How to read heatmaps: click, scroll and move maps for more conversion
6 min read
Heatmaps are the most intuitive view in web analytics — and the most frequently misread. Red blobs look like proof, but at first they are just concentrations. What they mean depends on which map you are looking at, how many sessions sit behind it, and what you put next to it.
This guide explains the three map types — click, scroll and move — the way they are technically built, and derives from that how to read them without fooling yourself. It ends with what this is actually about: turning observation into a conversion improvement.
The core principle: aggregate, not individual observation
A heatmap is an aggregate over many sessions — not tracking of individuals. visz.ai never follows a single visitor around your site; it sums interactions across visits and renders the result on top of the page.
From that follows the first reading rule: a heatmap gets more meaningful as more sessions arrive. A page needs roughly 30 sessions for a reliable scroll curve — below that, visz.ai shows an insufficient-data hint instead of a seemingly precise curve.
And the second: segment before you judge. Desktop and mobile visitors click and scroll differently — the filter bar builds you a heatmap for a single device class, traffic source or segment. A blended picture of both worlds leads to conclusions that are true for neither.
Click maps: where intent becomes visible
Click heatmaps show where attention and intent concentrate — what visitors actually click. The most valuable find is often not the big blob on your button, but the blob next to it: clicks on non-interactive elements that visitors expect to be clickable.
An image, a heading or an icon that repeatedly collects clicks is unpaid design feedback: either a link belongs there, or the element should look less clickable. The reverse holds too: if your primary call-to-action gets fewer clicks than decorative elements near it, it is competing visually — and losing.
Scroll maps: fold line, 50% drop-off, attention cliff
A scroll heatmap is built from the maximum scroll depth per session: for each 5% depth band (21 bands from 0 to 100%), the curve shows the share of sessions that scrolled at least that far. If the page fits entirely in the viewport, visz.ai honestly shows “Nothing to scroll” instead of a gradient.
Three markers carry the interpretation:
- Fold line — this is how far visitors see the page without scrolling, derived from the median viewport height. Everything that is supposed to sell competes for the space above it.
- 50% drop-off — the page depth at which only half of your visitors keep reading. Content below it structurally reaches a smaller audience; an important CTA below this line has had its audience halved before it even becomes visible.
- Attention cliff — the band with the steepest drop between two consecutive depths. Directly above the cliff sits the content after which interest breaks off: the most precise place to rethink order or length.
Move maps: hesitation zones and drain zones
Move heatmaps use mouse movement as an attention proxy. The tracker samples cursor positions (throttled to ~120 ms) in page coordinates — the same reference frame as click maps — and aggregates them server-side into 20-pixel cells. Important: move maps are desktop-only, because touch devices have no hover; for mobile attention they are simply the wrong source.
From dwell time and speed, visz.ai derives two zone types: hesitation zones — high dwell time with slow movement; visitors pause here, often because something is unclear or a call-to-action is missing. And drain zones — the cursor rushes straight through; when such a zone sits over a clickable element, it is likely being skimmed over.
Move data is a qualitative signal, not proof. There is deliberately no hard minimum threshold — on thin data you get a hint, and the map grows more meaningful as more sessions arrive. Treat hesitation zones as questions to your page, not as answers.
How to read heatmaps without fooling yourself
The most common misreadings can be avoided with five checks:
- Segment first, conclude second: look at desktop and mobile separately — a blended picture distorts both.
- Check the data base: below ~30 sessions, a scroll curve is anecdote, not evidence.
- Combine the maps: few clicks on an element below the 50% drop-off usually mean “hardly anyone sees it” — not “nobody wants it”. Only the scroll map makes the click map readable.
- Don't transfer move to mobile: move maps only arise with a fine pointer (mouse/trackpad) — mobile attention belongs in scroll maps.
- Verify suspicions on the individual case: real sessions sit behind every pattern — a session replay behind the drop-off shows what the curve only hints at.
From observation to conversion
A heatmap only becomes valuable through the change you derive from it: the click-collecting image becomes a link, the CTA moves above the 50% drop-off line, the section before the attention cliff gets tightened, the hesitation zone gets clearer wording. Then: same map, same segments, new time window — and the comparison shows whether the thesis held.
In visz.ai you find all three views under Heatmaps → Clicks, Scroll and Move, segmentable by device, time window and traffic source. Like everything in the tracker, heatmaps are consent-gated — nothing is captured before consent, and “Do Not Track” wins — and the data is EU-hosted in Germany. visz.ai is free.
See it on your own website
Session replay, funnels, heatmaps and frustration signals. GDPR-first, EU-hosted in Germany, and visz.ai is 100% free.
Start for free