Analytics numbers tell you what happens: how many people abandon a form, how many click a button. They rarely tell you how. Heatmaps and session recordings partly close that gap, showing visually where people click, how far they scroll and how they move around the page. They are powerful tools, but easy to misuse and sensitive from a privacy point of view.
What they are
Heatmaps aggregate the behaviour of many visitors on a page and represent it with colours: warmer where clicks or attention concentrate. There are three main kinds:
| Type | What it shows | Useful for |
|---|---|---|
| Click | where people click or tap | elements ignored or mistaken for links |
| Scroll | how far down the page each share of visitors gets | placement of important content |
| Movement | where the pointer goes | very rough hints, desktop only |
Session recordings reconstruct, like a video, the journey of a single visitor: scrolling, clicks, fields filled in, pages visited.
When they are useful
They work best as a diagnostic tool, after the numbers have pointed to a problem. Some typical cases:
- clicks on non-clickable elements: an image or heading people believe is a link;
- rapid repeated clicks on the same spot, a sign of a button that does not respond or responds too slowly;
- important content below the scroll depth reached by most visitors;
- unexpected behaviour on mobile, such as a menu covering content or a button too close to another.
If the funnel shows heavy abandonment on a page, watching twenty or so recordings of that page often reveals the reason within minutes.
How to read them without being misled
- The movement map is not gaze. The pointer and the eyes don't always move together, and on mobile there is no pointer.
- Maps depend on the device. A page has a different layout on desktop and phone; look at them separately.
- Dynamic content confuses. Pop-up menus, carousels and banners change position, and clicks may be attributed to the wrong element.
- A few sessions are not a sample. One curious recording is still an anecdote; you need to see the same behaviour several times before drawing conclusions.
- Seeing is not explaining. A recording shows that a person hesitates; it does not say why. For the why you need user testing or direct questions.
Privacy
This is the most delicate point. Session recordings can capture everything that appears on screen, including text typed into fields and information shown in private areas. Before enabling them:
- mask every input field, not just those that look sensitive;
- exclude private pages: customer area, checkout, profile, messages;
- check in practice what gets recorded, by watching a few test sessions;
- limit retention to a short period and access to a few people;
- assess the legal basis and consent and update the privacy notice, because this is detailed observation of individual behaviour.
Many tools offer masking by default; it still needs checking, because custom fields or content loaded after the page starts can slip through.
Sample instead of recording everything
You don't need to record every visit. A limited sample, or recordings active only on specific pages for a defined period, reduces the data collected, the weight on the page and the time needed for analysis. Performance benefits too: recording scripts are not lightweight.
Targeted use
In my work I use them like this: a precise question, raised by quantitative data; a specific page or step; a limited collection period; a viewing session with notes on recurring behaviour; a hypothesis for a change to be verified with the numbers. Once the question is answered, collection is switched off.
In short
- Heatmaps and recordings show how people behave, not why.
- Use them to diagnose problems the numbers already point to, not to explore at random.
- Read desktop and mobile separately and be wary of movement maps.
- Mask every field, exclude private pages and limit retention and access.
- Record only a sample, on specific pages and for a defined time.
