Attention

In short

Foresight predicts attention; it does not record it. A model trained on real eye-tracking studies estimates where a typical viewer is likely to look. Heatmap, fogmap, gaze plot, attention build-up and areas of interest are all views of that prediction.

Predicted attention, not eye-tracking

Foresight predicts attention; it does not record it. No one is filmed or tracked when you run an analysis. An attention model, trained on large collections of real eye-tracking recordings from research studies, estimates for every point of your creative how likely people are to look there during free viewing.

The result is a probability map. It describes a typical viewer population, not a specific person, and it is based only on what is visible: shapes, contrast, faces, text and composition. It does not account for sound, the viewer's goals, brand familiarity or the surrounding media context, unless you choose a platform frame. Each map is scaled to the creative's own strongest point, so the colours show where attention goes within that creative.

Attention is predicted on what the viewer would see, including the platform screen if you chose one.

What you see in Foresight

All attention overlays are views of that one prediction:

• Heatmaps: where attention concentrates.
• Fogmap: what stays visible and what fades.
• Gaze plot: the main hot spots of attention, with an indicative order (images only).
• Attention build-up: how attention spreads over the first seconds (images only).
• Elements and areas of interest: rectangles you draw and label around a face, logo, headline or product to measure the attention each receives (images only).
• Element benchmarks: how areas of type Face, Logo or Text compare with a typical element of their type and size.

The scored numbers, Attention, Clarity, Focus, Spread and Mental Demand, are explained in the KPIs section.

Platforms

If you pick a platform, Foresight places your creative inside a simplified version of that platform's screen, with the interface softened so it does not dominate, and predicts attention on the whole screen. Every overlay then reflects your creative in context, and some attention may go to the interface around it. The algorithm stays the same; the picture being analysed changes.

Image vs video

	Image	Video
Heatmap	Yes	Yes, as a video
Fogmap	Yes	Yes, as a video
Gaze plot	Yes	No
Attention build-up	Yes	No
Areas of interest	Yes	No

For videos, Foresight predicts attention on frames sampled throughout the video and fills in the moments between them. Each frame is analysed on what it shows; sound and the story so far are not taken into account.

How to read it

Overlays are visual aids for understanding where attention goes. The comparable numbers are the attention KPIs and the area-of-interest values.

Common pitfalls

• It is a simulation of a population average, not proof that your audience looked at something.
• Do not compare two creatives by colour intensity. Each map is scaled to its own creative.
• Predicted attention cannot capture intent. A viewer searching for a price or loyal to a brand may look differently.