Which heatmap answers which question?
One honest note before we start: vendors split heatmaps into six or more sub-types, and each sub-type gets its own marketing page. Mtrix ships two — click maps and scroll maps, by device. This page explains what every type is genuinely for, including the ones we don't sell, so you pick the map for your question instead of the question for someone's map.
That's the definition. The mistake that wastes most heatmap time comes next: opening "the heatmap" without deciding which question you're asking. Different map types answer different questions, and reading the wrong one produces confident, wrong conclusions.
The map-to-question table
| Map type | The question it answers | Reach for it when | What it can't tell you |
|---|---|---|---|
| Click map | Where do people actually click — and is it where I think? | CTAs underperform; you suspect people click things that aren't links | Why they clicked, or what happened after |
| Scroll map | How far down do people get? Does anyone reach this section? | Long pages; deciding what deserves the top; "is the fold a problem" | Whether the content they passed registered at all |
| Move (hover) map | Where does the cursor travel? | Rarely — cursor position is a weak, desktop-only proxy for attention, and it means nothing on touch screens | Anything reliable about reading or intent on mobile |
| Rage-click view | Where do people click repeatedly in frustration? | Something feels broken; support tickets mention "the button doesn't work" | What the visitor experienced — for that, watch the session |
| Engagement / zone maps | Which page zone earns the most interaction per view? | Enterprise-style zone reporting across templated pages | The individual behaviour underneath the aggregate |
The first two rows answer almost every question a store operator actually has. The other three exist, and here is the honest context for each.
Click maps: the "are they clicking what I think?" map
A click map earns its place with three findings, over and over:
- The click you designed for isn't happening. The add-to-cart gets a cold spot; a competing link right above it glows.
- Non-link clicks. Visitors click product photos, underlined text and badges that aren't clickable. Every one is a shopper asking for something the page won't give — often the cheapest fix on the list.
- Click order. Not just where people clicked, but what they clicked first — which tells you what the page made them want to do.
Mtrix click maps show density, accuracy, order, non-link clicks and rage clicks, split by device — because your mobile page and desktop page are different pages and their clicks should never be read blended.
Scroll maps: the "does anyone see this?" map
A scroll map answers the question every long product page raises: does anyone reach the section you're arguing about? Depth, attention and exit points show you where the audience thins, which content sits below the point most visitors abandon, and — the classic finding — whether your add-to-cart lives in a zone half your mobile visitors never see.
Two habits keep scroll maps honest. Read them per device, always; mobile and desktop scroll behaviour differ too much to blend. And don't judge your page against a universal scroll benchmark — the "average scroll depth" figures that circulate are mostly unsourced. We keep a sourced, dated set here: scroll depth benchmarks.
The maps Mtrix doesn't ship — and what to use instead
Plainly: Mtrix ships click maps and scroll maps, by device. That's all. No move maps, no standalone engagement-zone analytics. If your requirement is specifically those, other tools ship them — our Shopify heatmap tools comparison covers who.
Why we're comfortable with that line: the questions the other map types answer usually have a better instrument.
- "What are people paying attention to?" — the move-map question. Cursor position is a desktop-only proxy, and most of your traffic is on a phone, where there is no cursor. A scroll map's attention view plus watching real sessions answers it with evidence instead of proxy.
- "Where are people frustrated?" — the rage-click question. Mtrix surfaces rage clicks inside the click map, and — more usefully — detects rage clicks, dead clicks, error clicks and form abandonment as session replay filters. A heatmap shows you where frustration clusters; the recording shows you what it was like, which is the part you can fix.
- "Which zone performs?" — the enterprise-zone question. For a DTC store this is usually the click map plus analytics drill-down wearing a more expensive name.
Heat map analysis: a five-step method
"Heat map analysis" isn't staring at colours until inspiration strikes. The version that produces fixes:
-
Start from a metric, not a map.
A page with high traffic and weak conversion, from your analytics. The map is the second step, never the first.
-
Pick the map for the question.
Clicks misfiring → click map. Content ignored → scroll map.
-
Split by device before reading anything.
The blended map is a fiction — it averages two different pages.
-
Confirm in a handful of sessions.
The map found the where; two or three replays of affected sessions give you the what happened. Aggregate plus instance is the whole method.
-
Ship the fix as a test.
A heatmap finding is a hypothesis. Move the block, fix the dead element — then A/B test it and read the result in revenue, not vibes.