The short answer: Publish a comment policy before the campaign starts, then hide only what breaks it: threats, slurs, doxxing, spam and coordinated fake accounts. Leave criticism up. Staff the hours that actually get busy, mainly debate nights and election day. Keep a record of every action so you can answer censorship claims with evidence.
Why does a campaign need its own moderation plan?
Outside a campaign, a politician's page gets a steady trickle of comments. During one, three things change at once. Volume jumps, often by an order of magnitude on a single post. The comments arrive in bursts tied to events you can predict: a debate, an interview, a scandal, the close of polls. And the stakes of a mistake rise, because a hidden comment that looks like silencing a critic becomes a story in itself.
So the plan is simple: decide the rules in advance, apply them the same way to supporters and opponents, and be able to show what you did.
What should your published comment policy say?
Write it before the campaign, pin it, and link it from the page's About section. A short policy does more work than a long one. Cover:
- What gets removed or hidden. Threats and incitement, slurs and dehumanizing language about protected groups, doxxing (home addresses, phone numbers, workplaces, photos of private homes), sexual harassment, spam and scam links, and impersonation of the candidate or staff.
- What stays, explicitly. Criticism of the candidate, the party, the program and the campaign. Say this out loud. It is the sentence you will quote later.
- What happens to repeat offenders. Blocking an account after repeated violations, and that the block is about behavior, not opinion.
- Who to contact if someone thinks a comment was hidden by mistake, and that you will look again.
Keep the policy in every language your voters actually comment in. If your audience switches between two languages mid-sentence, a single-language policy reads as if it was written for someone else.
What should you hide, and what should you keep?
Hiding is reversible and keeps the evidence, which makes it the right default on Facebook and Instagram. On Facebook a hidden comment stays visible to its author and their friends; on Instagram it stays visible to the person who wrote it. Either way it rarely triggers the "you deleted me" reaction that a removal does. Deleting is for content you never want attached to your page at all.
| Comment type | Default action | Why |
|---|---|---|
| Threats of violence, incitement | Hide, screenshot, report to the platform, escalate to police if credible | Unlawful in most jurisdictions and a safety matter for staff and volunteers |
| Slurs, dehumanizing language about a group | Hide | Breaks platform rules and your policy; leaving it up invites more |
| Doxxing: addresses, phone numbers, private photos | Hide immediately, then report | Real-world risk to a named person; speed matters more than process |
| Spam, scam links, fake giveaways, crypto bots | Hide automatically | No judgment call needed, and volume is highest during traffic spikes |
| Coordinated fake accounts repeating the same text | Hide, block the accounts, report the network | Meta's Community Standards treat networks of inauthentic assets as a separate violation |
| Rude, sarcastic or hostile criticism | Keep | Hiding it is the single fastest way to turn one angry voter into a story |
| Policy questions, including hostile ones | Keep, reply where useful | Public answers reach everyone reading the thread, not just the asker |
| A false factual claim about your candidate | Keep and correct in a reply | See below |
The hard cases sit between rows. A comment can be offensive and still be political speech. If you find yourself arguing about whether something is "too much", that is usually a sign it belongs in the keep column. For a fuller treatment of where that line sits, see how public-facing pages can reduce toxic comments without silencing real debate.
How should you handle disinformation in your own comments?
Resist the urge to hide false claims. A hidden claim cannot be corrected in public, and the person who posted it can still see it, screenshot it, and say you suppressed them.
Three responses work better. First, reply once with the correction and a link to a primary source, then stop; repeating the false claim in order to deny it mostly spreads it further. Second, keep a short, pre-written set of replies for the five or six claims you already know are circulating about your candidate, so a staffer at 11pm is not drafting policy at speed. Third, report content that breaks the platform's own rules and leave it at that.
The exception is volume. Treat one person with a wrong belief as a conversation worth having. Treat forty accounts posting identical text within an hour as spam, because that is what it is. The same pattern logic that works on scam links and fake giveaway comments applies to copy-pasted political spam.
How do you staff debate nights and election day?
