Cookieless Audiences
Home Database API Docs Pricing Live Demo Taxonomy
Use Cases
Media Planning by Persona Inventory Curation & Deal Packaging Seller-Defined Audiences CDP & Analytics Enrichment ABM Account Profiling
Industries
SSPs DSPs Publishers Agencies Curation Platforms
Company
Contact Login
Try Live Demo
Use case · Market research & diligence

Market Sizing & Research from Domain-Level Audience Data

Cookieless Audience treats the web as a census: 102 million domains, each described with coded demographics, interests, purchase intent and firmographics from fixed, versioned vocabularies aligned with IAB Audience Taxonomy 1.1. Market maps, category sizing and due diligence become dataset operations instead of survey projects.

285 interests, 283 purchase-intent segments, 1,667 personas and B2B firmographics — same coded vocabulary for every domain.
Turn market questions into queries: how a category's media supply splits by audience, which segments are crowded or underserved.
Profile who a target company's properties actually reach — 8 age brackets, 6 income bands and 14 life stages per domain.
102MDomains in the census
29/285Interest groups / sub-interests
34/283Intent groups / segments
v1.0Versioned vocabularies
The problem

Market research inputs haven't kept up with the questions

Today's questions demand segment-level resolution — not “how big is fitness media” but “which properties reach high-income endurance athletes versus budget-conscious beginners?” Traditional inputs fall short in different ways.

Syndicated reports age fast

They aggregate to whole categories on annual cycles. By the time the report ships, the market has already shifted beneath it.

Surveys are narrow

Precise about the few hundred people asked and silent about everyone else. Scaling from a sample to a market-wide view requires big assumptions.

Panels blur the mid-tail

Panel-based web measurement resolves the head of the market but blurs out the mid-tail — often exactly where interesting entrants, acquisition targets and niche communities live.

Domain-level data changes the unit

Each domain becomes a row with coded attributes from a fixed vocabulary — the same 8 age brackets, 6 income bands and 14 life stages for every domain. Distributions aggregate cleanly without re-coding. Coverage does not depend on panels, cookies or site cooperation.

Research outputs

Three research products this dataset supports

The same rows and codes back three very different deliverables. The complete field reference lives on the audience segmentation taxonomy page.

Market & category maps

Slice a vertical's domains by audience attributes and lay out who serves whom: which properties reach income_level: affluent audiences, which own life_stage: family_young_children, where B2B and B2C supply overlap. The map shows crowded cells and white space in the same view.

Category sizing

Count and characterize the supply side of a category: how many domains carry a given interest or intent segment at high confidence, and how that supply distributes across demographics and geographies via vertical and country slices. Supply-side structure is the piece traditional sizing models usually have to assume.

Commercial due diligence

Profile a target's domains and its competitors' on identical coded attributes: does the audience match the equity story? Is claimed differentiation visible in the interests and intent segments the properties actually attract? Pairs naturally with competitor audience benchmarking for the head-to-head view.

Workflow

From research question to defensible answer

A typical engagement runs in four steps, all on the same coded fields.

STEP 1

Define the universe

Start from a vertical or country slice of the database, a list of category keywords, or a set of seed domains. The universe is explicit and reproducible — anyone can re-run the same filter on the same dataset version and get the same rows.

STEP 2

Filter on coded attributes

Express the market definition as attribute filters: INT.healthy_living.fitness_and_exercise at high confidence, audience_type: b2c, a target income band. Confidence bands let you run a strict cut and a broad cut of the same market.

STEP 3

Aggregate into distributions

Group the qualifying domains by any attribute to build the market's audience distribution — age mix, income mix, life-stage mix, intent-segment presence. Fixed vocabularies mean the group-by is trivial and comparable across markets and dataset versions.

STEP 4

Drill down where it matters

For named companies in a diligence or map, drop to page level with the real-time API — profile a target's key sections individually rather than trusting the domain average, and attach the coded evidence to the report.

Worked example

Mapping the home-fitness media category

An illustrative diligence question: a fund is evaluating a portfolio of fitness content sites and wants to know which audience pocket each property occupies. One row of the dataset looks like this — raw codes on the left, rendered labels on the right.

