Content Integrity, Scoring & IP Policy

_Last updated: 3 August 2026_

This policy is TrustRating's constitution. It states the commitments that make the platform worth using — how scores are kept honest, how manipulation is treated, who owns what, how our brand may be used, and how copyright complaints are handled. It binds us as much as it binds you, it is part of the Terms of Service and the Business Terms of Service, and everything else on this platform is built on it. Review platforms live or die on one question — can the numbers be bought? — and this document exists so that our answer is not a slogan but a set of enforceable commitments with named remedies.

1. The no-pay-to-play covenant

We commit, contractually and publicly: no payment to TrustRating buys a score, and no payment removes truth. No subscription tier, advertising spend, partnership, courtesy grant or personal relationship: raises or protects a TrustScore; removes, hides, reorders or de-weights a genuine review; changes how TrustGuard or human moderation treats a company's content; or buys placement that is presented as merit. Paid visibility, where it exists (for example featured placement), is a plan feature that is always distinguishable from organic ranking, and the score shown in a featured slot is the same score shown everywhere else. The distinction is structural, not cosmetic: featured slots are separate surfaces, organic rankings and category orderings are computed without reference to plan status, and search results do not know what anyone pays. A featured company with a poor score displays its poor score in the featured slot — paying for the spotlight buys the spotlight, never the makeup. The inverse holds too: not paying us never harms a company — unclaimed and free-tier companies are scored by the same methodology, moderated by the same rules, and protected by the same fraud systems as the highest-paying subscriber. Any employee, partner or affiliate who offers to bend these rules for money is defrauding you — report it via the contact page and we will act on it.

2. How scores work — and what they are not

TrustScores combine verified customer reviews with independent assessments from multiple third-party AI models analysing a company's public footprint, weighted and aggregated as described on the methodology page, which is the public specification of the system. The multi-model design is itself an integrity measure: no single AI vendor's blind spots, biases or failures decide a company's standing, and models that disagree sharply about a company are a signal to weigh, not an inconvenience to average away silently. Scores are recomputed automatically as evidence changes — new reviews, removed reviews, fresh AI scans — and they move in both directions without human hands on the dial. Legally, a TrustScore is a statement of opinion produced by an automated system from the evidence available at the time: it is not a certification, a guarantee, an audit, financial advice, or a verifiable fact about a company's future conduct. We stand behind the honesty of the process, never behind the future behaviour of a scored company; the corresponding disclaimers in the Terms of Service apply. Where AI-generated summaries or assessments are displayed, they are labelled as such — imperfection is disclosed, not hidden, because a reader who knows what produced a number can weigh it, and a reader who does not is being managed rather than informed.

3. Score disputes and corrections

A company that believes its score reflects an error of process — reviews that break the Reviewer Guidelines, a coordinated attack, outdated AI analysis of a rebranded or reformed business — has real remedies, all free, and none requiring a plan, a lawyer or a relationship: report rule-breaking reviews with evidence under the Moderation & Appeals Policy; request a re-scan of the AI assessment from the business panel so the models evaluate the company as it is today; and correct factual company data via profile management or the contact page. What is not a remedy: disliking a lawful score. We do not negotiate scores, and threats — legal or otherwise — do not move numbers; process does. Where we find a genuine defect in our own computation, we fix it retroactively and say so; corrections of substance are documented in our public trust reports. The remedies are also honest about their limits: a re-scan re-evaluates the evidence, it does not promise a better verdict, and companies sometimes request one and score lower — the process serves accuracy, and accuracy is not always flattering.

4. Manipulation — the integrity offences

The following are integrity offences whoever commits them — consumer, business, agency, affiliate, competitor: fake, purchased or incentivised reviews (in either direction); review gating and cherry-picked invitations; coordinated attacks and review bombing; flag abuse against honest content; self-reviews and competitor sabotage; AI-targeted manipulation — publishing or structuring content designed to deceive the models that feed TrustScores, prompt-injection attempts against our systems, or flooding channels with synthetic material to skew analysis; widget and badge falsification under the Widget & Embedding Terms; and buying or selling accounts, claims or moderation outcomes. Consequences follow the enforcement framework of the Moderation & Appeals Policy and Section 11 of the Business Terms of Service, and for grave commercial fraud we reserve civil action and referral to consumer-protection authorities — fake reviews are illegal in most jurisdictions, not merely against our rules.

Two properties of integrity enforcement are worth underlining. It is symmetric: fake praise and fake attacks are the same offence, a business inflating itself and a competitor deflating it face the same machinery, and removing fraudulent five-star reviews lowers scores exactly as removing fraudulent one-star reviews raises them. And it is retroactive: when a fraud operation is uncovered, everything it touched is unwound — its reviews come down, the scores it distorted recalculate, commissions it generated are voided — because an integrity system that only polices the future ratifies every past cheat.

