Libel, slander, and where a social media post actually sits
Almost everyone who contacts us about slander on social media has been libelled. The distinction is old and it still decides things. Libel covers defamatory material published in permanent form: writing, print, images. Slander covers the transient spoken word. A tweet, a Facebook post, a Reddit comment, a one-star review, a caption under a photograph: all permanent, all libel. So online slander and social media slander are, nine times in ten, the wrong label for the right problem, and the label does not change the remedy you need.
There is one genuinely arguable case, and it is worth knowing because nobody writes it down. A live stream is spoken and transient at the moment it goes out. Section 166 of the Broadcasting Act 1990 treats words published in a programme service as publication in permanent form, which makes broadcast defamation libel — but a personal TikTok Live or an Instagram Live is not obviously a licensed programme service. The argument rarely survives contact with reality. The moment that stream is saved, clipped, replayed or captioned, the saved version sits in permanent form and you are back in libel. Worth noting too: since the Defamation Act 2013 imposed a serious harm test on every defamation claim, the old procedural gulf between libel and slander matters far less than it once did.
The vocabulary is a mess, and the mess costs people weeks. Cyber libel, online libel, cyber defamation, digital defamation, website defamation, internet libel, libel on social media, defamation on the internet, defamation of character on social media: the same civil wrong wearing different coats. The phrase people type varies just as much, from defamation social media and defamation online through to social media defamation of character, defamation of character online and online defamation of character. Search engines treat them as one question. English law treats them as one tort. Only one of those terms carries a genuinely different meaning. Cyber libel is a specific criminal offence in the Philippines under section 4(c)(4) of the Cybercrime Prevention Act of 2012, carrying a term of two years, four months and one day up to eight years. If you are in Manila, the phrase means something precise. In Manchester it does not.
Does the label matter in practice? Only in two places, but they are places where getting it wrong costs money. The first is the pleading: a claim has to be brought as libel or slander, and slander historically required proof of special damage except in a small set of categories actionable in themselves. The second is the evidence you gather. Libel is proved by the artefact: the post, the archived page, the screenshot with its URL and timestamp. Slander is proved by witnesses, which means people willing to state what they heard and when. If someone stood in a warehouse and accused you of theft in front of colleagues, you need those colleagues; if they typed the same accusation into a group chat, you need the chat. People arrive with the wrong kind of evidence surprisingly often, having spent weeks collecting it.
One more piece of terminology worth retiring. Defamation of character is not a separate tort in England and Wales — it is defamation, and the character is the reputation the law already protects. American sources lean on the phrase heavily, which is why it dominates the search results, but no English pleading contains it.
What matters far more than the label is what the post is attached to. A defamatory allegation sitting inside an impersonating profile is two problems, and the impersonation half moves faster — the triage we run in LinkedIn defamation and impersonation takedowns. A defamatory caption under a stolen photograph is a copyright problem wearing a defamation coat, and copyright is the faster lever. The same principle splits an Instagram case four ways, because a profile, a post, a Reel and a Story each take a different route through Meta, as the Instagram account removal service breakdown sets out. Grasping why a TikTok defamation claim isn't a guidelines report saves the fortnight most people lose filing the wrong form. And once material spreads, duets, stitches and the republication problem decide whether removal ends anything at all.
The four things a defamation claim has to clear
A statement is not defamatory because it is unkind, unfair, or untrue. Under English law it has to clear four separate hurdles, and most of the cases that reach us fail on the fourth rather than the first.
Before the tests themselves, one clarification that saves a great deal of confused reading. Online defamation law is not a separate body of rules: it is ordinary defamation law applied to a newer medium. There is no distinct internet defamation law in England and Wales, no standalone social media defamation law, and no special statute for posts as opposed to newspapers. The Defamation Act 2013 governs all of it. What changes online is evidential rather than doctrinal, and it changes in ways that mostly favour a claimant: the words are captured verbatim, the timestamp is recorded, the audience is measurable, and the republication is traceable. A pub rumour leaves nothing behind. A post leaves everything behind.
It must be published to at least one person other than you. Social media clears this trivially. A public post is published to everyone who sees it, and even a locked account with forty followers has published to forty people. It must identify you, though not necessarily by name; if people who know you would reasonably understand the post to be about you, that is enough, and a nickname, a workplace, a photograph or a recognisable detail will do it. It must carry a defamatory meaning, which means the ordinary reasonable reader must take it as lowering you in the estimation of right-thinking people. And under section 1 of the Defamation Act 2013, its publication must have caused, or be likely to cause, serious harm to your reputation.
