Terms

Cluck the Web is an independent editorial directory. Its rankings are a compilation of available third-party estimates, public charts, product information and editorial judgement. They are not an official worldwide league table, an independently audited usage census, or an endorsement by a listed service.

The initial edition covers a selected set of 100 websites and 100 apps, allocated to five broad categories for each medium. “Overall” means position within the relevant covered set of 100, not a measured position among every website or app in existence. Category rank means position within the 20 entries covered in that category. A service may appear once in the website set and once in the app set; these are separate rankings.

The covered selection focuses on general-audience consumer services. It is not exhaustive of every country, language or adult, gambling, enterprise-only or preinstalled-system category. A category's number-one entry is first in this directory's covered category, not necessarily the number-one service in a measurement provider's different category definition.

For websites, the initial order uses exact-domain worldwide estimates from Semrush's September 2026 overall chart where available. Remaining entries are ordered using visible estimated visits, then available category-chart positions and editorial tie-breaking. Category positions from different charts are not inherently comparable; weaker tail positions are editorial and require further review. A subdomain is not assigned its parent domain's traffic as if both were the same measured property. Regional domains are not automatically summed into an invented global brand total. Source-specific worldwide positions, where available, remain separate from Cluck the Web's covered-set positions.

For apps, the initial order is an editorial synthesis of cross-platform download reports, product reach signals and limited public app-usage charts. Cross-referencing does not turn downloads into active users. Download reports can include different dates, countries, stores and taxonomies; preinstalled applications and unavailable regional applications create further differences. Public Apple store metadata is used for product identity, not as a global usage ranking. Some entries have no directly comparable chart point in the collected evidence and their order remains provisional. No fabricated download or active-user total is assigned to those entries. The finance-app report contains conflicting period labels; its figures are retained with that limitation rather than silently relabelled.

Rankings are snapshots and can be revised as better evidence becomes available. A lower directory position does not imply lower quality, safety or financial strength. Measurement-provider estimates are not necessarily first-party counts.

Financial figures retain their currency, period and company scope. Revenue, profit, payment volume, merchandise value, funding, debt and valuation are different measures. Rounded figures produce approximate ratios. Market prices are dated snapshots, not live quotes. Annual recurring revenue is a run rate, not necessarily recognised annual sales. Private share-sale valuations and proposed transactions are dated references rather than public-market prices.

Owners sections describe corporate families and selected disclosed holders, investors or founders; they are not a complete current ranked register unless expressly identified as one. A founder, executive, operator and shareholder are different roles. Voting rights can differ from economic ownership; proxy interests can overlap. Fund managers generally act for portfolios and clients. Historic aggregate Vanguard disclosures must not be treated as a current combined position following its 2026 disaggregation.

Equity percentage, percentage of a specified share class and total voting power use different denominators. Disclosure dates and older underlying position dates remain separate. Reddit proxy interests overlap and must not be added together.

Our editorial 100-point rubric assigns up to 25 points each to claims versus practice, data use, content quality and user control. It is a comparative opinion based on the information available on the assessment date, not a probability of safety, guarantee or technical audit. Where a comparable assessment is incomplete we omit a numerical total, while still providing the qualitative guide. Named higher-scoring alternatives must be relevant to the same use and have a higher completed score under this framework; practical comparisons without a score claim are labelled separately.

The approved editorial Trust rubric (cluck-trust-1.0, approved 2026-10-18) totals 100 points.

What they say vs what they do — up to 25 points. Clarity and scope of claims; evidence of delivery, accountability and any substantiated contradiction.

What they know—and what they do with it — up to 25 points. Data minimisation, secondary uses, sharing and readable limits; necessary financial records are not equivalent to advertising profiling.

Quality vs Quantity — up to 25 points. Provenance, transparent selection, correction and relevance to the service's actual purpose; entertainment taste is not factual accuracy.

How much control you really have — up to 25 points. Practical documented choice over privacy, recommendations, permissions, spending, export, complaints and exit, as applicable.

Component bands: 0 = Substantiated severe failure on the assessed issue, without a meaningful documented safeguard; never a substitute for missing evidence.; 5 = Weak documented protections or substantial substantiated mismatch.; 10 = Material limitations; some documented protections.; 15 = Mixed/standard documented provision with material trade-offs.; 20 = Strong, specific documented provision with clear remaining limits.; 25 = Exceptional, well-corroborated provision on the assessed issue; does not mean perfect safety.

Use five-point component bands, not spurious one-point precision. Sum the four components without extra weights. Preserve source dates and component rationales.

Evidence gate: All four components require sufficient relevant, inspected evidence and an explicit rationale. No numeric score from keyword counts, download rank, ownership wealth or unread pages. A missing component means the total remains null/Not rated. Unknown is not zero or a negative finding.

An editorial public-document assessment of the listed consumer service and stated account/region scope. Not a penetration test, security certification, content-wide fact check or financial-product recommendation.

The fifth Trust heading is not scored. Suggest only assessed, higher-total alternatives with a genuinely comparable use case. Explain the trade-off; a higher total need not win every component. Do not suggest a messaging app as a replacement for a game or a different regulated product as suitable financial advice.

Scores are dated and must be revisited if product terms, controls or substantiated findings change. Sources and rationale remain accessible in Trust details; the full rubric belongs in Terms. No unsupported correction contact address is invented.

We use company disclosures, product documentation, financial filings, regulators, reporting and other public material. Sources are selected for relevance; not every detail can be independently cross-verified. Product features, permissions, prices, ownership and laws change and can differ by location, device, plan and account. Check the linked source and current product before acting. Behavioural explanations interpret documented design features; they do not claim access to private intent, prove an addiction mechanism or predict every person’s response. The absence of a feature from an article does not prove that it does not exist.

Editorial summaries simplify complex subjects and can contain errors or become outdated. Linked third parties remain responsible for their own products, terms and content. Nothing here recommends an investment, purchase or disclosure of sensitive information. We welcome specific corrections supported by sources. These general limitations belong here rather than being repeated after every observation.