Engagement rate (video)

Engagement rate explained: three formulas, one name, and why the number needs a footnote

Video engagement rate is the percentage of an audience that engaged with a video, and it is the least standardised metric in video reporting. At least three incompatible formulas share the name, and the same video can score 4 percent under one and 62 percent under another without anything changing.

Treat it as a family of metrics rather than a metric. A video engagement rate quoted without its formula is not a benchmark, a target, or a comparison. It is a number with no units.

How is engagement rate calculated? The three formulas in circulation

• Interactions over impressions. Reactions, comments, shares and clicks divided by impressions, times 100. This is the paid-social convention and the one LinkedIn Campaign Manager uses. Typical results run in the low single digits.
• Interactions over followers or reach. The same numerator against audience size rather than servings. Common in organic social and influencer reporting, and it produces a higher number because the denominator is smaller.
• Percentage of the video watched. Seconds watched divided by video length, averaged across views. Wistia, the video hosting platform, defines engagement rate this way, as how much of a video people watch on average. It has no interactions in it at all, and it routinely lands between 40 and 70 percent.

The third is not a variant of the first two. It measures attention, where they measure action. Sharing a column heading does not make them the same quantity.

Engagement rate explained: a worked example

A dealer group posts a 60-second walkaround. It gets 40,000 impressions, reaches 12,000 people, and collects 520 reactions, comments, shares and clicks. Viewers watch an average of 37 seconds.

Interactions over impressions gives 1.3 percent. Interactions over reach gives 4.3 percent. Percentage watched gives 61.7 percent. All three are correct. Put any one of them in a board pack without the formula beside it and you have told your leadership something they cannot act on and cannot verify.

What does engagement actually mean here?

The word covers two genuinely different behaviours, and confusing them is what makes the metric so easy to abuse. Video engagement in the social sense means a public, countable action: a like, a share, a comment. In the one-to-one sense it means a private one: replying, booking, approving work, clicking the call to action.

For a dealership, the second kind is the one attached to revenue. A hundred likes on a walkaround is a marketing result. One customer who approves the work from the page is a repair order. If your engagement rate counts the first and ignores the second, you are measuring applause.

Why a number without its formula is not comparable

Three checks before you accept anyone's engagement rate, including your own tool's:

• What is in the numerator? Some definitions count a play as an engagement, which folds play rate into engagement rate and roughly doubles it.
• What is in the denominator? Impressions, reach, followers, sends or page loads. Each swap changes the answer by multiples, not percentage points.
• Are interactions and attention being mixed? Some dashboards blend both into one composite score with undisclosed weights. Those cannot be compared to anything, including their own previous version after a product update.

Where you cannot resolve the definition, report the raw counts instead. Six replies from 40 sends is a fact that survives translation between platforms. A percentage does not.

What to measure instead, or alongside

Pick the specific behaviour you care about and name it. Reply rate, click rate on the primary button, approval rate, appointment rate. Each one has a single obvious formula and none of them can be quietly redefined by a vendor between quarters.

Write the formula into the report header, not the footnotes, and keep it fixed for at least a full quarter. Most of the movement teams celebrate in engagement rate is a definition changing underneath them, either because a platform updated its counting rules or because somebody rebuilt the dashboard and picked a different denominator.

Then keep the attention measures separate from the action measures. Watch time and the retention curve tell you whether the video held people. View rate tells you whether it was seen. An engagement rate that has swallowed all of these tells you nothing about any of them, which is precisely why it appears so often in vendor marketing and so rarely in a report anyone uses to make a decision.