Serious harm is the hurdle that ends most claims, and it is where honest advice diverges hardest from what a distressed person wants to hear. The statutory wording is exact: a statement is not defamatory unless its publication has caused or is likely to cause serious harm to the reputation of the claimant. For a business the bar sits higher still, because section 1(2) provides that harm to a body trading for profit is not serious harm unless it has caused or is likely to cause serious financial loss. A company therefore needs evidence of lost contracts, cancelled orders or measurable revenue damage, not merely a bruised brand. A post with eleven views and no engagement rarely clears the threshold. A post that circulated through your industry, cost you a named contract, and can be shown to have done so, generally does. This is why the first thing a competent desk asks for is not the URL but the reach.
What does serious-harm evidence actually look like on paper? Not a feeling, and not a follower count. In the files that clear the threshold we typically see four things assembled together: platform analytics or archived engagement figures showing genuine reach into the claimant's own community rather than to strangers; dated third-party messages in which someone repeats or asks about the allegation, which proves it travelled; a documented consequence with a date attached, such as a rescinded offer, a cancelled booking or a supplier putting an account on hold; and, for a company, a revenue or pipeline comparison spanning the publication date. A single one of those is an argument. All four together is a case. The claimants who struggle are almost never the ones with the nastiest post. They are the ones who can prove it existed but cannot prove anybody who mattered read it.
The mirror image of these tests is the defence side, and you should know what the poster will say before you spend anything. Section 2 gives a complete defence if the statement is substantially true. Section 3 protects honest opinion, provided the statement is recognisably opinion and indicates the basis for it. Section 4 protects publication on a matter of public interest where the defendant reasonably believed that publishing was in the public interest. Between them these three account for most of the material people ask us to remove — and none of them can be argued away by filing more reports.
Volume in particular does nothing. Platforms assess whether a specific rule was broken, not how many people complained, a threshold model set out in Meta counts violations, not report volume and again in TikTok judges the violation, not the tally. Where material genuinely does breach a rule, the question becomes which lawful instrument fits, and which of the three lawful TikTok routes fits your case walks that decision through on a single platform. None of this is legal advice, and we say so plainly in what these briefings are, and what they are not.
Why the platform is almost never the defendant
Here is the fact that reorganises everything else, and the one most guides skip: in the United States you cannot sue the platform for what a user posted. 47 U.S.C. § 230(c)(1) provides that no provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider. X, Meta, Reddit, Google and TikTok are all US companies. A defamation claim aimed at the company hosting the post fails at the threshold, and subsection (e)(3) blocks states from legislating around it.
That single provision explains a pattern that otherwise looks arbitrary. Why does a copyright complaint move a US platform within days while a defamation complaint earns a form letter? Because section 230(e)(2) carves intellectual property out of the immunity. Copyright has a statutory notice-and-takedown machine behind it; defamation, in the US, has nothing equivalent. When a case contains a genuine copyright or trademark hook — a stolen photograph, a lifted logo, a copied bio — we use it, not because it is clever but because it is the only lever with a legally mandated response attached.
What section 230 does not do is worth stating just as clearly, because a lot of people read it as a general amnesty and give up too early. It does not protect the person who wrote the post; the author remains fully liable, and suing the author is the ordinary route in the United States. It does not protect a platform for content the platform itself created or materially developed. It does not cover federal criminal law, and it does not cover intellectual property. Nor does it stop a platform choosing to remove something voluntarily under its own rules — which is precisely why platform-policy reporting works at all, and why the practical American strategy is to pursue the poster in court and the content through policy, in parallel rather than in sequence.
The United Kingdom took the opposite route, and this is where a UK claimant has an option their American counterpart lacks. Section 5 of the Defamation Act 2013 gives website operators a defence for material they did not post themselves — but that defence is defeated where the claimant could not identify the poster, served a compliant notice of complaint, and the operator failed to respond in accordance with the regulations. The effect inverts section 230: it creates a real incentive to act. An operator that ignores a properly drafted notice loses its shield.
The catch is jurisdictional reach, and honesty requires stating it. Section 5 binds operators the English courts can reach. A US platform served with an English section 5 notice will usually treat it as correspondence rather than obligation, because the immunity it enjoys at home is stronger than the incentive we can apply from here. Where section 5 genuinely bites is on UK-hosted forums, review sites, local news comment sections and small independent websites — and those account for a surprising share of the damage in a typical case, because they rank.
