# Defamation vs Opinion Online: Where the Line Falls

Defamation vs opinion online: how to tell a false statement of fact from protected opinion, what search engines and platforms do, and what to try instead.

---


Defamation vs opinion online comes down to one question: does the statement assert something about you that a reader could check and find to be false? If it does, it may be a false statement of fact and a candidate for removal or a legal claim. If it only expresses a judgment, such as "terrible service" or "I would never hire him", it is an opinion, and in many legal systems an honestly held opinion is protected from a defamation claim.

The line matters because the two kinds of content are treated in opposite ways. A platform, a search engine and a court will each look for a false fact before they act. A harsh review, a critical thread or an angry blog post that contains no checkable fact usually stays online, however unfair it feels, and a legal threat about it tends to make it spread.

This guide gives you the tests to sort a post statement by statement, shows how the line moves between the United Kingdom, the United States and Australia in outline, and explains what search engines and platforms actually do with opinion. It also says plainly when a route will fail, and it does not help anyone hide accurate or honestly held criticism on a matter of public interest. It is information, not legal advice: a lawyer in the right country decides what applies to your facts.

<div class="box"><b>The short version</b>
<ul><li>Break the post into single statements. Only a statement that a reader could prove true or false can be defamatory, and only if it is false.</li><li>Adding "in my opinion" does not turn a factual claim into an opinion, and calling a harsh opinion "false" does not turn it into a fact.</li><li>Search engines and platforms do not decide who is right in a dispute. They act on a policy violation, a clear legal basis or a court order.</li><li>When the content is opinion, the useful routes are a public reply, a correction of any facts inside it, and lawful work to move accurate pages above it.</li></ul></div>

## A statement is a fact if a reader could prove it wrong

The working test is checkability. A statement of fact asserts something that happened or exists: you took a payment and did not deliver, you were fired for theft, your license was revoked, your company is insolvent. Someone with the right documents could show whether each one is true.

An opinion is a judgment, a reaction or a conclusion that has no single provable answer. "The room was dirty" is close to the line, because cleanliness can be argued. "The staff were rude" is a judgment. "The hotel is a rip-off" is a judgment too, unless the writer adds that they were charged twice, which is a fact.

Three features carry most of the weight when a lawyer or a platform reviewer reads a post.

First, whether the claim can be checked at all. Dates, amounts, named acts, criminal allegations and statements about licenses, qualifications or finances can be checked. Taste, tone and overall verdicts cannot.

Second, whether the writer shows the basis for the view. A review that tells the story behind its verdict lets readers weigh it for themselves, which is the reason many legal systems protect honest opinion. A bare label such as "fraud" or "scammer" suggests there are undisclosed facts behind it, and a reader may take it as an accusation of something specific.

Third, the setting. Readers treat a heated reply in a forum thread differently from a signed article that presents itself as reporting. Courts and reviewers look at how an ordinary reader would understand the words in that place.

## Five checks that sort a post before you act

Run these checks on each statement in the post, one at a time. A post that looks like a single attack is usually a bundle of three or four separate statements, and only some of them are the problem.

<figure class="fig">
<figcaption><b>Sorting one statement at a time</b></figcaption>
<ol class="fig-steps">
<li><b>Copy the exact words</b>Capture the post, the address, the date and the account name, then copy each sentence into its own line. Paraphrasing hides the difference between a fact and a feeling.</li>
<li><b>Ask what a reader would take away</b>Write the one-sentence meaning an ordinary reader would draw. If the meaning is "this person committed a specific act", you are looking at a factual claim.</li>
<li><b>Test if it can be checked</b>Ask what document, record or witness would prove it right or wrong. If you cannot name one, treat the line as opinion.</li>
<li><b>Look for the basis</b>Note whether the writer states the facts behind the view. A stated basis that is true makes the opinion safer for the writer. A stated basis that is false moves the problem to the fact.</li>
<li><b>Check your own proof</b>Gather the evidence that the factual claim is false before you contact anyone. If it is partly true, say so to yourself now, because the other side will say it.</li>
</ol>
</figure>

The fifth check is the one people skip. A claim is only removable if it is false, and a post you cannot disprove is not a defamation case, whatever it does to your search results.

