Corporate reputation management: what a company can actually remove online

01Actionable, or merely negative

A company usually arrives with a mixed list: two articles, nine reviews, a forum thread, an ex-employee post, and a search page that will not move. Procedurally they have nothing in common, and a request sent down the wrong route is refused on the spot.

02Opinion against a false statement of fact

In Milkovich v. Lorain Journal Co., 497 U.S. 1 (1990) the US Supreme Court held that the First Amendment requires no separate privilege for anything labelled opinion. Two principles do that work instead: on a matter of public concern a statement must be provably false, and statements that cannot reasonably be read as stating actual facts, such as rhetorical hyperbole, are protected. Calling something an opinion does not strip out its factual implications.

The Defamation Act 2013 works from the other direction. Section 1: a statement is not defamatory unless publication has caused or is likely to cause serious harm to reputation, and for a body trading for profit only where that is serious financial loss. Section 3 gives an honest opinion defence on three conditions: it was a statement of opinion, it indicated in general or specific terms the basis of the opinion, and an honest person could have held it on that basis.

The working test on each line is whether a neutral reader would take it as a checkable assertion. The invoice was never paid is checkable. They are a nightmare to deal with is not. Most damaging posts contain both, which is why they are argued line by line.

03What review platforms will remove

Google removes content breaching its prohibited and restricted content policies: fake engagement, impersonation, misrepresentation, harassment, personal information, profanity, advertising, repetitive content, off-topic posts. Fake engagement covers content not based on a real experience, paid or incentivised reviews, and multiple accounts posting for one person. Businesses are also barred from soliciting fake content or offering incentives for reviews, edits or removal of negative ones.

Google states it allows content describing negative experiences in a respectful manner. A review can still go if it is offensive or profane, contains personal information, or makes unsubstantiated allegations of unethical behaviour or criminal wrongdoing. That last ground is the one usually available, and it is narrow. Reviews are reported from the profile under Read reviews or in the Reviews management tool; Google says evaluation typically takes several days, and the tool allows a one-time appeal for up to ten reviews at once.

Trustpilot requires a review to rest on the reviewer's own genuine experience, which may be a call, email or chat rather than a purchase, treated as recent within roughly twelve months. A business may flag on defined grounds: harmful or illegal content including defamation, which only businesses may report; other people's personal information; advertising; no genuine experience the business recognises; or a review about a different business. Flagging does not guarantee removal, and Trustpilot states businesses have no control over it. Yelp publishes its own content guidelines, and the grounds differ again.

04Fake reviews are now a regulatory matter

The FTC's Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, was published on 22 August 2024 and took effect on 21 October 2024. Section 465.2 covers fake or false reviews and testimonials; 465.4, buying reviews where the incentive is conditioned on a particular sentiment; 465.5, undisclosed insider reviews; 465.6, company-controlled sites presented as independent; 465.8, trading in fake indicators of social media influence.

In the UK, paragraph 13 of Schedule 20 to the Digital Markets, Competition and Consumers Act 2024, in force from 6 April 2025, lists among practices unfair in all circumstances: submitting or commissioning a fake review, meaning one that purports to be but is not based on a person's genuine experience; concealing that a review was commissioned, where commissioning includes incentivising by any means; publishing reviews in a misleading way, such as selectively removing negative ones; publishing without reasonable and proportionate steps to prevent or remove fake reviews; and offering such services to traders.

Under Part 3 of that Act the CMA may impose, for a relevant infringement, a penalty of a fixed amount not exceeding £300,000 or, if higher, 10 per cent of turnover.

05News articles: correction, reply, or nothing

Publishers almost never delete. A deletion demand is refused in a line; the same facts framed as a specific factual error often are not. Clause 1 of the IPSO Editors' Code of Practice is the standard a UK complaint is measured against: the press must take care not to publish inaccurate, misleading or distorted information or images, including headlines not supported by the text; a significant inaccuracy, misleading statement or distortion must be corrected promptly and with due prominence, and where appropriate an apology published; a fair opportunity to reply to significant inaccuracies should be given when reasonably called for; and comment, conjecture and fact must be clearly distinguished.