Two adjacent regimes are worth knowing. The EU Digital Services Act imposes notice-and-action duties on platforms operating in Europe, a lane most guides ignore and one we use routinely; the European lane most guides skip shows how it changes a Facebook filing, and why Snap publishes no reporting API covers the same duties on a platform with no automation surface at all. The UK Online Safety Act 2023 is the other, and it is the one people most often misread. Its illegal-content duties have bound platforms since 17 March 2025, with Ofcom able to fine up to £18 million or ten per cent of qualifying worldwide revenue. But the priority offences in Schedule 7 are criminal offences: threats to kill, harassment, stalking, fraud, intimate-image abuse. Defamation is a civil tort. It is not on that list, and the Online Safety Act gives you no defamation takedown route. If your case also involves harassment or intimate images, it gives you a very good one.
Reporting defamatory posts on social media, platform by platform
Every major platform has a reporting flow, and none of them has a button marked defamation that does what you want. What they have instead are rule categories, and your job is to find the one the post genuinely breaches. Harassment, impersonation, privacy violations, threats, hate speech, non-consensual intimate imagery and coordinated abuse all carry enforcement weight. A false factual allegation about you, standing alone, usually does not, because the platform will not adjudicate whether it is true.
The practical technique is the same everywhere. Report the specific post rather than the account when one post is the problem, and the account when the whole profile is. Pick the narrowest category that actually fits. In the free-text box, quote the exact words and name the rule they breach, because reviewers weight specific evidenced reports far above vague ones. Then stop. Filing the same report from six accounts does not increase the odds; on several platforms it lowers them, and on a few it gets the reporter actioned instead.
Where the platforms diverge is in what a report physically transmits and who reads it. On Instagram, the routes Meta actually acts on are narrower than the menu suggests, which is why the Instagram report routes Meta actually acts on matters more than report volume. Snapchat's problem is time rather than category, since the content frequently expires before review — the four lawful Snapchat takedown routes work around that. Telegram is unusual in routing different harms to different addresses entirely, and Telegram channel takedown routes sets out which one compels what. WhatsApp transmits far less than people assume — what a WhatsApp report actually sends is a short slice of recent messages, not your case. And on X, blocking is the only outcome you directly control, with the X report categories that work doing the rest.
What happens after you file is the part nobody describes, and the silence causes needless panic. Most reports are triaged by automated classifiers first, and the overwhelming majority of enforcement decisions on the large platforms are made without a human ever seeing the report. You will usually receive an acknowledgement within minutes and an outcome within one to several days, though a queue can stretch to weeks when a platform is dealing with a surge. A no-action decision is not the end: every major platform now runs an appeal path, and appeals are reviewed by different systems and sometimes by people. In our experience an appeal that adds specific new context — naming the rule, quoting the words, explaining the real-world consequence — converts materially more often than the original report did, because the first filing was scored by a classifier and the second is read by a reviewer. What does not help is refiling the identical report from the same account, which is deduplicated and, on some platforms, counted against you.
One tactical note from our own filings. Where a defamatory post also contains a photograph of you that you took, or that was taken from your private account, the copyright route frequently outruns the defamation route by a wide margin, days against months, because it lands in a queue with a statutory clock attached rather than a discretionary one. We do not manufacture copyright claims to exploit this, and neither should anyone else; a bad-faith notice is perjurable in the US and reversible everywhere. But where the hook is real, it is the fastest lawful door in the building.
How to remove defamatory content from Google
Online defamation removal has two halves, and conflating them is the single commonest mistake we see. Removing a post and removing the search result are two different jobs, and people routinely finish the first and assume the second happened. It does not happen automatically. Even after the original comes down, the URL can sit in the index for weeks, and cached copies, syndicated reprints and scraper sites keep it alive far longer.
Google's position on defamation is narrower than most reputation-management sales pages admit, so here it is plainly. Google does not adjudicate whether a statement is defamatory. It has neither the authority nor the machinery to decide truth. In the United States its legal removal process does not accept a private individual's bare defamation claim as grounds for de-indexing. What it does accept is a court order: once a court has determined the material is defamatory, that order can be submitted through the Google legal removal process and acted on. Enforcement is then geographically scoped, restricting access in the country where the content was held unlawful rather than worldwide.
