denominator is how the two figures in the worked example above end up in one column. Three accounts free with no card. Which figures across the whole category are measured and which are modelled is audited in best competitor analysis tools for social media.

Frequently asked questions