## How the same sentence is classified: a decision table

The wording below is invented, with generic parties, to show how the line usually falls. Real outcomes depend on context, country and the full post.

<div class="tablewrap"><table><thead><tr><th>Example wording</th><th>Usually read as</th><th>Why</th><th>Likely route</th></tr></thead><tbody>
<tr><td><strong>"The owner kept my deposit and never answered"</strong></td><td>Fact</td><td>Checkable with payment records and messages</td><td>Evidence file, platform report, then a letter</td></tr>
<tr><td><strong>"He is a fraud"</strong></td><td>Depends</td><td>Reads as opinion in a rant, as an accusation if it implies undisclosed facts</td><td>Ask what facts sit behind it before choosing a route</td></tr>
<tr><td><strong>"Worst service I have ever had"</strong></td><td>Opinion</td><td>A personal verdict with no single provable answer</td><td>Public reply; removal usually fails</td></tr>
<tr><td><strong>"She was dismissed for stealing"</strong></td><td>Fact</td><td>A specific act and an employment record can prove it</td><td>Correction request, platform report, legal letter if false</td></tr>
<tr><td><strong>"I think the firm cuts corners, judging by the delays I saw"</strong></td><td>Opinion with a basis</td><td>A conclusion tied to stated facts the writer observed</td><td>Reply with your side of the facts; removal usually fails</td></tr>
<tr><td><strong>"In my opinion, he was convicted last year"</strong></td><td>Fact</td><td>The opinion label does not change a checkable claim</td><td>Treat as a factual claim and test it</td></tr>
<tr><td><strong>"The company is a joke"</strong></td><td>Opinion or hyperbole</td><td>No reader takes it as a literal claim</td><td>Leave it, or address the underlying complaint</td></tr>
</tbody></table></div>

Two rows deserve attention. The "fraud" line is the ambiguous case that most disputes turn on. And the last factual example shows why the label "opinion" protects nothing by itself: the court or reviewer looks at what the words assert, not at the writer's preface.

## Why a label does not decide it in either direction

People make the same mistake from both sides. Writers think that "allegedly" or "in my view" makes a factual claim safe. Targets think that "false" or "defamatory" makes an opinion removable.

<aside class="callout">
<span class="callout-label">Rule</span>
<h4>The assertion decides, not the label</h4>
<p>A reviewer or court asks what the words claim about you and whether that claim can be proved wrong. A prefix such as "I believe" does not shield a false fact, and a heated tone does not turn an opinion into one.</p>
</aside>

A related trap is the post that mixes both. A review might read: "Awful experience, rude staff, and they charged me twice." The first two parts are opinion. The third is a fact you can check against your billing records. If you were charged once, that one sentence is where your effort should go, and a report that attacks the whole review will be refused.

The reverse also happens. A post that is accurate but unflattering, such as a correctly reported dispute or a true account of a past event, has no false statement to challenge at all. If the facts are true, no route in this guide will remove them, and we would not help anyone try.

## How the line moves between the UK, the US and Australia

Honest opinion is protected in different ways in each system, and the details differ.

In England and Wales, section 3 of the <a href="https://www.legislation.gov.uk/ukpga/2013/26/section/3">Defamation Act 2013</a> makes honest opinion a defense with three conditions: the statement was an expression of opinion, it indicated, generally or specifically, the basis for that opinion, and an honest person could have held the opinion on the facts that existed when it was published. The defense fails if the claimant shows the defendant did not actually hold the opinion. The section replaced the older common law defense of fair comment. Scotland and Northern Ireland have their own rules, which we did not review for this guide, so check them if your case sits there.

In the United States, the First Amendment shapes the line more than any single state statute. In <a href="https://www.law.cornell.edu/supct/html/89-645.ZS.html">Milkovich v. Lorain Journal, 497 U.S. 1 (1990)</a>, the Supreme Court declined to create a separate opinion privilege. Framing a statement as opinion does not dispel the factual implications it carries, and a statement on a matter of public concern must be provable as false before liability can be assessed. We read the Court's published summary of the case for this guide, not the full opinion. Individual states add their own rules, and a private dispute may be judged differently from a matter of public concern.

Australia also has an honest opinion defense, but we could not read its statutory text for this guide. The states and territories each have their own defamation statutes, so the conditions and wording should be checked in the statute of the state that applies to your case.