Two further routes are frequently confused with each other and with defamation removal, and separating them saves real time. The European and UK right to be forgotten, established by Google Spain in 2014, is a data-protection mechanism: it delists a URL from name-based searches on European Google domains where the information is inadequate, irrelevant or excessive. It is delisting, not deletion, and it is not a defamation remedy — though for an old, inaccurate, low-public-interest page about a private individual, it often achieves what a defamation claim could not. Separately, Google's Results about you tool handles personal information rather than reputation: phone numbers, home addresses, email addresses, and since an expansion in February 2026, government-issued ID numbers such as passport, driving licence and Social Security numbers, along with an easier path for non-consensual explicit imagery. It will not remove a false allegation about your conduct. It removes your data, not your reputation.
Getting the URL out of the index once the page is gone has its own mechanics, and skipping them is why people tell us the removal did not work. If the page has genuinely been deleted, the fastest route is Google's Remove Outdated Content tool, which asks Google to recrawl a URL that now returns 404 or 410 and drop it. If the page still exists but the defamatory passage has been edited out, the same tool can refresh the stale snippet rather than remove the result. If the site owner cooperates, a noindex tag or a removal request from within their own Search Console is faster and more durable than anything you can file from outside. None of this touches copies. Syndicated reprints, aggregator mirrors, scraper sites and archive captures each carry their own URL, and each needs its own request — which is why a case that looks finished at one URL is often only a third done, and why the count of live copies, not the original, is the number we track.
Where a defamatory page cannot be removed or delisted, the honest remaining option is suppression: building and strengthening accurate, authoritative material that outranks it. That is slower, it is never guaranteed, and it does not delete anything. The mechanics run in reverse to ordinary search work, which is covered in removal versus suppression on Instagram, and the de-indexing lane specifically in de-indexing a Telegram channel from Google. On Google-owned property the picture changes again, since strikes and court orders behave differently there — see what actually terminates a YouTube channel and how to get a YouTube video taken down and keep it down, which is the difference between a removal and a durable one.
Website defamation, and the notice that actually carries weight
When the material sits on a website rather than a platform, the balance of power shifts in your favour, and most people never find out. Small sites have no legal department, no immunity they can rely on in England, and a genuine incentive to make a properly served complaint go away.
The instrument is the section 5 notice of complaint, and its power comes from the timetable attached to it. Under the Defamation (Operators of Websites) Regulations 2013, an operator that receives a compliant notice has 48 hours to act on it. It must pass the complaint to the poster, who has until midnight at the end of the fifth day after notification to respond. If the poster does not respond, responds incompletely, or consents to removal, the operator must remove the statement within 48 hours of that deadline or that response. If the operator cannot contact the poster at all, it must remove within 48 hours of receiving the notice. Miss those steps and the section 5 defence is gone.
Those deadlines are set out in the schedule to the Defamation (Operators of Websites) Regulations 2013, and they run in working hours rather than wall-clock hours, so weekends and bank holidays do not count against the operator. The design is deliberate: Parliament wanted a route that resolves an anonymous complaint quickly without forcing a small site to adjudicate truth. What the operator is being asked to decide is not whether the allegation is defamatory. It is only whether the poster will stand behind it and identify themselves. That is a question a site owner can answer in an afternoon, which is exactly why the mechanism works where a threatening letter does not.
A notice only works if it is compliant, and the regulations are specific about content. It must specify the complainant's name, set out the statement complained of and where on the website it appears, explain why it is defamatory, and state whether the complainant consents to their name and contact details being passed to the poster. Get one of those wrong and the clock never starts. This is the single most under-used remedy in UK online defamation practice, and it costs nothing but care to serve.
Finding who to serve is usually easier than people expect. A public WHOIS lookup gives the registrar and often the hosting provider; the site's own terms or privacy page frequently names an operating company; and a UK company can be checked at Companies House in a minute. Where the operator ignores you, the hosting provider is the next lever, because most hosts have acceptable-use policies that prohibit unlawful content and a genuine commercial interest in not being party to it. Serving a host is not the same as serving the operator and it carries no statutory clock, but a well-evidenced complaint to a host has ended more small-site cases in our files than any letter to the site owner. Escalating to a registrar is a last resort and rarely proportionate. Pulling a domain over one page is an outcome most registrars will refuse, and asking for it tends to end the conversation.
Two practical cautions from doing this repeatedly. First, the notice route is only available where you cannot identify the poster; where you can, the operator keeps its defence and your argument is with the author. Second, removal is not always the end, because a counter-notice or a re-post can put material back — the put-back dynamic examined in which Facebook removals hold and which reverse. Where the poster is anonymous and the material is serious enough to litigate, the unmasking step comes before the claim, and the five lawful instruments that reach X covers how a section 5 notice and a disclosure order sit alongside each other.
