The human attention database.
Half of our waking life now happens on screens — and no one can say what actually happens there. ATTEX can: the first complete record of how we actually live online, with real eye-tracking at its core. A shared, auditable database that media, AI and the marketing industry can build on, compare against, and trust.
a black box. Until now.








Understanding attention is key to building better advertising and consumer experiences. ATTEX Data's unique technology and Tobii's heritage in eye tracking make this a game changer for how we plan, optimize, and measure media.
Billions in ad spend vanish every year — because human attention data isn't used.
85% is lost — ~$600bn a year buys ads no human ever sees.
And the flood keeps rising: AI makes content infinite — human attention stays fixed at ~8 hours a day. Infinite supply chasing a finite resource makes attention the scarcest asset in media — and the most valuable thing to measure.
Local, passive, always-on human eye-tracking panels — running quietly through people's real days.
Three interactive views on the ATTEX dataset. Numbers are illustrative. Not every view is in the public API today — but it's all in the data we're already collecting.
and the gaze between them, recorded.
One continuous record. Five connected layers.
Attention
Real gaze data — where eyes actually land
Sub-second eye-tracking. Fixations, saccades, gaze duration on every visible element.
Real gaze, on a real creative.

Built for Scandinavian winters.
520 km range · from €38,900
- Handle & Sponsored0.0s
- Headline0.0s
- Product hero0.0s
- Learn more band0.0s
- Caption & actions0.0s
With the real eye-tracking panel, every blob is millisecond-accurate gaze. Your cursor is just a stand-in — the metrics, joins and outputs are exactly the same.
A passive panel capturing the full digital day — on desktop and mobile.
Want to look at the real thing? One day of impressions, gaze and accessibility tree — yours after a 3-field form.
What the data unlocks.
Two reference workflows built on ATTEX data. The tools themselves aren't shipping products — they show what your team could build on top of the API.
Upload a campaign plan, see what actually drove the attention. Placement, creative and context — decomposed.
Ask a strategic question. We answer it with what your audience actually does — not what they say.
Interested in building your own use case on the ATTEX API? Talk to us about access →
Media buyers optimise against ATTEX data — so the ads actually get seen.
Uplift across categories| Category | Ad recall | Consideration |
|---|---|---|
| Retail | +18% | +10% |
| Insurance | +19% | +15% |
| FMCG | +19% | +15% |
| Average | +19% | +13% |








“Schibsted Attention Study 2023” — a published study on brand lift and ad fixation, built on ATTEX data.
Read the study (PDF)Independent academic supervision. The study compares fixation, dwell time and brand lift across Schibsted's desktop and mobile inventory using ATTEX's passive eye-tracking panel.
One database. Plug it into your stack.
We are not a dashboard vendor. ATTEX is the shared foundation — a transparent, auditable database that media, AI and the marketing industry can build on, compare against, and trust.
Stream to S3, BigQuery or Snowflake — or query the JSON API directly. Same schema either way.
Push events into your bidder, MMM, MTA model or LLM training set as they happen.
Stable panelist IDs, creative IDs, surface and content IDs — built to join with your stack.
Consented panel, GDPR-grade documentation. Auditable lineage on every metric.
// Pull verified attention + behaviour for a creative
const r = await attex.query({
audience: { age: [25, 44], gender: "female", market: "SE" },
surface: ["ig_feed", "tiktok", "ctv"],
creative: "volvo_ex30_q2_15s",
metrics: ["attention_rate", "fixation_ms", "co_exposure"],
window: "last_14d",
});
// → joinable to your DSP, MMM, brand-lift study, ID graph or LLM
attex.streamTo(s3 ?? bigquery ?? snowflake);
Same calls, same schema — whether you stream raw events, batch-export to your warehouse or query the JSON API.