flowpeekresources / guides / UX
How to read heatmaps and turn observations into store improvements
You'll plan one improvement to your store and how to check its result.
Short answer
The key answer
A heatmap shows the distribution of recorded interactions on a selected page. Set the device, map type and period, then check specific areas in the live store. An intense color doesn't prove interest or sales; combine the result with the funnel, events and a page test.
- Choosing the right page, device and map type
- Careful interpretation of clicks and scrolling
- One improvement idea and a plan to check it
1. Come with a question about a specific page.
Pick a page that matters for purchases: a category, a product or a cart step. Start with a question, for example whether users try to open the shipping information. Without a question like this, it's easy to stare at colored spots and never reach a decision. A good topic comes from the funnel, a customer report or something your team noticed.
Check that the store is collecting heatmap data. The “Collect heatmap data” toggle controls this; switching it on starts collecting new data and doesn't recreate history. Recording covers visits within the available tracking and consent scope. An empty map can mean there's no data for the selected filters, not a lack of interest in the page.
2. Set the page, device and period.
In the Heatmaps module, choose Desktop (computers) or Mobile (phones). Then pick an item from the “Page” list, the map type and the period: 7, 30 or 90 days. Click “Load heatmap”. When comparing views, change one element at a time. That way you know whether a difference comes from the device, the time range or the type of interaction.
The screenshot shows a real Clicks view for desktop and 30 days. The background is a simplified view drawn by the app, so we don't attribute clusters to the cart or shipping. The numbers next to the page selector and the metrics after loading can describe different ranges. To describe the selected view, use its own metrics.
3. Read clicks as actions, not intentions.
A more intense, golden area indicates a higher density of recorded interactions in that view. It doesn't automatically mean more buyers. Several clicks can come from the same person, and a repeated action can stem from interest as well as from a problem with an element.
When you spot a cluster, open the actual store page on the same type of device. Find out what element sits in that spot and what it does. If it looks like a button but doesn't respond, you have a hypothesis to test. The map alone, however, doesn't confirm what the user expected or whether the click ended in a purchase.
4. Check scrolling and cursor movement separately.
The “Scroll” mode helps you see recorded activity at different heights of the page. Use it when important information sits far from the top. Don't derive an exact percentage of people who read the content from the color alone. Scrolling past a section doesn't prove the user noticed or understood it.
Mouse movement and Hover describe the behavior of the cursor, the on-screen pointer. They aren't eye tracking. On a phone, interaction looks different than on a computer, so don't treat these views as interchangeable. When comparing periods, also check that the page layout hasn't changed: the same spot can correspond to a different element.
5. Connect the map with behavior after the click.
If your question is about a button, check the available event or the next funnel step. The heatmap helps you find an area of attention, and the event report gives the scale of the recorded action. Don't assume every click on a point of the map automatically has its own event name. The scope of such actions depends on the setup.
Write down what you actually know: people click in a given spot, the visible information sits lower down, or a given step is recorded less often. Call the cause a hypothesis. You can check it with a hands-on page test, an available journey and the relevant report. Each of these sources answers a different question.
6. Make one change and write down how you'll assess it.
A good improvement has a specific reason: clarifying a label, moving information or removing an element that blocks an action. Write down what should become easier and how you'll know. The fact that the colors look different after the change isn't a sufficient measure of success.
Go back to a comparable device, page and period. Account for changes in traffic and promotions. Look at the relevant event or funnel step and repeat the store test. If the offer changed at the same time, don't credit the whole result to moving a button. Also note observations that didn't confirm the original hypothesis.
Common mistakes
Color means attention or a purchase
The map shows interactions. Motivation, understanding of the text and sales results have to be assessed by other methods.
Comparing maps of different page layouts
After a design change, the same area can represent a different element. Note the release date and the observation range.
For your next review
Checklist
- I have a question about a specific page.
- I've checked data collection, device, type and period.
- I don't equate clicks with people or attention.
- I've checked the element directly in the store.
- I have a hypothesis, one improvement and a way to assess it.
Now your store
Turn one observation into an improvement brief
- 1
Choose a page and note the current heatmap range.
- 2
Point out the area to check without adding the customer's motivation.
- 3
Reproduce the action in the store and choose a supporting metric.
From report to decision
Practice example — a hypothetical situation
This is a suggested way of working, not a description of the clusters visible in the screenshot.
- Observation
- Clicks on a section heading that looks expandable
- Hypothesis
- The design suggests an action the element doesn't perform
- Check
- A usability test and a fix to the element's design or behavior
The brief covers one element. It doesn't promise higher sales based on the heatmap alone.
Questions and answers
Does a heatmap show where the user was looking?
No. Clicks, cursor movement and scrolling aren't eye tracking. Don't treat them as confirmation that information was read.
Will turning on heatmaps recreate last month?
No. Collection covers new, available interactions. Selecting 30 days doesn't create data from before the feature was switched on.
Why shouldn't I compare colors alone?
Intensity is shown relative to the given view. Align the ranges and check the numbers and the relevant actions before you assess a change.
From knowledge to report
Open the right module.
Heatmaps and session recordings (coming soon)
Coming soon: click and scroll heatmaps, plus session recordings with AI analysis. Sign up for the beta and be one of the first to try the module.
Conversion funnel
The conversion funnel shows each step of the purchase: product, cart, checkout form and purchase. See how many visits reach each one and choose where to start making improvements.
Event analytics
See how often customers add products to the cart, click key elements and submit forms — how many people do it and how that changes day by day.
Read the next guide.
All guidesConversion / 7 min
Conversion funnel: how to find the problem with mobile purchases
Pick a purchase step, compare devices and check user journeys. Plan a specific checkout test instead of guessing why conversion is lower.
Analytics / 7 min
How to read the Flowpeek dashboard and pick your first topic to analyze
Learn the difference between a visitor, a page view and an order. Read each result against the period, your own goals and the detailed reports.
Setup / 7 min
Setting up Flowpeek: how to prepare your store and sign off the setup
Gather your store details, check the first visit and order, agree on the data scope. A simple sign-off checklist — no self-installation needed.
Free · 30 minutes · online
GA4, your store and your ads show different numbers? We'll check it with you.
We compare one month of sales across your tools and point out three fixes to your measurement. No strings attached. What we check
30 days on your own data
Check what happened next.
Go back to events, the funnel or journeys and assess whether the observation matters for a specific action in the store.
- No fixed-term contract
- We set everything up for you in 48 h
- App servers in Germany (EU)
- Cancel anytime
30 days for PLN 0 · first payment after the trial · cancel anytime
See pricing after the trial