If a false allegation about you is live right now, get a confidential case review. We map the lawful route before any fee is named, we never forge a notice or bluff a lawsuit, and we will tell you plainly when the honest answer is suppression rather than removal.
When you need a defamation solicitor, and when you do not
Search for an online defamation lawyer, a defamation lawyer online, an internet defamation lawyer, a social media defamation lawyer or an online defamation attorney and you will meet the same twenty firms, all of them competent and none of them cheap. Anyone hunting a defamation attorney online is usually doing it at two in the morning, three days into the problem, and about to spend money on the wrong tier. Before you book anything, it is worth knowing which tier you are actually in, because two of the three do not need a lawyer at all.
Tier one — the content breaches a platform rule. Impersonation, harassment, privacy violations, intimate images, a stolen photograph. Route: a correctly evidenced report or a copyright notice. No solicitor needed, no fee to a law firm, and the fastest outcomes we see. Most cases that arrive convinced they need litigation are actually here.
Tier two — the platform will not act, but the material is removable another way. Route: a section 5 notice to a UK operator, a Digital Services Act notice in Europe, a right-to-be-forgotten delisting request, a hosting-provider complaint, or an approach to the poster. This tier needs precision rather than a practising certificate. It is where a removal desk earns its fee.
Tier three — you need a court, or you need money. You want damages, an injunction, an order to identify an anonymous poster, or a formal determination that the statement is false. That requires a solicitor, and no amount of takedown work substitutes for one.
We should be exact about our own position, because the keyword you typed may have implied otherwise. Obscura is not a law firm. We do not provide legal advice and we do not litigate. Our founder read media and defamation law, which shapes how we triage a case, and where a matter belongs in tier three we coordinate with media-law counsel who bill you separately and directly — the Obscura takedown team and its method sets out how that works, and our terms of engagement set out what an engagement does and does not cover. Any firm telling you a legal letter guarantees removal is selling you something the law cannot deliver, which is the same arithmetic as what a TikTok ban service charges, and for what.
If you do instruct someone, arrive prepared and choose deliberately. Bring the archived captures with dates, the reach figures, the documented consequences, a timeline of what was published when, and a note of everything you have already filed and the responses you received. That pack is the difference between a first meeting that scopes your case and one you pay for while the file is assembled in front of you. On choosing: defamation is a specialist field, and a high-street firm that does conveyancing and family work is not the right instruction. Check that the firm actually runs defamation and privacy matters, ask directly what proportion of their caseload it represents, and ask for a costs estimate broken down by stage rather than a single headline number. Ask, too, what they think the realistic outcome is. A solicitor who will tell you your claim is weak at the first meeting is worth more than one who will not.
What does a solicitor cost in England and Wales? Published ranges put pre-action work at roughly £3,500 to £9,500, early settlement at £12,000 to £25,000, and a matter that runs to trial anywhere from £59,000 to £312,000 or beyond, with the High Court issue fee alone at £10,000 to £10,500 for claims over £200,000. Conditional fee agreements exist for strong cases and after-the-event insurance can cap some exposure, but neither removes the central risk: lose, and you may pay the other side's costs as well as your own.
Suing for defamation on social media: the clock, the cost, the outcome
Can you sue for defamation on social media? Yes, and the claim is ordinary defamation: the medium changes the evidence, not the tort. Whether you should sue for online defamation is a different question, and the honest answer for most people is that suing is the slowest, dearest and least certain of the routes available, which is why it belongs at the end of a sequence rather than the start of one. Reported online defamation cases also skew what people expect, because the ones that get written up are the ones that went to trial — a tiny and unrepresentative slice. The social media defamation cases we see resolve most cleanly never generate a judgment at all.
The clock is the first thing to check, because it is short and it is unforgiving. Under section 4A of the Limitation Act 1980 no defamation action may be brought after one year from the date the cause of action accrued. The single publication rule in section 8 of the Defamation Act 2013 adds a second trap: where the same publisher republishes substantially the same statement, limitation runs from the first publication, not the latest one — so a post from 2024 that is still live does not give you a fresh year. The exception is where the manner of the later publication is materially different, and the court retains a discretion to extend time. One year, from first publication. People lose good claims to that sentence every year, and once it has run no amount of merit brings them back.
Then the arithmetic. Add the cost ranges above to an adverse-costs risk of similar size, and a claim needs to be worth pursuing in more than an emotional sense. Set against that, the outcomes a court can deliver are ones no takedown route can: damages, an injunction restraining repetition, an order for a statement in open court, and a public determination that the allegation was false. If what you need is vindication on the record, litigation is the only thing that produces it.
