Somewhere in most content reports there's a line like "average engagement rate: 4.2%" computed across every platform the brand posts on.
That number is not a summary. It's an artefact of the mix.
The counting rules are different
The core problem is that platforms do not agree on what a view is, and they never promised to.
They differ on how long something has to be watched before it counts, whether a replay counts as another view, whether the number is impressions or plays, whether autoplay in a feed qualifies, and how a scroll-past is treated. Each platform documents its own definitions in its analytics help pages, and each changes them from time to time without harmonising with anyone else.
So a "view" is a platform-specific unit, like a currency. Averaging views across platforms is averaging dollars and yen because both are numbers.
The same applies to the numerator. What counts as an engagement differs too — which reactions exist, whether saves are reported, whether shares to private messages are counted. You're dividing an inconsistent numerator by an inconsistent denominator and reporting the result to one decimal place.
Why the blended number moves for the wrong reasons
Even setting definitions aside, a blended average is dominated by whichever platform contributed the most volume.
This is the failure mode that matters most, because it's invisible. A metric that moves when your posting mix moves will get celebrated as improvement and investigated as decline, and in both cases the team will look for the cause in the content.
The audience isn't the same either
The counting is only half of it. The other half is that the audiences behave differently on purpose.
Discovery-led feeds show content mostly to people who don't follow you. Subscription-led surfaces show it mostly to people who do. A follower converts to engagement at a much higher rate than a stranger, so a platform that shows you to strangers will produce a lower rate from identical content.
That means a lower rate on a discovery platform can accompany far more actual reach and far more new audience. Judged by the blended average, the better outcome scores worse.
Intent differs too. Someone searching is further down the funnel than someone scrolling. Comparing a search-driven surface to a scroll-driven one on engagement rate compares two different psychological states, not two pieces of content.
What to do instead
Report per platform. Always. One row per platform, never a total row. If someone insists on a single number, give them the one that survives aggregation — total reach, or total engagements — and label it as a volume figure, not a quality one.
Benchmark each platform against its own history. Your YouTube performance this month against your YouTube performance last quarter. That comparison is valid because the definitions are held constant. This is also the only benchmark that knows your audience, which is why outliers are defined against your own distribution.
Compare like with like when you compare rivals. A competitor's YouTube against your YouTube. Same platform, same window, same metric — that's what makes competitor research load-bearing rather than decorative.
Index rather than average, if you need one line. Express each platform's current performance as a percentage of its own trailing baseline, then look at the set of index numbers. "YouTube 118, Instagram 94, TikTok 103" is honest and readable. It compares each platform to itself and puts the results side by side without pretending they're the same unit.
The comparison that is legitimate
Cross-platform comparison is fine — as long as what you're comparing is directional, not numeric.
Useful cross-platform questions:
- Which platform is growing fastest, each measured against its own baseline?
- Does a topic that performs well here also perform well there?
- Which platform sends the most traffic to the thing we care about?
- Where does our audience actually spend time?
Each of those compares patterns or absolute outcomes, not rates. Patterns travel. Rates don't.
The most interesting version is the disagreement: a topic that works on one platform and fails on another is telling you something about format fit, and it's a genuine finding — see comparing YouTube, Instagram and TikTok honestly for how to run that read without smuggling a blended average back in.
What about "engagement rate" benchmarks in industry reports?
Treat them as conversation, not input.
They aggregate across account sizes, categories, content types and platform definitions, and they're usually computed on whatever sample the publisher could access. They tell you roughly what order of magnitude is normal. They cannot tell you whether your 3.1% is good, because they don't know your denominator, your category or your audience mix.
Your baseline is your content. It's the only benchmark that controls for the things that matter.
Where Acumin fits
Snapshots are per platform by construction, and the comparisons are same-platform against your own trailing history. Where numbers come from a platform's own API for an account you've connected, they're labelled measured; where they're derived from public data, they're labelled as estimates — and the two are never averaged together into one confident figure.
There's no blended cross-platform engagement rate anywhere in the product. That's a deliberate omission rather than a missing feature.
How to use this today
Open your most recent content report and find any number computed across more than one platform.
Work out what would happen to it if you kept performance identical and simply posted twice as much on your weakest platform. If it moves, it's measuring your mix.
Split it into one row per platform. The report gets longer and starts being true.
Related: Engagement-per-view, explained covers what the rate is good for once you're computing it correctly. Measured vs reported is the distinction between a number from an API and a number from a claim.