Eye tracking marketing: what it measures, who uses it, and the cheaper ways to get the answer
What an eye tracking study records, the marketing questions it answers well, what it costs you in time and setup, and the lower cost methods that settle most page and ad decisions.
Published by AI Knows Us · Updated 12 October 2026
Eye tracking in marketing means recording where people look, for how long and in what order, while they view an advert, a web page, a pack or a shelf. A camera and software follow the pupil and turn the result into fixation points, gaze paths and heatmaps. It answers one question precisely: did the thing you wanted noticed get noticed, and when. It does not tell you whether people understood it, liked it or bought.
What the equipment records
An eye tracker measures where the gaze rests and where it jumps. The pause is a fixation. The jump between two fixations is a saccade. From those two events, the software builds the outputs a marketer actually reads.
- Heatmap. A colour overlay showing where fixations clustered across many viewers.
- Gaze plot. The path one viewer took, numbered in order.
- Time to first fixation. How long it took before a viewer looked at an area you care about, such as a price or a button.
- Dwell time. How long viewers kept looking at that area.
- Areas of interest. Zones you draw around a headline, a logo or a call to action so each can be compared with the others.
Screen based trackers sit under a monitor and suit web pages, emails and video. Wearable glasses suit shops, shelves and packaging, because the person walks and the camera records the scene they see.
The questions it answers well
| Marketing question | What the study shows |
|---|---|
| Does the headline get read before the image | Order of fixations on the layout |
| Is the call to action seen | Time to first fixation on the button |
| Does the logo register on the pack | Dwell and first fixation on the logo area |
| Which product on the shelf draws the eye | Glasses footage across the shelf, by position |
| Does a face in the advert pull attention away from the offer | Share of dwell time on the face against the offer |
Packaging and in store work is where eye tracking earns its cost, because there is no other cheap way to see what a shopper noticed in the three seconds they stood at a shelf.
What it does not measure
Looking is not the same as reading, and reading is not the same as agreeing. A viewer can stare at a confusing price for a long time. A long dwell can mean interest or it can mean the text was hard to follow. So pair the eye data with a question afterwards, such as asking the person to say what the offer was, and read the two together.
The reading patterns people mention
You will meet the F pattern in almost every article on the subject. Nielsen Norman Group published its original eye tracking visualisations and described that users often read web pages in an F shaped pattern: two horizontal stripes followed by a vertical stripe. The practical lesson is for text heavy pages. Put the point in the first line, put meaning in the first words of each heading and paragraph, and do not hide a key fact in the middle of a long block. It is a pattern for pages full of plain text, and a page with strong layout, images and clear sections is read differently.
What a study takes
- A question written before you start. "Where do people look" is not a question. "Do people see the delivery charge before they start checkout" is.
- Participants who match your customers. A dozen or so people from your real audience usually reveal the main attention patterns in a layout, and a bigger sample is needed when you want to compare two designs with confidence.
- Tasks, not staring. Ask each person to do something, such as find the price or choose a product, so the gaze reflects intent.
- Two versions to compare. A heatmap of one design says little. The same measurement on two versions says which one works.
- A follow up question. Recall and comprehension finish the story the gaze data starts.
The lower cost ways to settle the same decision
Most small businesses never need a lab. For a web page, a click and scroll heatmap tool shows where visitors click and how far they scroll, and session recordings show what a real visitor did in order. A five person usability test, where you watch someone try to complete a task and ask them to talk while they do, finds most layout problems for the price of an afternoon. Split testing two versions measures what people did instead of where they looked. For a pack or a poster, a squint test works as a first pass: blur your eyes or shrink the design to the size of a thumbnail and see what still stands out.
When the question is how a page performs for conversions, the conversion rate page gives the order of tests, and the landing page agency page covers the brief for a designer.
When a proper study is worth it
- You are changing the pack of a product that sells through shelves, and a wrong design costs a print run.
- You run a high traffic page where a small gain pays for the study.
- You are choosing between advertising layouts for a national campaign.
- You need evidence for a decision that people inside the company are arguing about.
For everything else, run the cheap method first. If it settles the matter, you saved the budget. If it leaves the question open, you now know exactly what to ask the lab.
Questions about eye tracking in marketing
Is eye tracking the same as a heatmap?
A heatmap is one output of eye tracking. Web heatmap tools that record clicks and scrolls measure a different thing, behaviour on the page, and they are not eye tracking, though the pictures look alike.
How many people do I need in a study?
For a read on one layout, a small group of real customers shows the main patterns. For a comparison you want to rely on, plan a larger group and ask the research supplier to state the sample size in the proposal.
Can a webcam do the job?
Webcam based tools estimate gaze roughly and suit large scale, rough attention checks. Dedicated hardware records gaze more precisely and suits packaging, shelf and interface work where small differences matter.
Does attention predict sales?
Attention is the first step in a chain, and a product has to be noticed before it can be chosen. Treat the gaze data as the first measurement in a chain that ends with a purchase measured in your own sales figures.
Where the eye goes when the reader is an assistant
Eye tracking measures how a person scans a page, and a growing number of buyers now skip the page and ask an assistant, which reads it for them. What the assistant lifts from a page is the first clear sentence under a heading, a table, a price or a named fact, and that is close to the thing eye tracking teaches: lead with the point and keep the evidence close. You can test the result directly by asking ChatGPT, Gemini, Perplexity and Claude the question your page answers and seeing which sentence they quote. AI Knows Us (aiknowsus.com) runs that comparison for a business across the assistants, and the lines they quote tell you which parts of your page are doing the work.