Heatmaps: what they show, and what they do not prove
A heatmap shows where visitors click, how far they scroll and where their cursor travels. It never says why. The Nielsen Norman Group measures that 74% of viewing time happens in the first two screenfuls of a page, which decides where to place what sells.
- More than 42% of viewing time falls in the first fifth of a page, and more than 65% in the first two fifths.
- People scroll more than they used to, but the drop in attention after the first screenful stayed the same between 2010 and 2018.
- A heatmap describes a behaviour, never a reason. It serves to form a hypothesis, not to settle a budget decision.
- The Nielsen Norman Group recommends cross-referencing heatmaps with session recordings, funnels and form analytics before concluding.
- The F-shaped pattern is described by its authors as a rough general shape, not a uniform behaviour. Citing it as a law is a mistake.
On this page
Heatmap
A heatmap is a visual representation of visitors' aggregated behaviour on a page, produced by a measurement tool. It takes three main forms: the click map, which shows where people tap, the scroll map, which shows how far they go down, and the movement map, which tracks the cursor. It describes what happened and never explains why, which makes it a hypothesis-forming tool rather than a decision instrument.
What they actually measure
The tool is inexpensive, often free, and that is precisely what makes it dangerous: a management team sees visual proof in it and makes a budget decision on a coloured image. The data is good; the reading is the problem.
The most useful figure comes from the Nielsen Norman Group, which redid its eye-tracking measurements in 2018 to check whether long pages had changed habits. The result: people scroll more than before, but attention still drops sharply after the first screenful. About 74% of viewing time concentrates in the first two screenfuls, more than 42% in the first fifth of the page and more than 65% in the first two fifths.
What actually decides something is not the percentage itself, it is its stability. In 2010, the share above the first screenful was 80%. Eight years later, with longer pages and a cleaner design, the pattern holds. A business that places its selling argument a third of the way down the page therefore pays for content the majority of its visitors will never look at, however good the writing.
What to settle before buying a behavioural measurement tool
The cost of a heatmap tool is low, which hides the real cost: the analysis time and the decisions made on a too-quick reading. A misread heatmap gets a working page rebuilt.
- What precise question do we want to answer, and can a heatmap do it?
- How many visits does the page we want to observe receive, and over what period?
- Who will read these heatmaps each month, and how much time will they spend on it?
- Are our pages similar enough to be grouped into templates?
- If the tool changes no decision within six months, what do we stop funding?
The useful answer starts from a precise question and states in advance what a heatmap will not be able to settle. An evasive answer sells the tool before the question, presents a heatmap as proof, or draws a conclusion from a low-volume page.
Deciding whether your site loses sales to friction, to placement or to the offer is a trade-off that requires cross-referencing several sources. A 90-minute consultation settles it, with a written summary your team can execute.
The three reading traps
They come up in almost every engagement we take over, and each one costs a pointless project.
Confusing the behaviour with the reason
A cold zone can mean that nobody sees the element, that nobody understands it, or that nobody needs it. The three causes call for three different fixes, only one of which is a move. Deciding without knowing which one leads to moving a button nobody wanted to click, then starting over.
Concluding on too few visits
A heatmap built on two hundred visits shows clusters with no statistical value. The Nielsen Norman Group stresses the need to account for sample size and to group pages into templates to reach a sufficient volume. A monthly report on a low-traffic page mostly measures chance.
Taking the F-shaped pattern for a law
Its authors themselves describe it as a rough general shape rather than a pixel-perfect uniform behaviour. On a product page, the eye also goes to the price and the buy button, outside the pattern. Designing a page by applying the F to the letter without looking at your own data means applying an average to a specific case.
What stays in-house: the business question you want to answer, the list of pages that carry revenue, and the authority to change those pages. A heatmap without a precise question produces a pretty image and no decision. What gets delegated: the setup, grouping pages into templates, cross-reading with recordings and funnels, forming hypotheses and prioritizing by expected effect. A business that names three precise questions gets more value than a business that installs the tool across the whole site.
How to use them without going wrong
The rule fits in one sentence: a heatmap tells you where to look, never what to conclude. The Nielsen Norman Group puts the same idea forward by recommending you cross-reference heatmaps with session recordings, conversion funnels and form analytics before trusting an observation.
| What you observe | What it suggests | What to check elsewhere |
|---|---|---|
| Scrolling that stops high | The selling argument is too low | The conversion rate by device |
| Clicks on a non-clickable element | An unmet expectation | The session recordings |
| A cold zone on a primary button | A visibility or desire problem | The funnel and the button label |
| Repeated clicks in the same spot | A slowdown or a block | The page response times |
| Abandonment in a form | A field that snags | The field-by-field analysis |
Heatmap tools record the behaviour of identifiable people and often drop tracking cookies. In Quebec, this collection falls under personal-information protection rules, a subject covered in Law 25 and marketing. Installing a tool without checking consent creates a regulatory risk for a modest analytical gain.
The order of the fixes to make before you even measure is detailed in conversion rate optimization, the experimentation method in the calculation to run before testing, the continuity between the ad and the page in continuity between the ad and the page, and measurement reliability in measuring conversions properly.
Turning a site into an asset that sells rather than a storefront is at the heart of the Improve your site's conversion goal.
Already running a marketing team? See how we plug in as reinforcement on conversion rate optimization.
Frequently asked questions about heatmaps
Does a heatmap prove anything?
No. It describes aggregated behaviour and never gives the reason for it. A cold zone can mean the element is not seen, not understood or not wanted, and those three causes call for three different fixes. A heatmap tells you where to look, not what to conclude.
Where should you place what matters on a page?
High. The Nielsen Norman Group measures about 74% of viewing time in the first two screenfuls, more than 42% in the first fifth of the page and more than 65% in the first two fifths. People scroll more than before, but the drop in attention after the first screenful stayed stable between 2010 and 2018.
How many visits does a heatmap need to be readable?
Enough that the clusters are not chance. The Nielsen Norman Group recommends accounting for sample size and grouping similar pages into templates to reach a sufficient volume. A heatmap built on a few hundred visits mostly measures noise.
Is the F-shaped pattern reliable?
It is a rough general shape, described as such by its authors, not a pixel-perfect uniform behaviour. On a product page, the eye also turns to the price and the buy button, outside the pattern. Designing a page by applying the letter without looking at your own data amounts to applying an average to a specific case.
- Nielsen Norman Group, Scrolling and Attention, 2018 eye-tracking measurements compared with those from 2010, accessed July 2026.
- Nielsen Norman Group, F-Shaped Pattern for Reading Web Content, the original research and the authors' caution on the pattern's scope, accessed July 2026.
- Nielsen Norman Group, Heatmaps and Visual Analytics for Website Optimization, on sample size, grouping into templates and cross-referencing methods, accessed July 2026.
Gabriel almost always takes your first call and carries out your audit. He builds the strategy starting from your growth goal: where to put your budget, which market to test and how to connect each lead to a real sale in your CRM. He mainly leads engagements for three goals: Optimize the profitability of your digital campaigns, Develop a new market, and Generate demand and growth. With Geneviève, he also works on organic search (SEO), AI visibility (GEO) and conversion rate optimization (CRO). The sales a Google Ads or Meta Ads campaign brings in depend on the page that receives the click. He writes mainly about marketing strategy, paid advertising and measurement.
About Falia →
SEO
Paid advertising
Conversion
AI visibility
Web design