The short version

Engagement rate is engagements divided by impressions or reach, where engagement bundles together likes, comments, shares, saves, clicks, and sometimes video views.

javascript
Engagement rate = (engagements / impressions) x 100

The problem is the bundling. A like and a save are not the same signal, and the cheap actions are far more numerous, so they dominate the number. Engagement rate is consequently the easiest performance metric to raise without improving business results.

Why the bundle misleads

Rank the actions by what they cost the person:

ActionEffortWhat it indicates
LikeAlmost nonePassing approval
ViewNone, often automaticPresence
CommentModerateReaction, not always positive
ShareModerate to highWillingness to be associated with it
SaveHigh intentIntention to return
ClickDirectInterest in what is behind it

A post with 900 likes and 2 saves and a post with 100 likes and 80 saves can report similar engagement rates. The second is far more valuable and the metric hides it.

This is why engagement rate is a starting question rather than an answer: the useful information is the composition, not the total.

Where it genuinely helps

Early creative signal. Before conversion volume accumulates, engagement gives a rough read on whether a concept lands. It is directional, not conclusive.

Comparing creatives in the same format and placement. Same bundle, same conditions, valid comparison.

Detecting negative reaction. A comment spike with poor click-through often means the ad provoked rather than persuaded. Read the comments rather than the count.

Organic and paid social where reach depends on response. Here engagement has a mechanical effect on distribution, which makes it worth something in itself.

Where it misleads

As a performance target. Optimizing for engagement selects for content that provokes reaction, which is not the same as content that sells. Controversy engages well and converts badly.

Across formats. Different formats have structurally different engagement rates. Comparing a video's rate to a static image's is comparing two bundles.

As a proxy for brand health. High engagement with negative sentiment is worse than low engagement.

What to look at instead

  • Saves and shares specifically. These cost the person something and mean more.
  • Comment sentiment, not comment count.
  • Click-through rate, when the goal is traffic. It is unbundled and unambiguous.
  • The downstream metric. Engagement is upstream of everything that matters.

Practical rule

Break the bundle before you use the number. If you cannot see the composition, engagement rate tells you that something happened, not whether it was worth anything.

Related reading

Automate marketing in less than 2 minutes