5. Who owns what

Clean ownership lines prevent both overreach and confusion, so here they are. Yours: your reviews, replies and other User Content — you license them to us as described in the Terms of Service, and you keep ownership; nothing in any of our documents transfers your copyright to us, ever. Ours: the platform — software, design, the compilation and structure of profiles and reviews (protected by database rights, which is why bulk extraction is prohibited under the Acceptable Use Policy), TrustScores and their presentation, AI summaries we generate, our documentation, and the TrustRating name, logo and trade dress. Licensed to you: what the Terms of Service, Widget & Embedding Terms and API terms each grant, no more. Facts remain free: quoting a score with attribution and date in journalism, research or comparison is fair use of information we publish — wholesale copying of our database or dressing your product in our design is not.

6. Brand rules

Brand rules exist for the same reason score rules do: a mark that anyone can counterfeit certifies nothing. You may, without asking: name TrustRating truthfully ("we are on TrustRating", "TrustScore 4.7 as of 3 August 2026"), link to any public page, and use the official widgets and assets your plan provides under their terms. You may not: use our name or logo to imply endorsement, partnership or certification that does not exist; register domains, social handles or products confusingly similar to ours; alter our logo, recolour badges, or fabricate "TrustRating approved" seals — no such seal exists; use our marks in advertising keywords in the ways restricted by the affiliate terms; or attach our marks to unlawful or deceptive material. Press and researchers are welcome to our public data with attribution; for anything beyond, ask via the contact page — we answer quickly and usually say yes to honest uses.

7. Copyright and IP complaints — notice and takedown

If you believe content on TrustRating infringes your copyright, trademark or other IP rights, send a notice via the contact page including: identification of the protected work and of the infringing material (URLs); your name, contact details and relationship to the rights holder; a good-faith statement that the use is unauthorised and that your notice is accurate; and your signature (electronic suffices). We remove or disable clearly infringing material expeditiously, notify the user who posted it, and accept counter-notices from users who believe the removal was mistaken — where a valid counter-notice is not followed by court action from the complainant within the statutory window, content may be restored. The procedure mirrors the DMCA's architecture because it is the best-understood framework worldwide, and we apply its equivalent balance in jurisdictions with their own regimes (including the EU's notice-and-action rules under the DSA). Repeat infringers lose their accounts. Knowingly false notices and counter-notices carry legal liability, and abusive takedown campaigns against honest reviews are treated as the integrity offence they are: copyright law is not a tool for deleting criticism, and a review quoting your marketing copy to criticise it is not infringement.

8. Research, press and public data

We want the platform's data to serve the public conversation about commercial trust, not just individual purchase decisions. Journalists may quote scores, review excerpts and our published statistics with attribution and date, and may contact us for comment or aggregate figures — we answer press quickly and do not condition access on favourable coverage. Academic and public-interest researchers can request structured data access for defined projects; we are generous with aggregates and careful with anything touching personal data, which stays governed by the Privacy Policy. What neither press nor research licences cover is republishing the database itself or laundering bulk extraction through a "study" — the line, as everywhere in this policy, is between using facts and taking the collection. Our own aggregate findings — fraud patterns, category trends, enforcement statistics — are published in the trust reports for anyone to use with attribution.

9. Reporting integrity violations — including ours

Integrity reports get priority handling. Anyone — user, employee, business, competitor of a cheater — can report review fraud, score manipulation, badge abuse or bribery attempts through the contact page; reports are confidential, reporter identities are protected as described in the Acceptable Use Policy, and reports about TrustRating's own staff or processes go to people with authority over those staff and processes. We do not retaliate against good-faith reporters, and where the law provides whistleblower protections (as EU law does), those protections apply on top of our own commitments, not instead of them. If you were offered a score-for-money deal by anyone claiming to represent us — report it with everything you have; that claim is either fraud against you or a breach we urgently need to know about, and both demand the same response.

10. Our own conduct — ethics commitments

The integrity rules bind us first — a platform that polices everyone's honesty except its own is running a protection racket, not a trust service. We commit that: TrustRating employees and moderators do not write reviews of companies on the platform, do not moderate cases where they have a conflict of interest, and do not trade on non-public score information; our sales team has no access to moderation tools and no influence on scores — the wall between revenue and integrity is organisational, not just rhetorical; we correct our mistakes visibly rather than quietly (Section 3); we publish how the system works (methodology), what we enforce (this policy and the Moderation & Appeals Policy) and aggregate enforcement outcomes in trust reports; and we submit ourselves to the same appeal discipline we offer you — every enforcement decision names its rule and its appeal route. If we ever fall short of this section, we want to hear it: contact us, and the report goes to people with the authority to fix it.

11. Changes

Amendments to this policy follow the change process of the Terms of Service, with one addition fitting its rank: we will never amend Section 1 to permit what it prohibits. The date above identifies the version in force.