Safari and Firefox block third-party cookies by default, and iOS restricts cross-app identifiers — 40%+ of traffic is already cookieless today. Content-inferred audience attributes close the gap: profile what a domain publishes, then buy the audiences its content demonstrably attracts. No PII anywhere in the pipeline.
Third-party cookies remain on Chrome. But the browsers and devices that already block them account for a large, stable and often premium slice of every campaign's addressable traffic.
Retargeting pools and third-party audience segments depend on a cross-site identifier. On Safari, Firefox and iOS in-app inventory, that identifier doesn't exist — DSP segments silently under-deliver.
When buyers can't attach an audience to an impression, they bid run-of-network. Publishers with Apple-heavy readerships monetize below comparable cookied inventory.
What a site publishes is observable on every browser, every time. Inferring the audience from content moves targeting from the user to the inventory — no consent-dependent identifier required.
Join your inventory to pre-computed audience attributes, filter on fixed vocabulary values, and package the result the way your buying stack already activates deals.
Pull domains where your delivery logs show high Safari, Firefox or iOS share — the placements your audience segments currently skip.
Match those domains against the 102M-domain dataset. Each row carries demographics, interests, purchase intent, personas and banded confidence values.
Every attribute comes from a fixed, versioned vocabulary — INT.* interests, PI.* intent, enumerated demographics. Filters are exact, repeatable and diffable between refreshes.
Export as a curated inventory package: an allowlist for a PMP or Deal ID, an SDA description on the sell side, or a planning input for direct buys. For per-URL depth, enrich through the real-time API.
Honest scope note: this is planning-time and curation-time targeting. The dataset and API describe domains and pages for selection and packaging — not an impression-level, in-bidstream pre-bid classifier.
Your logs show a mid-size cooking site delivering 70% of impressions from Safari and iOS. Behavioral segments can't describe those readers — but the domain record can.
A meal-kit advertiser filtering on INT.food_drink.cooking + PI.food_beverage.food_delivery_services at high confidence surfaces this domain into a curated cookieless food package.
Hundreds of comparable domains are surfaced the same way — without touching a single user identifier. Every code is enumerable in advance.
INT.food_drink.cookingHealthy Cooking and Eating INT.food_drink.healthy_cooking_and_eatingPI.food_beverage.food_delivery_servicesRestaurants PI.food_beverage.restaurantsEach approach trades off differently on cookieless coverage, audience meaning and operational effort. They are complementary more often than competitive.
| Approach | Works on Safari / iOS web | Carries audience meaning | Requires user data / consent signal | Where it fits |
|---|---|---|---|---|
| Third-party cookie audiences | Blocked by default | Yes | Yes | Chrome-side of the plan only |
| Keyword / page-context targeting | Yes | Topic, not audience | No | Brand-adjacency at page level |
| Publisher first-party segments | Yes (logged-in scope) | Yes | Publisher-held, per-publisher | Large single-publisher deals |
| Content-inferred audience attributes | Yes — browser-independent | Demographics, interests, intent, personas | No — no PII in the pipeline | Cross-publisher curation, PMP packaging, planning |
Publishers describing their cookieless audience with neutral, taxonomy-coded attributes.
Building curated domain packages and Deal-ID allowlists from audience filters.
The same content-inferred method applied to streaming and CTV planning.
Planning against 1,667 deterministic personas instead of opaque segments.
Page-level audience segmentation for per-URL planning and analysis.
Every enumerated vocabulary value, aligned with IAB Audience Taxonomy 1.1.
No — third-party cookies remain on Chrome. The cookieless problem exists today because Safari and Firefox already block them by default, and iOS restricts cross-app identifiers — together roughly 40%+ of traffic. Content-inferred attributes work identically on cookied and cookieless browsers, so one methodology covers both.
No. The domain dataset is built for planning, curation and enrichment: select and package cookieless inventory into curated lists, PMPs and Deal IDs before campaigns run. The real-time API adds per-URL granularity for planning and analysis. We never claim impression-level pre-bid classification.
Attributes are inferred from a domain's published content, not its visitors. Every attribute comes from a fixed, versioned vocabulary (v1.0) aligned with IAB Audience Taxonomy 1.1 — 8 age brackets, 5-point gender skew, 6 income bands, 285 sub-interests, 283 purchase-intent segments — each with a banded confidence value. No PII is used. The full vocabulary is published on the taxonomy page.
The top 100k domains cost $490 one-time ($190/quarter refresh); the top 1M is $1,990 ($590/quarter) with instant checkout and immediate download. Vertical and country slices run $190–$490. Larger corpora up to 102M and custom feeds are quoted individually, with custom licensing from $15,000/year. The real-time API uses standard credit plans, e.g. Pro at $99/month for 10,000 credits. See pricing.
Run any domain through the audience demo and see the exact attributes a cookieless targeting filter would use — then download the corpus slice that fits your plan.