In the systems we read, the practical result points the same way: a post made up only of judgment, with a stated basis, is a poor candidate for a claim, and a post that alleges a specific false act is a better one.

## What search engines and platforms do with an opinion

A search engine is not a judge. It does not weigh whether a review is fair or whether a dispute was your fault. Google's <a href="https://support.google.com/legal-help-center/answer/13948866?hl=en">legal help page</a> lists defamation, copyright, personal data and court orders as common legal reasons to report content, and advises identifying the correct legal basis. For content that breaks a local law, it says Google typically restricts access only in the country or region where the content appears illegal. Restricting a result in a search engine does not take the page itself off the web.

What this means in practice is simple. Search engines tend to act on a clear legal basis, a court order or a policy violation. They tend not to act on a request that says "this is unfair" or "this is false" with nothing to prove it. For a post that is plainly opinion, expect a refusal, and expect that a long exchange will not change it.

Platforms behave in a similar way, with their own rules in place of the law. A review site, a forum or a social network usually has policies on harassment, personal information, impersonation, spam and fake accounts, and many have a route for reporting factual falsehoods that can be shown to be false. Few will remove a one-star rating because the business owner disagrees with it. The reliable test is to read the platform's own rules before you file, and to report under the rule that the content actually breaks, not under the word "defamation" as a general complaint.

<div class="tablewrap"><table><thead><tr><th>Who can act</th><th>Acts on opinion?</th><th>Acts on a false fact?</th><th>What it needs</th></tr></thead><tbody>
<tr><td><strong>Platform or site</strong></td><td>Only if it breaks a platform rule</td><td>Sometimes, if its policy covers it</td><td>The exact address and the rule broken</td></tr>
<tr><td><strong>Search engine</strong></td><td>Rarely</td><td>Sometimes, with a clear legal basis</td><td>A legal basis, evidence and the exact address</td></tr>
<tr><td><strong>The author</strong></td><td>If they choose to</td><td>If they choose to, or after a letter</td><td>A calm request with proof</td></tr>
<tr><td><strong>A court</strong></td><td>Almost never</td><td>Yes, after a full process</td><td>A claim that fits the local test and the clock</td></tr>
</tbody></table></div>

## A mixed post, worked through: an invented scenario

This scenario is invented, with generic parties, and the numbers are assumptions. A small contractor finds a long forum post by a former customer. It says the work was "shoddy and a disgrace", that the contractor "took a 30 percent deposit and disappeared for six weeks", and that the contractor "has no license". The contractor did take a deposit, and did go quiet for two weeks because of illness, and does hold a license.

Break it into statements. "Shoddy and a disgrace" is opinion. "Took a 30 percent deposit" is true. "Disappeared for six weeks" is partly false, since the gap was two weeks, though a reader may feel the difference is small. "Has no license" is false and checkable, and it is the most damaging line, because it suggests the contractor operates illegally.

The useful move is narrow. The contractor asks the forum to correct or remove the license claim, attaching a copy of the license, and posts a short public reply that sets out the real timeline of the gap without insulting the customer. The contractor does not ask for the whole post to come down, because most of it is opinion or true, and a request that includes those parts invites a refusal.

Here is a realistic sequence for this kind of case, written as ranges from experience, not promises.

<div class="tablewrap"><table><thead><tr><th>Stage</th><th>What happens</th><th>Usually takes</th></tr></thead><tbody>
<tr><td><strong>Sort and evidence</strong></td><td>Statements separated, license record and message history gathered</td><td>One to three days</td></tr>
<tr><td><strong>Public reply</strong></td><td>Calm, factual reply posted with the license detail</td><td>Same day</td></tr>
<tr><td><strong>Platform report</strong></td><td>Report filed on the false license claim only, with proof</td><td>Days to a few weeks for a reply</td></tr>
<tr><td><strong>Direct request to the author</strong></td><td>Polite message asking for a fix of one line</td><td>A week or two for any answer</td></tr>
<tr><td><strong>Legal letter</strong></td><td>Considered only if the false line stays and harm is real</td><td>Weeks, with counsel</td></tr>
</tbody></table></div>

The result you can expect is partial. The false license line may be fixed or removed. The rest stays, and the best response to the rest is a visible, accurate answer.

