Heatmaps and Session Recording
A scroll map settles arguments about user behaviour that no meeting has ever settled. Ivan Sp sets up heatmap and session recording tools to replace internal opinion about user behavior with actual observed data. An assumption stated confidently in a meeting is not evidence, and these tools exist specifically to test assumptions that have never actually been checked.
Validating or Disproving Team Assumptions
Internal teams routinely hold confident claims about how users behave, such as “users don’t scroll past the fold,” with no data backing the claim. Heatmap and session recording tools test these claims directly against real visitor behavior.
A claim confirmed by the data is worth acting on with confidence. A claim disproven by the data saves the team from optimizing around a belief that was simply wrong.
What Heatmaps Actually Show
I read aggregate views first. Heatmap tools show where visitors click, how far they scroll, and which elements draw attention across many sessions. The result is a visual pattern rather than a single anecdote. Scroll-and-click mapping shows where traffic actually engages after arriving from a search result, which often differs from where a page’s design intended attention to go.
What Session Recordings Add
Session recordings play back individual visitor sessions, showing where a specific visitor hesitated, moved their cursor, or abandoned a form. Form analysis identifies precisely which field caused a visitor to give up, rather than only that the form underperformed as a whole.
Feedback polls and short surveys layered onto the same tooling add a visitor’s own stated reason alongside the behavioral data. That closes the gap between what a visitor did and why they say they did it.
Baseline Before Change
Heatmap and recording tools get set up as part of an initial audit phase, establishing a baseline before any change gets made to a page. Without a baseline, proving that a later change actually improved behavior is impossible.
Quantitative Data Compared to Qualitative Research
Heatmaps and recordings show what users do at scale, without explaining why a user behaved a certain way. Qualitative methods take over at that point, covered under usability testing. The two get paired rather than treated as substitutes for each other.
Where the Standard Behaviour Advice Breaks Down
- The usual advice
- Nobody scrolls, so put everything important above the fold.
- What holds up in practice
- Scroll behaviour varies enormously by page and by audience, and the claim is testable in two weeks for nothing. On many pages a majority scroll well past the fold when the content earns it.
Consider a team that compresses a long service page to fit above the fold, on the strength of a rule somebody read years ago. The page loses most of its detail.
The scroll map, installed afterwards, shows 71 percent of visitors had been reaching the halfway point. The detail they were reading is now gone.
I install the tool and write the assumption down before looking. Reading the data first reliably produces a memory of having believed it all along.
Design rules applied without measurement remove content that was working. The removal is invisible unless a baseline was captured first.
Four Things Commonly Assumed About Behaviour Data
What does your team believe about visitor behaviour that nobody has ever measured?
Ivan Sp meets these four assumptions whenever a design decision rests on a rule somebody read once.
| Commonly assumed | What actually holds |
|---|---|
| Nobody scrolls below the fold. | Scroll behaviour varies enormously by page and audience. On many pages a clear majority scroll well past the fold. |
| A week of data is enough. | A week can miss an entire weekly cycle. Two weeks is the practical minimum for a representative sample. |
| Session recordings are a privacy problem by nature. | Input masking is configured before collection starts, which prevents form contents being captured at all. |
| Heatmaps explain why visitors behave that way. | Heatmaps report what happens at scale. Motivation comes from usability sessions and from asking people directly. |
| Rage clicks are a browser bug. | Repeated rapid clicking usually means something looks interactive and is not. The finding is a design one rather than a technical one. |
Writing the assumption down before you look at the data is what makes the exercise honest. Reading the data first reliably produces a memory of having believed it all along.
What Each View Reports
- Click map
- Where visitors click, aggregated across sessions, including clicks on things that are not links.
- Scroll map
- How far down the page each percentage of visitors reached.
- Move map
- Cursor movement, which correlates loosely with attention on desktop.
- Session recording
- A replay of one visit, showing hesitation, backtracking, and form abandonment.
- Rage click
- Repeated rapid clicking on one element, usually a sign something looks interactive and is not.
- Minimum useful window
- Two weeks, so the sample covers a full weekly cycle.
- Privacy requirement
- Input masking, so form contents are never captured in a recording.
