Engagement rate is the number everyone quotes and almost nobody defines. It is genuinely useful, but only once you know what it is measuring and what it quietly leaves out.
There is no standard formula
The first problem is that "engagement rate" is not one metric. Depending on who is calculating it, you might be looking at:
- Likes and comments divided by followers
- Likes, comments, saves and shares divided by followers
- Any of the above divided by reach or by impressions
These produce very different numbers for the same account. An engagement rate calculated against reach will almost always look healthier than one calculated against followers, because reach excludes the followers who never saw the post.
So the first question about any engagement figure is not "is it good" but "what is the denominator". A creator quoting 12% and a tool reporting 3% for the same account may both be correct.
Why a high rate can be a bad sign
Engagement rate correlates inversely with audience size, and that is normal. Smaller accounts have audiences that know them; large accounts accumulate passive followers. A 9% rate on a 5,000-follower account and a 2% rate on a 500,000-follower account can represent similar audience quality.
A genuinely high rate on a large account is worth a second look, not immediate excitement. Things that inflate it:
- Giveaway posts, which attract comments from people with no interest in the product
- Comment pods, where groups reciprocate engagement regardless of content
- Purchased engagement, which is cheap and increasingly hard to spot from the outside
- A single viral post dragging an otherwise flat average upward
None of these are proof of anything on their own. They are reasons to look at the distribution rather than the average.
An average hides the thing you need to see. Ask for the spread.
What to look at instead
Engagement rate is a useful filter and a poor decision-making tool. Once a creator passes the filter, these tell you considerably more.
Comment quality. Read thirty of them. Real audiences ask specific questions and reference things from earlier posts. Pods and bought engagement produce generic approval - emoji strings, "nice", "great post".
Saves and shares. Saves indicate genuine intent, and shares indicate the content was worth someone's own reputation. Both are far harder to fake than likes and correlate much better with purchase behaviour.
Consistency across recent posts. Ten recent posts with similar performance describes a reliable audience. Nine flat posts and one enormous spike describes a creator who got lucky once.
Audience overlap with your customer. A fitness creator in Pune with a 7% rate is irrelevant if you only ship to Delhi NCR. Location, age band and language decide whether the engagement is reachable at all.
Past collaborations in your category. Someone who has run brand work before knows how briefs, approvals and disclosure work. That is worth real money in coordination time saved.
How we treat it
On WebRectangle, engagement rate is the first metric on every creator card, ahead of follower count. That ordering is deliberate: the product exists on the premise that relevance beats reach, and putting follower count first would contradict it.
But it is a starting filter, not a verdict. Where the data supports it we surface signals - patterns worth a closer look. We do not label creators as fraudulent, because no data available from outside a platform's own systems supports that claim with any confidence, and anyone telling you otherwise is selling something.
The number narrows the field. Reading the comments makes the decision.
Running a campaign soon?
Tell us the product and the audience, and we will come back with creators worth approaching.
Talk to us