Benchmarks
Foresight uses benchmarks in two ways. Every score is positioned against a large reference library of creatives, and, if you pick a comparison group, the group's average is shown next to your score. Colour bands are fixed and do not depend on the benchmark.
The word "benchmark" covers two different things in Foresight. Keeping them apart makes your results much easier to read.
Two ways Foresight uses benchmarks
1. The reference library behind every score. Foresight's models produce raw measurements, such as how concentrated predicted attention is or how strongly a creative is predicted to move people. To turn those raw measurements into a 0–100 score, Foresight positions your creative against a large reference library of real-world creatives that were analysed with the same models. That is why a score above 50 means "higher than typical" (for Mental Demand the middle is best, and Valence is not a ranking). Every score is compared with creatives of the same media type: images with images, videos with videos (see how scores are positioned).
2. The comparison value next to your score. When you set up an image analysis without a platform frame, you can pick a comparison group: a creative type and, optionally, a sector. Foresight then shows the group's average next to each score, as a marker on the score bar and as a "+N" or "−N" chip. This tells you whether your creative sits above or below that group.
Your score never depends on the group you pick. Changing the comparison group changes only the reference line you compare against.
When no comparison value is shown
Your scores are still calculated exactly the same way, against the reference library of the same media type. The only difference is that no comparison value is shown: no marker, no "+N / −N" chip and no "vs benchmark" wording. This applies to:
- an image with no creative type or sector;
- an image analysed inside a platform frame (Instagram, Facebook or X), which has no creative type or sector choice;
- every video.
Colour bands are separate
The green, amber and red colours are fixed score bands, the same for every media type, sector and platform. They are not calculated from your comparison group. A creative can beat its group average and still sit in the amber band. See Score tiers.
In this section
- The benchmark universe: what the reference library contains.
- Sectors and creative types: which comparison groups exist and when they apply.
- How scores are positioned: how a raw measurement becomes a 0–100 score, and how the comparison value is calculated.
- Score tiers: the green, amber and red bands and what they mean.
- Image vs video benchmarks: why an image score and a video score are not on the same ladder.
- Platforms and benchmarks: what choosing Instagram, Facebook, X or TikTok does to the benchmark.
- Benchmark versions and updates: why comparison values can shift slightly over time while your scores stay fixed.
Faces, logos and text inside an image have their own, separate references. See Element benchmarks.