- Cost floor
- Microsoft Clarity provides heatmaps and recordings at no cost.
A Worked Example: The Assumption and the Scroll Map
Consider a team convinced that visitors never scroll. The figures below are illustrative, written to show what a scroll map does to that claim. No client engagement is described here.
| Page depth | Team assumption | Scroll map reading |
|---|---|---|
| Reached the fold | 100 percent | 100 percent |
| Reached 50 percent depth | Very few | 71 percent |
| Reached the pricing section at 68 percent | Nobody | 54 percent |
| Reached the form at 91 percent | Nobody | 38 percent |
| Rage clicks on the static price table | Not considered | 212 in two weeks |
I treat the final row as the finding. Two hundred people tried to click a price table that was never a link.
A heatmap is a worn path across a lawn. Nobody designed it, everybody made it, and it shows exactly where the paving should have gone.
The assumption in the middle column was not unreasonable. The assumption was simply untested, and testing it took two weeks and no budget.
How Behaviour Data Gets Collected and Read
I set up before changing anything, so a baseline exists to compare against.
- Install the tracking snippet through Google Tag Manager rather than hard-coding it into the theme.
- Enable input masking before collection starts, so form contents are never recorded.
- Confirm the tag fires on every template that matters, using Tag Manager preview mode.
- Collect for at least two weeks, so the sample covers a full weekly cycle.
- Read the scroll map first, and record the depth reached by half of visitors.
- Compare that depth against the position of the main call to action.
- Read the click map for clicks on non-interactive elements, which reveal misleading design.
- Filter recordings to sessions that abandoned, and watch at least ten of them end to end.
- Write findings as observations rather than as fixes, so the hypothesis stays separate from the data.
- Re-collect after any change, and compare against the stored baseline rather than against memory.
How This Work Is Scoped
I scope behaviour research by template count rather than by page count.
- What arrives
- A configured tool, a stored baseline, and a written findings document per template.
- Typical duration
- Three to four weeks, of which two are collection rather than analysis.
- What the client supplies
- Tag Manager access, and agreement on which templates to instrument.
- Included
- Installation, masking configuration, verification, collection, and the findings document.
- Not included
- Design changes, and any interpretation of why a visitor behaved as they did.
- What changes the estimate
- Low traffic, which lengthens collection, and pages behind a login.
- How it is quoted
- After a short scoping call, once template count and traffic are known.
Tools and Primary Sources
The references below cover the tools and the privacy requirement. Each opens in a new tab.
- Microsoft Clarity provides heatmaps, scroll maps, session recordings, and rage click detection at no cost.
- Hotjar adds on-page survey prompts alongside heatmaps and recordings.
- Google Tag Manager deploys the tracking snippet without editing theme files directly.
- Nielsen Norman Group on scrolling publishes research on how attention distributes down a page.
Test One Assumption This Week
I suggest a way to settle a recurring internal argument for free.
- Write down one claim your team repeats about how visitors use the site.
- Install Microsoft Clarity on the page that claim is about.
- Wait fourteen days without looking at the data.
- Open the scroll map and the click map, then compare both against the written claim.
Writing the claim down first is what makes the exercise honest. Reading the data first almost always produces a memory of having believed it all along.
Common Questions
Do heatmap tools slow down page load speed?
A well-implemented tracking script adds minimal overhead. The tool gets configured to load asynchronously, so it does not block the page’s own content from rendering.
How much traffic does a page need before heatmap data becomes useful?
Enough sessions to see a genuine pattern rather than noise from a handful of visits, typically at least a few hundred sessions for a meaningful read on one page.
Is session recording a privacy concern?
Recordings capture behavior, not personal identity. Sensitive form fields such as passwords or payment details get masked automatically by standard tooling before any recording is stored.
Do session recordings create a privacy problem?
They can, which is why input masking is configured before collection starts. Masking prevents form contents from being captured, and recordings should still be covered by the site’s privacy notice and its consent mechanism.
Discuss This Work
The first call is a free 30-minute consultation with no pitch attached. You receive an honest assessment of where the site stands and what can realistically be built.