Gaze cueing and conversion

You have a strong image, a clear message and a button leading to the next step. You still need to know whether visitors view those elements in the intended relationship. The gaze direction of a person in the image is one design variable worth investigating. What might it contribute, and when can you claim a conversion improvement?

Gaze cueing is the orienting of attention in response to another person’s gaze direction. In communication, you can investigate it by showing a person looking towards relevant information. This is an attention hypothesis; additional enquiries or purchases need separate measurement.

In a classic 1998 study, Friesen and Kingstone found faster responses to targets in the direction of a schematic gaze, even when that direction did not predict target location. This concerned attention without a necessary eye movement, not purchasing. Friesen and Kingstone, 1998

A cue is a signal that can guide attention. The term describes an attentional effect; it is not a conversion model in its own right.

A meta-analysis supports the gaze-cueing effect in experimental attention tasks. Its magnitude varied with factors including task, timing and facial expression. McKay et al., 2021

Advertising has also been studied. Hutton and Nolte found longer viewing of the product, logo and remaining advertisement when a model looked at the product rather than the viewer. The study measured viewing behaviour. Hutton and Nolte, 2011

In a banner experiment, Palcu and colleagues found effects on product attention and reported purchase intention. This did not establish actual purchases; product evaluation and purchase intention were also distinct outcomes. Palcu et al., 2017

The practical inference is to use the evidence to motivate and test a design choice. Do not import a universal uplift percentage into a forecast for your website.

This table is an editorial measurement framework. Choose the outcome that fits your decision. A heatmap is not a conversion report, and clicking “enquire” is not a completed enquiry.

Start with one question: which information should someone be able to find at this moment? It could be a product detail, evidence, a price or an explanation of the next step.

Create two variants whose differences you can explain. Keep copy, offer and placement as similar as possible. If you change the model, rewrite the headline and enlarge the button at the same time, you cannot attribute the outcome solely to gaze direction.

Check each relevant screen width. Does the face remain visible alongside its target after cropping, scaling or stacking? Also document whether you change only the eyes or the whole head: these are different changes to the image.

A person is not compulsory. If the photograph contributes little to the message, consider a simpler design without a face as a separate hypothesis. This is a design option to test, not a claim that faces or arrows always perform better.

A fictional retailer wants to know whether visitors understand a chair’s adjustability before ordering. In variant A, a depicted person looks towards the visitor. In variant B, the same person looks towards the product detail. The rest of the page stays as similar as possible. This is a proposed test, not a client case or observed finding.

If B attracts more attention without a clear change in orders, report both findings. Then investigate whether the information is relevant and understandable. An inconclusive conversion result does not by itself demonstrate that no effect exists; uncertainty and sample size belong in the assessment.

1. Define the problem. Which information might visitors be missing, and what supports that suspicion?
2. State the hypothesis. Describe the expected change and the outcome you will measure.
3. Check the variants. Document every difference, including on mobile.
4. Choose suitable research. Use eye tracking for viewing behaviour and a randomised field experiment for the chosen conversion outcome.
5. Specify the analysis beforehand. Justify sample size, duration, exclusions and stopping rules. Avoid repeatedly checking until a favourable result appears.
6. Report the effect and uncertainty. Include counts and the relevant denominator. Identify the comparison agreed in advance.

This is a practical framework, not a universal research protocol. It does not prescribe a participant minimum that applies to every situation.

• Showing a hotspot in a heatmap and interpreting it as “more persuasion”.
• Presenting purchase intention as actual revenue.
• Changing multiple image elements and attributing the result to one.
• Applying desktop findings to mobile without checking.
• Reporting only the best-performing subgroup after the analysis.
• Making important conditions less visible to generate more clicks.

Use attention guidance to make relevant information accessible. Keep costs, conditions and choices clear.

Gaze direction is a concrete variable to include in image selection and testing. First identify the information supporting the decision, then measure whether the design helps. Report attention, understanding and conversion separately. This keeps demonstrated findings distinct from hypotheses that still need testing.