## What you can do with an opinion you cannot remove

When the content is opinion, the goal changes from taking it down to reducing its weight. Three lawful approaches work, and none involves pressuring the writer.

**Answer in public.** A short, factual reply, written for the next reader and not for the author, often does more than a removal would. State what happened, what you changed and how to contact you. Do not argue about tone, and do not attack the writer. Readers judge the reply as much as the post.

**Correct any fact inside it.** If one sentence is false, deal with that sentence alone, with proof. This is the narrowest and most successful request, because a reviewer can verify it quickly.

**Strengthen what people find first.** If the opinion ranks for your name, lawful and accurate pages of your own can sit above it over time: a clear profile, a business listing, independent coverage that is true. This is suppression, not removal, and it works slowly and without a guarantee. Our <a href="/suppress-negative-results">suppression service</a> works on this basis, and it never involves fake content or paid manipulation.

Some approaches should be avoided. Fake reviews to dilute a bad one break platform rules and consumer law in many places. Pressuring the writer, or contacting their employer, can turn a weak complaint into a stronger one against you. A threatening letter about a clear opinion is often posted online as a screenshot, and the second story then becomes the story.

## When removal should fail, and why we would not help

Some content is not removable, and some of it should not be. A true account of something you did, a truthful report on a court case, an honest review of a real experience, and a good-faith opinion on a matter of public interest are outside what any legitimate route removes. A request that tries to hide them risks a refusal, a public note about the attempt, and sometimes a counter-claim.

We apply the same line in our own work. We help where a statement is false, outdated, unlawful or a breach of privacy, and we decline files that aim to hide accurate public-interest information about wrongdoing. Our published <a href="/ethics">ethics page</a> describes how we draw that line.

Three signals suggest a route will fail. The post is signed and reads as the writer's account of a real event. The facts in it are largely true. Or the writer is a public critic, a journalist or a customer commenting on a matter that affects others. In each case, a reply and a correction of any single false fact are better than a takedown attempt.

## Your next step: sort one post, then pick one route

Take the post that worries you and copy it into a document. Split it into single statements, mark each as fact, opinion or mixed, and note the proof you hold for each factual line. Then choose one route that matches the strongest false statement, and leave the rest.

<div class="box hot"><b>If this is happening to you now</b>
<p>If the post contains a false, checkable claim that is damaging your work or your name, send us the address and your proof, and we will tell you in plain terms which part is removable and which part is not. Start with the private <a href="/audit">audit</a>.</p></div>

Reputation Partners reviews each file this way before anything is filed: the false statement first, the route second, and an honest answer when the content will stay.

## Questions people ask

### Can a negative review be defamation?

Yes, if it contains a false statement of fact about you, such as an invented charge or a denied service that did happen, and it causes real harm. A bad rating or harsh verdict alone is not. The decisive step is separating the checkable lines from the judgment, then acting only on the false facts.

### Is calling someone a liar defamation?

It depends on the setting and the words around it. In a heated exchange, a reader may take "liar" as abuse. In a signed account that lists what the person supposedly lied about, it can be read as a factual accusation. The label alone rarely decides it, so look at the sentences that explain it.

### Does "allegedly" protect a writer?

Not by itself. A word such as "allegedly" or "reportedly" may soften a claim, but a reader can still take the sentence as saying you did the act. What helps a writer more is a true basis, a fair account of the dispute, and a record of where the information came from.

### Can I force a site to remove an opinion I disagree with?

Rarely. Sites and search engines act on their own rules or on legal grounds, and disagreement is neither. If the post breaks a rule, for example by exposing private data or impersonating you, report it under that rule. Otherwise a reply and a correction of any false fact are the realistic options.

### What if part of the post is true and part is false?

Challenge only the false part. Ask for the single line to be corrected or removed, attach proof, and leave the true statements alone. A request that tries to remove everything is easier to refuse, and it signals that you are hiding the accurate parts. Reply publicly to the rest.

### Does it matter if the writer is anonymous?

It matters for who you can ask, not for how the post is classified. An anonymous false statement is judged by the same tests as a named one. Finding the writer is a separate legal step with its own conditions, and the content may still be opinion once you do.