So the complaint must identify a significant inaccuracy, not a tone, and the best realistic outcome is a correction attached to the article. For search purposes that is often better than deletion, which tends to be reported on elsewhere.

De-indexing is separate from changing a page. Google documents that its removals tool, used by the site owner, hides a page for about six months, and that the durable options are removing or updating the content, password-protecting it, or a noindex tag, which keeps the page out of results while leaving it reachable. For sites the company does not own, Google directs requesters to its legal and privacy routes, advises naming the specific URLs and explaining precisely what is violative and why, and notes that restricted content may still exist elsewhere.

06Copied content and the copyright route

Where a page reproduces the company's own material, the copyright route is the fastest that exists, because it is a statutory process with a timetable rather than a discretionary policy. Under 17 U.S.C. 512(c)(3)(A) a notification to the provider's designated agent must be written and include a signature of a person authorised to act for the rights owner; identification of the work and of the allegedly infringing material in enough detail to locate it; contact information; a statement of good faith belief that the use is not authorised; and a statement that the information is accurate and, under penalty of perjury, that the sender is authorised to act for the owner of an exclusive right.

The counter-notification path matters as much. Under section 512(g), where a subscriber files a counter notification with the required elements, including consent to federal court jurisdiction, the provider restores the material not less than 10 nor more than 14 business days after receiving it, unless the designated agent first receives notice that the complainant has filed a court action.

07The EU route: notice and action under the DSA

For platforms in scope of the Digital Services Act, Regulation (EU) 2022/2065, there is a documented process with an answer at the end. The rules applied to very large platforms and search engines from the end of August 2023 and to all platforms in scope from 17 February 2024.

Trusted flagger status is for entities established in the EU and awarded by a national Digital Services Coordinator, so it is not something a company obtains for its own disputes. The statement of reasons is the useful artefact: a refusal that must be written down and reported is one that can be appealed.

08Suppression through owned assets

Where nothing is removable, the remaining lever is what else the search page contains: pages the company controls and keeps accurate, content answering the search a reader actually runs rather than content aimed at the critic, named-person pages for executives whose names return the result, and a dated factual statement on the company's own site where an allegation is widely repeated. It is slower and less certain than removal, and the only honest option for genuine criticism.

Section 465.7 of the FTC rule draws the line inside this area: it prohibits misrepresenting that displayed reviews represent most or all reviews where reviews have been suppressed on the basis of rating or negative sentiment, while neutral criteria such as removing fake or off-topic content are treated differently. Suppressing in search is a publishing choice; suppressing on the company's own review display while implying completeness is prohibited conduct.

09Legal letters, and the risk of backfire

The most direct risk is written into regulation. Section 465.7 also prohibits unfounded or groundless legal threats, physical threats, intimidation, or knowingly or recklessly false public accusations used to prevent a review from being written or to cause its removal. A threatening letter aimed at a genuine review is not merely ineffective; it is the conduct the rule describes.

1. Confirm the statement is factual and checkable, not opinion.
2. Confirm it is false, with the record that shows it.
3. Confirm serious harm, and for a profit-making body serious financial loss, under English law.
4. Check limitation: one year from accrual for libel or slander.
5. Decide, before sending, what happens if the recipient publishes the letter.
6. Send the narrowest letter that states the error and what is sought.

The second risk is amplification: a letter about a post nobody read can become a story about a company threatening a critic, and that story will outrank the original. Where reach is small, doing nothing is often better reasoned, and should be recorded as a decision rather than as inaction.

10Measuring whether anything changed

This work fails most often in the reporting. Removal counts prove activity, not effect. What matters is what a person sees when they search the company name, captured the same way each time.

A removed page can stay in results until the engine recrawls, so persistence for a few days is not a failed removal. And a page that drops off the first results page has not been deleted, so the log must distinguish removed at source, de-indexed and outranked. A company that conflates them will believe a problem is closed when it is not.