Most claims never see a courtroom, and the shape of the process explains why. It opens with a letter of claim setting out the words complained of, the meaning attributed to them, the harm caused and the remedy sought. A defendant who recognises they are wrong can make an offer of amends under the Defamation Act 1996 — a correction, an apology, and agreed compensation — and where that offer is made and not withdrawn it substantially limits their exposure, which is exactly why it gets accepted. Negotiation follows. Only if that fails does a claim form get issued, and only a small fraction of issued claims reach trial. The practical consequence for someone weighing this up: the letter of claim is doing most of the work, and a well-drafted one from a specialist frequently resolves a matter that would have cost six figures had it run. That is also why sending your own threatening message first is so damaging: it is the letter you cannot take back.
Anonymity changes the sequence rather than blocking it. Where the poster is unidentified, a Norwich Pharmacal order can compel a third party holding the information to disclose it, which is the step that has to come before any claim — the unmasking work described in unmasking an anonymous Instagram impersonator. Encryption sets a harder ceiling: where the messages are end-to-end encrypted, no order and no fee unseals the content, only the metadata, as why no fee unseals an encrypted chat sets out. And in the United States a further hazard waits, because 40 states and the District of Columbia had anti-SLAPP statutes as of March 2026, with South Dakota's taking effect on 1 July 2026 and Missouri becoming the seventeenth state to adopt the Uniform Public Expression Protection Act in July 2026. A weak claim filed against protected speech can be struck out early with the defendant's costs awarded against you. Cases are won on evidence assembled before filing, not on the strength of the grievance, which is the argument running through the twitter account removal service briefing.
Online defamation UK: what changes when the claim is English
The UK is not a mild variant of the US position. It is close to its inverse, and knowing which side of the Atlantic your problem sits on changes what you should do first.
England and Wales gives you a codified serious-harm threshold, a one-year limitation period, a statutory notice procedure that pressures website operators, and no constitutional actual-malice hurdle. It also gives you a costs regime that can bankrupt an unsuccessful claimant, and since section 11 of the Defamation Act 2013 removed the presumption of jury trial, a judge rather than a jury decides. Section 9 restricts claims against defendants domiciled outside the UK and certain treaty states unless England and Wales is clearly the most appropriate place to bring the action, which is the provision most often overlooked when the poster turns out to live in Ohio.
The United States runs the opposite balance. Defamation is state law, limitation periods run around one to three years depending on the state, public figures must prove actual malice under New York Times v Sullivan, and section 230 keeps the platform out of reach entirely. The practical consequence for someone searching social media defamation UK versus the American equivalent: in Britain, the notice and the claim have teeth and the platform can be pressured; in America, removal work and search suppression carry the weight because the litigation route against the intermediary is closed.
British claimants have one further protection worth naming, and one gap. Sections 194 to 195 of the Economic Crime and Corporate Transparency Act 2023 brought in an early strike-out mechanism for economic-crime-related strategic lawsuits against public participation, in force from 18 June 2025 and applied in Kamal v Tax Policy Associates [2026] EWHC 551 (KB). A general anti-SLAPP statute covering all subject matter has not passed. Note also that Ireland's Defamation (Amendment) Act 2026, which commenced on 1 March 2026 and abolished juries in defamation actions, is Irish law only — it does not apply in England and Wales, and several English-language guides have already blurred the two.
The ordinary cross-border case has a sequence that works, and it is worth setting out because guides tend to present the two systems as alternatives rather than as stages. Start with the platform, wherever it is, because policy enforcement is jurisdiction-blind and costs nothing. Move next to whichever national mechanism actually binds someone reachable — a section 5 notice if a UK operator hosts it, a Digital Services Act notice if the service operates in the EU, a hosting complaint if neither applies. Deal with search separately and later, since delisting the result is a different filing from removing the page. Only then consider a claim, and consider it in the jurisdiction where the defendant has assets, because a judgment you cannot enforce is an expensive piece of paper. The mistake we see most often is people starting at the last step, instructing lawyers in their own country against a poster in another, and discovering the enforcement problem after the costs have been incurred.
