Measure a defined population, not a vendor category.
Night Analytics Research separates population design, source observations, calculations and interpretation so readers can see what a metric actually supports.
1. Define the population first
Each vertical starts with a functional eligibility rule. Discovery sources can suggest candidates, but traffic rank or a vendor category does not determine inclusion.
2. Separate Current Coverage and Comparable Panel
Current Coverage can evolve as the market changes. Longitudinal claims use a fixed Comparable Panel so coverage expansion is not mistaken for sector growth.
3. Resolve identity at three levels
Corporate group → Platform / brand → Property. Regional domains, mirrors and aliases are not automatically counted as separate platforms.
4. Keep measurement families distinct
Web visits, search interest, public disclosures and regulatory events answer different questions. We do not average incompatible inputs into a generic industry score.
Adult Video baseline protocol
Comparable Panel v1 contains 13 adult-first standalone web platforms whose primary public product is on-demand browsing or streaming of a multi-video catalog. The baseline uses the same three complete monthly Similarweb estimated-visit observations for each primary domain: July, August and September 2026.
Similarweb figures are third-party modeled estimates, not first-party analytics, unique users, revenue or market size. Night Analytics calculations may include share of estimated visits among tracked platforms, short-window change and concentration metrics when the denominator and period are explicit.
Ownership-adjusted analysis
Brand-level observations can understate concentration when multiple brands share a corporate owner. Night Analytics groups only ownership relationships supported by evidence and leaves unresolved entities separate. That is why the report uses the phrase known-ownership-adjusted concentration.
Access and regulatory events
Material events such as age-assurance deployment, geoblocks, access restrictions, acquisitions, closures or regulatory investigations are maintained as a separate dated layer. They help readers interpret discontinuities, but proximity does not establish causation.
Versioning and corrections
Monthly observations are preserved with their collection vintage. Public analytical refreshes are expected quarterly by default. Historical releases are not silently rewritten when a vendor revises estimates; release IDs and data-through periods remain visible.
Baseline release
The current Adult Video release is AV-MVP-BASELINE-2026-10, data through September 2026.