One of the most common mistakes in analytics is trying to reduce engagement to a single number.
- Time on page.
- Scroll depth.
- Clicks.
Pick one… build a dashboard… and move on.
It feels clean. It feels measurable. It feels actionable.
But it’s also misleading.
Because engagement doesn’t behave like a single metric.
It behaves like a pattern.
A combination of signals that only make sense when you look at them together.
Think about this: 👉 A user spends 2 minutes on a page but never scrolls
Was that engagement?
Maybe. Or maybe the tab was open in the background.
👉 Another user scrolls through the entire page in 10 seconds
High engagement?
Not really. That’s likely skimming… or even accidental scrolling.
👉 Another clicks on an element
That sounds like strong intent… right?
But without context:
Was it curiosity? Confusion? Misclick?
You don’t know.
Individually, these signals are weak.
They’re fragments.
And yet… most analytics systems treat them as if they were definitive answers.
That’s where the disconnect begins.
Teams start building interpretations on top of incomplete signals: “This content is engaging” “This section is working” “This feature is driving value”
But those conclusions are often based on one-dimensional views of behavior