If you are choosing counsel, the vocabulary differs as well. UK practitioners are defamation solicitors or libel solicitors; searches for defamation lawyers uk return the same firms under an Americanised label. A US attorney cannot conduct an English claim, and an English solicitor cannot appear in a US state court. Where the poster, the platform and the claimant sit in different countries — the ordinary case, not the exotic one — the first question is not who to instruct but which jurisdiction gives the fastest lawful route to the outcome you actually need. The commercial pressure created by the Online Safety Act regime is a live factor in that calculation, as the Facebook ban service price bands illustrate on the buyer-risk side, while ban-for-hire and buyer liability shows who carries the exposure when someone pays for an outcome the platform alone controls.
What cannot be removed, and what happens instead
This is the section every honest removal desk owes you, and the one most reputation pages leave out because it costs them sales.
Truth cannot be removed. If the allegation is substantially true, section 2 is a complete defence and no notice, letter or claim will move it — and pursuing it anyway hands the other side a story. Honest opinion cannot be removed. A review saying the food was cold and the service rude is opinion with an indicated basis, however commercially damaging, and section 3 protects it. Genuine journalism on a matter of public interest cannot be removed, and section 4 protects a publisher who reasonably believed publication served that interest. Public records, court listings, insolvency registers, regulatory decisions and companies filings are not defamatory for being unwelcome, and no takedown route reaches them.
Fake reviews sit in an awkward middle and deserve their own note, because they are the commonest business case we see. A review describing a real experience, however unfair, is opinion and stays. A review from someone who was never a customer is a different animal: it is a fabricated statement of fact, it breaches every major review platform's own content policy, and the platform route is genuinely effective — provided you give the reviewer-verification team what it needs, which is evidence that no transaction matches, not an assertion that the review is unfair. Coordinated review attacks leave patterns that platforms detect readily: clustered timing, new accounts, similar phrasing, reviewers with no other history. Present the pattern rather than the individual reviews. In the United States the Federal Trade Commission's rule on fake and deceptive reviews adds a regulatory lever that did not exist a few years ago, and it applies to the people buying the reviews as much as to the people writing them.
There is a second reason to leave those alone beyond the legal one. Filing an aggressive notice against protected material reliably produces the Streisand effect, and in the United States it can now produce an anti-SLAPP strike-out with costs. We have watched a two-hundred-view blog post become a national story because someone sent a threatening letter to a journalist. The letter was the story.
What we do instead, when removal is genuinely unavailable, is say so and then discuss suppression: building accurate, authoritative, well-structured material about you that ranks above the damaging result, so that the first page of your name reflects reality. That is a slower project measured in months, it never deletes anything, and it comes with no guarantee. Where your problem is broader than defamation, the route-by-route breakdown of what each lawful instrument reaches, what it costs and how long it takes is set out in content removal service routes, costs and real limits. Anyone promising guaranteed removal of a true news article is either misinformed or lying, and the same test applies to the shopfronts examined in what an Instagram ban service is actually selling. The report-volume version of the same false promise is dismantled in X has no bulk-report API.
We also draw hard lines around method. We do not forge DMCA notices or send bluffed legal threats, because a false notice is unlawful, gets the material reinstated, and exposes the client. We never ask for your password or your account access. We do not require upfront crypto-only payment. And for non-consensual intimate images we triage urgently and at no charge, without ever asking you to send us the material — StopNCII for adults and NCMEC's Take It Down service for anyone under 18 generate a digital fingerprint on your own device and share only that hash with participating platforms, so the image itself never leaves your possession and never reaches us.
A last word on expectations, since this is where trust is usually lost. Nobody can promise a removal, because the decision belongs to a platform, a court or a search engine, and none of them answers to a vendor. What an honest desk can promise is a route, a realistic timeline, a clear statement of the odds, and the truth when the odds are poor. If a proposal you have been sent contains the words guaranteed removal, or quotes a price before anyone has looked at what the content actually is, you are being sold a lottery ticket with a service agreement stapled to it.
What money and volume cannot buy
A recurring pattern deserves its own warning, because people arrive having already spent on it. Search any platform name plus mass report bot and you will find panels, GitHub repositories, APK downloads and Telegram sellers offering to bury an account under complaints for a fee. None of them does what it claims, and several of them are the actual threat.
The reason is structural rather than technical. No major platform publishes a bulk-reporting endpoint, so there is nothing for a legitimate tool to automate. What the sellers actually operate are pools of stolen or rented accounts, and what they most often want from a buyer is the buyer's own credentials. The mechanics are documented across the network: four sources, one dead end traces the supply chain, the Instagram spam report bot myth explains why floods get the reporter actioned rather than the target, three products, one missing API covers the Facebook variant, and buyers hand over their own account covers the WhatsApp one. Even where a real quota exists it terminates quickly, as 200 reports a day, then nothing shows.
