Engagement rate calculator

Engagements is the platform's own count, and it is the field that decides what the answer means. The event list behind it is usually long — reactions, comments, shares, saves, taps on the profile name, image expansions, short video views, clicks of every kind — and it mixes deliberate acts with accidents of scrolling. Take the narrowest definition your report offers if the question is whether people cared, and use the same one every time.

Impressions is the number of times the ad was served.

Reach is optional and is the number of distinct people it was served to. Supplying it produces the second rate, which is the one most people think they are already looking at.

Ad spend is optional and turns the rate into a price.

A rate against impressions and a rate against reach are different questions with the same label.

Per impression: of the times this ad appeared, what share produced a response. Per person: of the people who saw it at all, what share responded at least once. The two differ by exactly the frequency, so the same campaign at a frequency of four reports a per-person rate four times its per-impression rate. Neither is wrong and they are not interchangeable, which is why a report that does not say which one it used is not a report.

The gap between them is also informative on its own. A per-person rate far above the per-impression rate means the ad is being served repeatedly to a small audience, and the repeated impressions are producing far fewer responses than the first ones did. That is fatigue, visible in a metric most accounts read as a creative score.

Engagement measures match: whether the creative and the audience belong together. It moves within hours, it moves for reasons you can act on, and it is the earliest signal available that an ad is landing.

It is also the single easiest number on a platform to move dishonestly. Outrage, curiosity gaps, promises the landing page does not keep and images unrelated to the product all raise it. Every one of those buys responses from people with no intention of buying anything, and the campaign gets more expensive per outcome while the metric everybody watches improves.

There is one place where the number pays for itself directly. Most auctions discount inventory for ads people respond to, because a response is a better outcome for the platform as well. So a genuinely better creative lowers the price of the impression at the same time as it raises the response to it — a lever that no bid change can reach, and the reason this metric is worth watching even on a campaign whose objective is a purchase.

Different event lists. One platform's engagement includes three-second video views; another's does not. The rates are not comparable and the gap can be several multiples.

Different denominators. Per impression against per person, as above.

Different formats. A video, a carousel and a static image offer different numbers of things to do. A format change moves the rate without anything about the audience or the offer changing.

Different placements. A full-screen placement and a small in-feed unit produce different rates for the same creative, and a campaign on automatic placements is an average across a mix that drifts.

Warm against cold. Retargeting an audience that already knows you produces rates a prospecting campaign will never reach. Averaging the two produces a number that describes neither, and it always flatters the account.

Different objectives. A campaign the platform is optimising toward engagement is being shown to the people most likely to engage, so its rate is partly a property of the optimisation rather than of the ad. Comparing it to a conversion campaign's rate compares two different selections of people.

Start with the ask. A surprising number of ads with poor engagement never invite a response: no question, no offer, no reason to stop. The cheapest fix is usually to make the ad ask for something.

Then the first moment. On a feed the opening frame or the top third of the image decides whether anything else is seen, and most engagement problems are hook problems wearing a different name.

Then the match. A rate that is poor across every creative is more likely an audience problem than a creative one — the ad is being shown to people with no reason to care, and no amount of production fixes that.

Then frequency. A rate that decayed week by week on an unchanged creative is not a creative that got worse; it is an audience that has seen it enough. Widen the audience, extend the flight, or change what is being repeated.

And keep the cost per engagement in view rather than the rate alone. A campaign whose rate fell while its cost per engagement fell is reaching more people more cheaply, which is usually a better outcome than the rate suggests — and the ratio on its own would have reported it as a decline.