Find which domains over-index for any of 1,667 deterministic personas built from fixed, versioned vocabularies. The audience profile lives on the domain, not a cookie — so the same plan covers Chrome, Safari, Firefox and iOS alike, including the 40%+ of traffic where cookies are already blocked.
Cookie-based planning only sees traffic where third-party cookies function. Safari, Firefox and iOS block them by default — hiding a large slice of premium inventory from reach curves. Privacy regulation adds further pressure.
Persona planning asks “which properties does my target persona read?” instead of “which users can I follow?” A site’s audience profile is a stable property of the domain — it doesn’t vanish when the visitor arrives on an iPhone.
Demographics, interests, purchase intent, B2B firmographics and a resolved persona are pre-computed for every domain — all from a fixed vocabulary so two analysts filtering the same file always get the same answer.
Roughly 40%+ of traffic is cookieless today. Domain-level personas describe all of it equally.
A persona in this dataset is not a probabilistic lookalike. It is a deterministic combination of coded attributes, each from a versioned v1.0 vocabulary aligned with IAB Audience Taxonomy 1.1 — so a planning filter written today still means the same thing at the next quarterly refresh.
8 age brackets, a 5-point gender skew (male_strong → female_strong), 6 income bands, 7 education levels and 14 life stages such as young_professional or family_young_children.
Household composition, employment status, home ownership and urbanicity (urban, suburban, small_town, rural) round out who the readership is.
29 interest groups and 285 sub-interests carrying INT.* codes — from INT.sports.cycling to INT.personal_finance.personal_investing.
34 intent groups and 283 segments carrying PI.* codes, e.g. PI.travel.hotels_and_resorts or PI.auto_ownership.new_vehicles — the commercial layer on top of interest.
Company-size bands, seniority and job function on LinkedIn-standard scales, plus an audience_type flag (b2c / b2b / mixed) so B2B plans filter cleanly.
Every attribute ships with a banded confidence value — low, medium or high — so planners can trade reach against certainty explicitly instead of implicitly.
The full field reference lives on the audience segmentation taxonomy page.
The planning workflow is a sequence of filters over a flat file (or the same query via the API). No modelling step, no black box — the intermediate state at every stage is a readable list of domains and codes.
Take the brief’s target description and restate it in vocabulary terms: age bracket, life stage, income band, interests, intent.
Map each phrase to its fixed code — “young city renters into fitness” becomes four filterable values, not a paragraph.
Apply the filters to the domain file. Add a confidence floor (medium+) to keep only well-evidenced matches.
Sort the matches by how strongly the persona concentrates on each domain versus the corpus baseline, then sanity-check the head of the list.
Hand the ranked domain list to activation: direct buys, allow-lists, PMP curation or publisher outreach.
A brief asks for “younger urban professionals with disposable income who are actively considering gym memberships”. Here is that sentence as a database filter — codes on the left, human-readable labels on the right.
| Field | Filter value (code) | Reads as |
|---|---|---|
| age_bracket | 25_34 | 25–34 year olds |
| life_stage | young_professional | Young professional |
| income_level | upper_middle or high | Upper-middle to high income |
| urbanicity | urban | Urban readership |
| interest | INT.healthy_living.fitness_and_exercise | Fitness and exercise |
| purchase intent | PI.recreation_fitness.gyms_and_health_clubs | Gyms and health clubs |
| confidence | medium or high | Well-evidenced attributes only |
| Cookie / ID-based audiences | Panel-based site metrics | Domain-level personas (this dataset) | |
|---|---|---|---|
| Coverage of cookieless traffic | Blind where cookies are blocked (Safari, Firefox, iOS) | Covers panelled sites only; long tail thin | Full corpus — profile is a property of the domain |
| Long-tail domains | Sparse — few matched IDs | Mostly absent below the head of the web | 102M domains, head to tail |
| Granularity | User-level, but shrinking | Site-level, coarse demographics | Domain-level: demographics + 285 interests + 283 intent segments + persona |
| Reproducibility | Segment definitions vary by vendor | Panel weighting changes over time | Fixed v1.0 vocabularies; same filter, same meaning |
| Privacy posture | Identifier-dependent | Panel consent-based | No PII anywhere in the pipeline |
Each of the 1,667 personas is a named, fixed combination of vocabulary attributes — age bracket, life stage, income band, interests and intent drawn from the versioned v1.0 vocabularies. The same underlying attributes always resolve to the same persona, so a persona filter is reproducible across teams, tools and quarterly refreshes. Nothing is probabilistically matched to individual users, and no PII is involved.
No — and we say so deliberately. The domain database is built for planning, inventory curation and enrichment; the real-time API adds per-URL granularity for planning and analysis workflows. Neither product claims impression-level pre-bid classification in the bidstream. The output of a persona plan is a ranked domain list you activate through direct buys, allow-lists or Deal-ID packages.
Identically to the rest of it. Cookies remain on Chrome, but Safari, Firefox and iOS block third-party cookies by default, which puts roughly 40%+ of traffic outside cookie-based reach curves. Because the persona is computed from the domain’s content and audience characteristics rather than from visitor identifiers, cookieless inventory is described with exactly the same fields and confidence bands as everything else.
The top 100k domains with full audience attributes cost $490 one-time (or $190/quarter refresh); the top 1M domains cost $1,990 one-time ($590/quarter refresh) with instant card checkout and immediate download. Vertical and country slices run $190–$490, and larger cuts — 5M up to the full 102M corpus, custom enrichment or feeds — are quoted individually. See pricing for details.
Turn a persona domain list into curated Deal-ID packages.
Score how well each planned domain fits the brand, beyond blocklists.
Reach the traffic that cookie-based plans never see.
Compare your planned footprint against competitor audiences.
Filter live audience data in the demo dashboard, or download the top-1M domain file with full attributes and start planning in a spreadsheet this afternoon.
Open the audience demo See database pricing