Benchmarks

How scores are positioned

In short

Most Foresight scores are rankings. Your creative's raw measurement is placed within the reference library, and the score says how much of that library it beats, with 50 in the middle. Valence is the exception. The comparison value is the average score of your chosen group, calculated the same way.

What it is

Foresight's models first produce raw measurements: how concentrated predicted attention is, how far apart the main points of attention are, how strongly the creative is predicted to move people, how visually complex it looks. Raw measurements are hard to interpret on their own. Foresight makes them readable by positioning each one against the reference library.

Scores are rankings

Imagine lining up every creative in the reference library from lowest to highest on one measurement, then placing your creative in the line. The score tells you how much of the line your creative is ahead of:

• 50 means your creative is in the middle: about as many reference creatives score below it as above it.
• 80 means it is ahead of most of the library.
• 20 means most of the library is ahead of it.

Most scores show the share of the reference set your creative beats: 50 is typical.

Every ranking is made within the same media type: an image is ranked among images and a video among videos. This holds for attention scores and emotion scores alike.

How each KPI uses the ranking

• Emotional Engagement and Focus are straight rankings: higher beats more of the library.
• Spread is highest when attention gathers on one main area; more competing areas lower it, especially when they sit far apart.
• Clarity is built from Focus and Spread.
• Mental Demand is shown as a ranking too, where 50 means typical visual complexity for the library. Here the middle is best: the optimal zone is 25–75, and content that is too simple or too complex is flagged.
• Attention is built from Clarity and a complexity component: Attention also reflects whether the creative's visual complexity sits in a comfortable range.
• E-Score combines Emotional Engagement and Attention, so a high E-Score needs strength on both.
• Valence is different: it is not a ranking. It shows the emotional tone of the content itself, from −100 (negative) to +100 (positive).

How the comparison value is calculated

When you pick a comparison group, Foresight shows that group's average next to your score. For each KPI, it is the typical score of creatives in that group, on the same scale as yours: those creatives were scored with the same models and formulas, so the two numbers can be compared. Comparison groups exist for images analysed without a platform frame (see Benchmarks).

The difference (your score minus the group average, rounded) appears as a "+N" or "−N" chip. A few display rules:

• No chip or marker is shown when no comparison value applies (no group selected, a video, or an image in a platform frame) or no average is available.
• In some cases, for example Valence, no difference chip is shown.
• Mental Demand shows no chip. It shows a verdict (Too low, Optimal, Too high) and the group average as a marker on its scale.

For Valence, the comparison value is simply the group's average tone, on the same −100 to +100 scale.

Image vs video

The mechanism is identical for images and videos, but each media type has its own reference library, and every score ranks an image only among images and a video only among videos. See Image vs video benchmarks.

For a video, the overall score is ranked against whole reference videos, while the moment-by-moment curves are positioned against individual frames from the reference videos.

How to read it

A score of 65 on Focus means your creative's attention is more concentrated than that of most reference creatives. If your group's average is 42, the chip shows "+23": you are well above typical creatives in that group.

Because the group average is on the same scale as your score, it does not have to be 50. A group average well below 50 on Focus simply means creatives in that group typically rank low on Focus. It does not mean the group "fails".

Common pitfalls

• Reading a score as a percentage of something (for example "65% of viewers"). It is a position relative to the reference library.
• Reading Valence as a ranking. It describes the tone of your creative, not how it compares with others.
• Judging Mental Demand by the group average alone. Check whether it sits inside the 25–75 optimal zone.
• Assuming a score would stay identical if the library changed. Rankings are relative, but your stored scores are never recalculated. See Benchmark versions and updates.