Attention

Gaze plot

In short

The gaze plot's numbered stops are the main hot spots of predicted attention in your image, joined in a suggested sequence. Treat the order as indicative only. It does not show viewing time and has no benchmark.

What it shows

The gaze plot draws numbered circles on the main hot spots of predicted attention in your image, joined by lines in a suggested sequence. Hot spots are places where attention clusters and that are clearly separated from each other.

Numbers indicate where attention gathers; the order is indicative, and stop size does not show viewing time.

What feeds it

The predicted attention map of your image and the hot spots found in it. If you chose a platform, the gaze plot is computed on your creative inside that platform's screen.

How it's built

The gaze plot is a simulation built on the attention map. Foresight simulates many viewers who start, as people typically do, near the middle of the image and then move across it. The sequence linking the hot spots is derived from those simulated paths. It is calculated directly from the attention model's output, with no separate model and no reference data.

The number of stops equals the number of distinct hot spots, so it varies with the creative. A creative with a single dominant hot spot shows a single stop. The same hot spots feed the Spread score.

Scale and direction

Numbered stops, positioned on the image. The numbering is an indicative order only: it is not a measured sequence of where people look.

All stops are drawn the same size: the gaze plot does not show how long people look at each stop.

The gaze plot has no benchmark. It describes your creative alone and is meant for qualitative reading. To compare an element with a reference, draw an area of interest and give it the type Face, Logo or Text; see Element benchmarks.

Image vs video

Images only. Videos do not get a gaze plot.

How to read it

• Use the stops to see where attention gathers: are your product, brand and message among them?
• Few stops close together suggest a creative that guides the eye; many stops far apart suggest attention is divided.
• Read the numbers as a hint, not a guaranteed route. For how attention develops over time, use the attention build-up.

Common pitfalls

• It is not recorded eye-tracking. No one's eyes were followed; it is a simulation built on predicted attention.
• Stop size does not mean viewing time. All stops are the same size.
• Do not read stop numbers as scores. There is no gaze-plot benchmark.
• Weak secondary areas may not get a stop, even if they show faintly on the heatmap.
• It may not match the attention build-up's order. The two are separate simulations and can suggest different sequences.