The paid-takedown market has the same shape and the same problem. Sellers quote a price for an outcome only the platform, a court or Google can actually deliver: an account banned, a post removed, a result de-indexed. What the buyer receives is a filing, which is the one part of the process that was always free. Some of these sellers are competent people filing correct reports and charging for the expertise, which is legitimate work honestly described. Others are charging four figures to press the same button you can press, and a few are charging for nothing at all. The test is simple enough to apply from an initial email: ask what specific route they intend to use, what the realistic timeline is, and what happens if it fails. A real desk answers all three. A shopfront answers none of them and repeats the price.
For a defamation case specifically, these tools are worse than useless. They generate a pattern of coordinated inauthentic reporting that platforms detect and penalise, they contaminate the evidence trail you will need if the matter ever reaches a solicitor, and they can convert you from complainant into respondent. A defamation claim is won by a small number of correct, well-evidenced filings. It has never once been won by volume.
The first 48 hours: the evidence pack that decides everything
How to deal with online defamation, in the order that actually matters: preserve first, report second, escalate third. Almost everyone does this backwards, argues in the replies, gets blocked, and loses the record.
Preserve before you do anything else. Capture full-page screenshots showing the URL, the account handle, the timestamp and the visible engagement counts. Save the direct link to the post itself, not to the profile. Submit the URL to the Wayback Machine so an independent third-party archive holds a dated copy — this matters enormously later, because your own screenshot is evidence you made, while an archive capture is evidence someone else made. Record reach: views, shares, comments, and who in your world saw it. Where you can, document consequences with dates — the cancelled meeting, the withdrawn offer, the client who asked about it — because that documentation is what converts an upsetting post into the serious harm the statute requires. Do all of this before you report anything, because material frequently disappears the moment a report lands, and deleted content is still actionable but very much harder to prove. Ephemeral platforms compress this window to hours rather than days, which is why the evidence is already deleting itself is the first thing we say about Snapchat cases. The same discipline underpins every filing we make, as what a removal desk does with your first forty-eight hours describes in detail.
The pack itself is short, and it is the same every time. One: a full-page capture of each URL showing the address bar, the handle, the date and the engagement figures. Two: the direct permalink to each item, recorded as text so it survives a screenshot being cropped. Three: an independent archive capture of each URL, dated. Four: the account details of the poster — handle, display name, profile URL, any identifying detail in the bio, and whether the account is new. Five: a reach note, recording who saw it and how you know. Six: a consequences log, one line per event with a date. Seven: a copy of every report you filed, with reference numbers and the replies received. Keep it in one folder, named by date. If the matter ever reaches counsel, this pack is what they will ask for in the first ten minutes, and having it ready shortens everything that follows.
Do not reply to the post. Do not repost it to complain about it, because you republish it to your own audience and hand the other side an argument about who caused the harm. Do not threaten the poster in writing, since that message will be screenshotted and it can turn a clean claim into a messy one.
Then triage honestly, in this order. Does it breach a platform rule? Report it precisely, once. Is it on a UK website with an unidentifiable author? Serve a compliant section 5 notice and start the 48-hour clock. Is there a court order or an eligible delisting case? Take it to Google. Is it true, opinion or public-interest journalism? Then removal is off the table and the conversation turns to suppression. Is it within a year of first publication and serious enough to justify the cost? That is the point at which a solicitor earns their fee.
A practitioner note, since these pages are usually written by people who have never filed one. The cases that resolve fastest in our experience are not the most outrageous — they are the ones where the client arrived with a complete, dated evidence pack and had not yet contacted the poster. The cases that stall are the ones where a well-meant argument in the comments has muddied who published what, or where six months passed before anyone thought about the one-year limitation period. Speed of preservation beats severity of harm almost every time.
Where to get online defamation help depends entirely on which of those five answers you reached. Tier one needs nothing but a careful filing. Tier two needs someone who has drafted a section 5 notice before. Tier three needs a solicitor, and the earlier you know that, the more of the one-year limitation period you still have to work with. What nobody needs is to spend three weeks arguing in a comment thread first, which is where a great many otherwise strong cases quietly die. Suing for online defamation is the last door in the corridor, not the first.
We handle case material under our privacy policy, and you can see the rest of the removal work across the Obscura takedown briefings or start from the Obscura content-removal desk. If you want a route mapped rather than a promise sold, ask for a confidential case review: we tell you which of the five lanes above your case is in, what it realistically costs, and when the answer is that nothing lawful will move it.