Dataset row (trimmed)

{
  "domain": "example-strength-training-hub.com",
  "audience_profile": {
    "demographics": {
      "age_brackets": [
        {"code": "25_34", "confidence": "high"},
        {"code": "35_44", "confidence": "med"... }
      ],
      "gender_skew": {"code": "male_lean",
                      "confidence": "medium"},
      "income_level": {"code": "middle",
                      "confidence": "medium"}
    },
    "interests": [
      {"code": "INT.healthy_living.fitness_and_exercise",
       "confidence": "high"}
    ],
    "purchase_intent": [
      {"code": "PI.recreation_fitness.gyms_and_health_clubs",
       "confidence": "medium"},
      {"code": "PI.apps.health_and_fitness_apps",
       "confidence": "medium"}
    ],
    "audience_type": "b2c",
    "vocabulary_version": "v1.0"
  }
}

Rendered as segment labels

Demographics

25–34 high 35–44 med Male lean med Middle income med

Interests

Fitness and Exercise high

Purchase intent

Gyms and Health Clubs med Health and Fitness Apps med
Three audience clusters Properties cluster into young/male/strength-focused, female-lean/wellness-oriented, and affluent personal-training pockets.
Intent signals differ Gym-membership intent dominates the first cluster; fitness-app intent runs higher in the second.
Distribution answers the question Which pocket each portfolio property owns — and how crowded it is — comes from the distribution itself, with a confidence band on every attribute.
Comparison

How domain-level audience data complements traditional inputs

It does not replace primary research — it replaces the guesswork between primary data points.

DimensionSyndicated reportsSurveys & panelsDomain-level audience dataset
ResolutionWhole categoriesSampled individualsEvery domain, coded attribute by attribute
Coverage of the mid-tailRarely itemizedSample too thin102M domains, uniform schema
ReproducibilityMethodology summarizedDepends on instrumentVersioned vocabularies; same filter, same rows
FreshnessAnnual cyclesPer-waveQuarterly refresh; API on demand
What it measuresAnalyst synthesisStated attitudes and recallAudience each property's content is built for, with banded confidence
Best used forContext and benchmarksWhy questionsStructure: who serves whom, where the white space is
Syndicated work frames the category and provides high-level benchmarks.
Domain dataset gives the category structure at full resolution — coded outputs drop straight into your models.
Primary research validates the cells that matter to the decision. A market map is a pivot table away.
FAQ

Frequently asked questions

How do you size a market with domain-level audience data?

You define the market as a filter over coded attributes — for example, all domains carrying a given interest or purchase-intent segment at high confidence within a vertical or country slice — then aggregate the qualifying domains into distributions by age, income, life stage or firmographics. That yields the supply-side structure of the market: how many properties serve it, and how their audiences are distributed. It complements, rather than replaces, demand-side sizing from spend or revenue data.

Can this be used for M&A or investment due diligence?

Yes. A common pattern is profiling a target's domains alongside its stated competitors on identical coded attributes, then checking whether the audience the content actually attracts matches the audience in the equity story — demographics, intent segments, B2B reach. Because the method needs no cooperation from the target and covers mid-tail properties, it works pre-LOI. Page-level drill-downs via the real-time API let you profile individual sections of a target's sites rather than relying on the domain average.

How is this different from syndicated research or panel measurement?

Syndicated reports summarize categories; panels estimate visitor demographics for sites with enough panel traffic. This dataset describes every domain individually — 102 million of them — from its content, using the same fixed vocabularies for each, with a low/medium/high confidence band on every model-inferred attribute. It resolves the mid-tail that panels blur out, and its coded output feeds directly into your own analysis instead of arriving as prose. The approaches are complementary: panels measure observed visitors where they can; this measures what each property's content is built to attract, everywhere.

What formats and licenses are available for research use?

The top 100k domains with full audience attributes cost $490 one-time ($190/quarter refresh); the top 1M is $1,990 with instant checkout ($590/quarter refresh); vertical and country slices run $190–$490. Larger corpora — from 5M up to the full 102M domains — plus custom enrichment and feeds are quoted individually, with custom licensing from $15,000/year. See pricing or contact us for a research or redistribution license.

Related use cases

Adjacent research workflows

The same rows power narrower questions on either side of a market study.

Put a market on one screen

Profile a few category domains in the live demo to see the coded fields, then license the slice of the database your study needs — from a $190 vertical cut to the full 102M-domain corpus.

Stay in the loop

You are on the list!

We will send you updates that matter — no spam.