The Complete Shopify Heatmap Guide for 2026

Heatmaps are the most installed and least used tool in ecommerce analytics. A merchant adds one, looks at a pretty orange blob over the hero image, says "interesting," and never opens it again.

The problem is not the tool. It is that a heatmap answers a narrower question than people expect, and answering it usefully requires segmentation most merchants skip. Done properly, heatmaps are the fastest way to find out why a page underperforms. Done casually, they are wallpaper.

The four types and what each one is for#

"Heatmap" covers several distinct visualisations. They answer different questions and confusing them is the origin of most bad conclusions.

Click maps (tap maps on mobile). Where people click, aggregated. Best for: finding elements that get attention you did not expect, and elements you built that nobody touches. On mobile these are tap maps, and mobile behaviour differs enough that you must view them separately.

Scroll maps. How far down the page visitors get, usually as a gradient with a marked point where some percentage of visitors have dropped off. Best for: discovering that your reviews, specs, or FAQ are effectively invisible. Covered in depth in Shopify scroll depth analytics.

Move maps (desktop only). Cursor movement, used as a rough proxy for attention. Treat with caution — the correlation between cursor and eye is real but loose, and it does not exist at all on mobile, which is most of your traffic. Useful as a hint, never as evidence.

Attention or engagement maps. Time-weighted, showing which regions stayed in view longest. Better than move maps for "what did they actually look at", and unlike move maps they work on mobile.

A fifth category is not really a heatmap but shows up in the same tools and matters more than any of them:

Error and friction maps — rage clicks, dead clicks, and thrash. These localise actual broken things rather than showing distributions. If your tool has them, look at them first. They have a much higher fix rate per finding than any distribution map. DynoWeb's heatmaps put these on the same view for that reason.

What a heatmap can and cannot tell you#

It can tell you: which elements attract clicks, which are ignored, how far down the page attention reaches, and where people click things that are not clickable.

It cannot tell you: why. A heatmap is a distribution, not an explanation. It shows you where to point the microscope, and the microscope is session replay.

The standard workflow is: heatmap to localise, replay to understand, test to confirm. Skipping the middle step is how merchants end up "fixing" things based on a colour gradient.

Setting up so the data is actually usable#

Sample enough sessions. A heatmap built on a few dozen sessions is noise with a nice colour scheme. There is no universal minimum, but a practical rule is: if adding another day of data visibly changes the picture, you do not have enough yet. Keep collecting until the pattern is stable.

Separate mobile and desktop. Always. This is the single most important rule. Layouts differ, fold position differs, and tap behaviour differs from click behaviour. A blended heatmap on a responsive Shopify theme is close to meaningless, because the same screen coordinate is a different element on different devices.

Pick pages, not the whole site. Heatmaps are per-page instruments. Start with your top three landing pages by entry traffic, your top product page, your cart, and your collection page. That covers most of the decision surface.

Watch out for template aggregation. Many tools group all product pages into one heatmap by URL pattern. That is useful when your product pages are structurally identical and misleading when they are not — a product with four images and one with twelve will not align. Check what your tool is aggregating before you interpret.

Account for dynamic content. Personalised blocks, A/B tests running concurrently, popups, and lazy-loaded sections all shift element positions between sessions. If a heatmap shows clicks in empty space, this is usually why.

How to actually read one#

Work through these questions in order.

1. Is the primary action getting the clicks? On a product page, add-to-cart should dominate. On a collection page, product cards should. If something else is winning — a navigation link, an image, a promotional banner — your page's attention is going somewhere other than the conversion path.

2. What is getting clicked that is not clickable? This is the highest-value pattern on any click map. Shoppers tap product images expecting zoom, tap specification rows expecting expansion, tap a review count expecting to jump. Every one of those is a shopper telling you what they expected the page to do. Sometimes the fix is making it interactive. Sometimes it is making it visually not look interactive.

3. What did you build that nobody touches? Carousels past the first slide, tabs past the first tab, filters, mega-menu sections, the third call to action. Cold elements are candidates for removal, and removal is usually the cheapest conversion improvement available.

4. Where does the scroll map die? Find the point where most visitors have stopped. Then ask what is below it that you believed was doing work. Reviews and shipping information are the two things most commonly stranded below the drop-off point.

5. Is the fold doing its job? On mobile especially, check that the price, variant picker, and buy button are visible without scrolling — and check it with your announcement bar, cookie banner, and chat widget all present, because that is what real visitors see.

The segments that make heatmaps worth the effort#

An unsegmented heatmap averages away every interesting finding. These are the cuts that reliably pay:

• Device. Non-negotiable, as above.
• New versus returning. Returning visitors skip your explanatory content and go straight for the buy box. New visitors need it. A blended map shows neither behaviour clearly.
• Converted versus not converted. The most underused segment in the entire discipline. Compare what buyers did against what non-buyers did on the same page. The difference is your hypothesis.
• Traffic source. Paid social visitors land cold and need the value proposition. Branded search visitors already know you and want the price and the button. Same page, two jobs.
• Landing page versus mid-session. Someone who entered on a product page behaves differently from someone who arrived via a collection.

The converted-versus-not comparison deserves emphasis. If buyers overwhelmingly opened the size guide and non-buyers did not, you have found something specific and testable: make the size guide more prominent. That is a hypothesis with a mechanism, which is worth more than ten observations about where the orange is.

Five mistakes that waste the tool#

1. Reading a heatmap as an explanation. It localises; it does not explain. Always follow to replay before you decide what to do.

2. Blending mobile and desktop. Produces a map that describes no real user.

3. Drawing conclusions from a small sample. Random clicks look like patterns to human eyes. This is what we are built to do.

4. Trusting move maps as attention. Cursor position is a weak proxy on desktop and does not exist on mobile. Use engagement maps if you need attention.

5. Optimising the wrong page. A brilliant fix to a page with 200 monthly sessions is worth less than a mediocre fix to one with 40,000. Rank by traffic times drop-off before you open any heatmap at all.

A four-step routine that produces changes#

1. Pick the page by arithmetic. Largest traffic multiplied by largest drop-off. Not the page that bothers you most.
2. Look at friction first. Rage clicks and dead clicks give you concrete, fixable bugs. Then click and scroll maps for the distribution story.
3. Watch five to ten replays of sessions that failed on that page, filtered to the segment your heatmap flagged. Write down what you actually saw, not what you assumed.
4. Ship one change and prove it. Test if you have the volume (how to A/B test on Shopify); otherwise change one thing, log the date, and compare year over year.

Then go again. The merchants who get value from heatmaps are the ones who run this routine monthly, not the ones who bought the best tool.

Where heatmaps fit#

Heatmaps are a diagnostic instrument in a larger loop: find the leak, understand it, fix it, prove the fix was worth it. They are excellent at the first step and useless at the last three.

If you want the whole loop — the leak found and the fix proposed with the revenue attached — that is what DynoWeb does. If you want to run it yourself, the method above works with any tool. Start with 12 reasons your Shopify store isn't converting to know what you are looking for before you open the map.