Research and editorial standards

ReviewRemoval publishes reporting guides and secondary research for businesses dealing with online reviews. We also sell review-reporting services. That commercial interest should be clear when you read our work. A platform-wide statistic is not evidence of our service success rate or a promise about your case.

Who is responsible for an article?

Articles credited to ReviewRemoval are organizational publications. We do not invent individual authors, professional credentials, peer reviewers or independent endorsements. The ReviewRemoval team is responsible for the article text, calculations and corrections. The author link on an article identifies this publication responsibility, not a claim of academic peer review. Our About page explains the service.

Start with the primary source

We link to the platform report, regulator, official help page or research publication behind a statement. Activity year, publication date and date checked are different fields. A 2026 article describing 2025 activity does not make those activity data a 2026 measurement. Where a report is revised, we identify which version we used and retain a correction trail when it changes the analysis.

For example, our 2024–2025 Google data audit separates current transparency displays, superseded values and announcement thresholds. The source labels are part of the dataset, not optional footnotes.

What makes an analysis original?

Original work can be a documented compilation, a reproducible calculation, a comparison with compatible definitions or a mathematical model. It does not mean that we collected new customer observations. We distinguish source-reported observations from our arithmetic, assumptions and constructed examples. We do not describe a synthetic example as a survey, a client cohort or a platform experiment.

Our rating-impact study explains what can be calculated from a displayed star rating under stated rounding assumptions. Its ranges are mathematical results, not measurements of businesses that used our service.

Keep the population and denominator visible

“Removed reviews”, “blocked submissions”, “all contributions” and “appealed decisions” describe different populations. We do not combine their counts as if they were the same unit. Rounded source displays stay labelled as rounded. A strict upper or lower bound stays a bound rather than becoming an exact point value. We state whether a percentage change means percentage points or relative change.

Global platform data do not establish a Canadian, provincial or city-level removal probability. A statement about Toronto or another community requires evidence about that community. Our Canadian service coverage describes where remote assistance is offered, not a measured local success rate or a physical office.

Make the calculation checkable

Research pages provide the relevant source links, methods, limitations and downloadable tables or code when available. Charts have written descriptions and are accompanied by tabular values or data downloads. We preserve nulls, inequalities and separate record types so a spreadsheet user can interpret a figure correctly. Citation files help identify the specific release; they are not evidence that an external institution has endorsed it.

A useful citation includes the publisher, title, version, relevant activity period, source population and important limitation. Attribute the underlying observations to their original publisher and our compilation or calculation to ReviewRemoval. Original research explicitly released under CC BY 4.0 may be reused under that licence, with attribution and changes identified. Source publishers retain their own rights.

Service claims require service evidence

A removal count from Google or another platform cannot substantiate an agency success rate. A service-outcome claim would need a defined eligible cohort, request dates, accepted cases, successes, failures, unresolved cases, follow-up period and treatment of reviews that return. We do not infer our own results from public platform totals. Pricing examples are calculations, not observed case outcomes.

A reporting guide describes the route and evidence that may be relevant. It does not guarantee that a platform will remove criticism or treat a report as eligible. For a specific matter, use the evidence preparation tool and the platform's current official instructions.

Updates and corrections

Source-check dates state when the listed material was checked; they do not imply continuous monitoring. We change an article's modification date when its substance changes, rather than routinely refreshing dates to appear current. A changed number should be reflected in the text, chart, table and downloadable data together. Existing releases should remain identifiable when a later version replaces them.

To report an error, contact us with the article URL, the statement or record ID, the primary source and the correction you believe is needed. Put “Research correction” in your note. Please omit private customer records and passwords. We will examine the source and amend a material error; a disagreement or correction request is not a promise that the proposed interpretation will be adopted.

Examples of these standards in use

• Google removal statistics: versions, denominators and reporting corrections.
• Canadian tourism and Google Reviews: the scope of the survey figures.
• Trustpilot appeals: separate decision types and missing denominators.

Published October 27, 2026. These standards apply to ReviewRemoval publications. See our Terms of Service and Privacy Policy for the request and service conditions.