How Glanzz scores attention
Glanzz predicts where people are most likely to look first on a creative, page, email or pack — before you spend a euro running it. Here's exactly how the scores are produced, and how far to trust them.
Every score here is a modeled prediction, not a measurement. It's a computer-vision model trained on how people's eyes move — strong for comparing options, catching problems, and explaining the "why" of a design — but it is not real eye-tracking and not a guarantee of performance. For high-stakes decisions, treat it as a fast, directional first pass and validate the shortlist with real users.
What the engine actually does
When you submit an image or URL, the engine builds a saliency map— a per-pixel estimate of how likely each spot is to draw the eye in the first few seconds. It's an ensemble: a neural object-saliency model combined with classic visual cues that real eyes respond to — text & CTAs, faces, contrast, and layout structure. Those signals are blended into one attention map, which every score and overlay is then derived from.
The map runs on CPU in a couple of seconds — no per-image cost, no GPU — so the numbers are consistent and repeatable for the same input.
The Impact Score
The headline Impact Score (0–100) is a transparent weighted blend of the five attention metrics below. Nothing hidden — these are the exact weights:
Verdict bands (Weak / Mixed / Solid / Strong) are descriptive cut-offs from the score design — not a percentile against a population. Where you see a percentile in-app, it's computed against your own workspace history of the same asset type.
The Decision layer
On top of the raw scores, Glanzz turns the same attention map into the calls you actually make:
- Where attention lands — each detected element ranked by its share of attention, plus a density multiplier (×1 = an average pixel, ×5 = it pulls five times its weight).
- Order of attention — the predicted gaze sequence (peak → peak with inhibition-of-return), mapped onto your elements, so you can see if the CTA is seen early or late.
- Above the fold — for tall pages, the share of total attention that lands in the estimated first screenful.
Asset types
Attention works differently per medium, so your asset-type choice re-weights the cues: advertising favours thumb-stopping power, websites favour hero/CTA/nav scanning, email favours the top fold and a single CTA, and packaging favours shelf standout. It also scopes your benchmarks and tailors the AI recommendations to that medium.
What it can't tell you
It predicts attention, not comprehension, emotion, brand recall, or whether someone will convert. It doesn't read your copy's meaning or your offer. Use it to make sure the right things get seen, then let messaging, proof and price do the rest.