Comment traffic on a campaign page is not evenly spread. Plan around the peaks.
- Map the calendar. List every debate, major interview, rally and the final days before the vote. Those are your surge windows.
- Two people per surge, not one. One person watches the live post, one works the backlog on older posts, where the worst comments usually land because nobody is looking.
- Agree an escalation path. Who decides on a borderline comment at 10pm? Who calls the police about a threat? Write down one name and one phone number for each.
- Automate the obvious. Keyword filters and blocked accounts catch slurs and spam without waking anyone. Keep humans for judgment calls.
- Cover the night. Comments posted at 2am sit on the page until morning. Automated hiding of clear violations overnight, reviewed by a person at 8am, is the realistic version of a night shift for a campaign that cannot afford one.
- Treat election day as a safety shift. Expect threats and voter-suppression claims. Have the reporting links open before you need them.
If your electorate comments in more than one language, read how to moderate comments in multiple languages before relying on keyword lists, which break badly across inflection and diacritics.
Are you legally responsible for what people post on your page?
Possibly, and the bar is higher for politicians than for ordinary users. In Sanchez v. France (Grand Chamber, May 2023), the European Court of Human Rights found no violation of free expression in the criminal conviction of a French politician who failed to promptly delete clearly unlawful third-party comments from his public Facebook wall. In the United States, the Supreme Court's 2024 decision in Lindke v. Freed set out when an official's social media account counts as state action, which can make blocking critics a First Amendment problem. Both points, and what they mean day to day, are covered in our article on legal responsibility for comments on your Facebook page.
What changed for political ads?
Paid and organic are different regimes. Meta requires advertisers running ads about social issues, elections or politics in designated countries to complete an authorization process and carry a "paid for by" disclaimer; ads with that content and no disclaimer are disapproved in review, and these ads are kept in Meta's public Ad Library for seven years. In the European Union the picture is different again: the Transparency and Targeting of Political Advertising Regulation (EU 2024/900) became fully applicable on 10 October 2025, and Meta responded by stopping political, electoral and social issue ads across the EU from early October 2025. Organic posts were not affected. Practical consequence for a campaign in the EU: your comment sections on organic posts carry more of the load than before, because that is where the conversation still happens.
How do you answer accusations of censorship?
With records. The accusation usually arrives as a screenshot and a claim that you hide everyone who disagrees. You cannot disprove that from memory.
What works: a published policy that predates the complaint, a log showing what was hidden and why, and the ability to say how many comments on that post are still visible. If a comment was hidden by mistake, unhide it and say so. One corrected mistake costs far less than a defended one.
Keep the log even for manual actions. "A volunteer hid it, we do not know who or why" is the answer that turns a complaint into a story.
Where a tool fits
Platform filters are free and you should turn them on first: Facebook and Instagram both offer keyword blocking, and Instagram's Hidden Words covers a lot of the routine abuse. They are rule-based, though, so they do not tell criticism from a slur-free insult, and they do not explain their decisions.
Social Moderator AI is built for the part after that: you write the rules in plain language, the AI applies them to each new comment, and every action comes with a written reason and an audit log entry you can show. It covers Facebook, Instagram, TikTok and YouTube. Note that political Facebook Pages and Instagram accounts are supported on the Pro plan, not the entry plan; there is more detail on the page for political teams. Whichever route you take, the policy and the record are yours, not the tool's.
Frequently asked questions
Should we turn comments off during the campaign?
Usually not. Closed comments on a candidate's page read as avoidance, and they remove the main place where supporters answer critics for you. Limiting comments on a single sensitive post, for example about a death or an ongoing investigation, is a defensible exception.
Is it censorship to hide a comment?
Hiding a comment that breaks a published policy is moderation. Hiding a comment because it criticizes the candidate is the thing you were accused of. The difference only exists if the policy exists first and you apply it to your own supporters too.
Should we block accounts during a campaign?
Block for repeated behavior, not for one bad comment, and never for disagreement alone. In the United States, officials should be aware of Lindke v. Freed before blocking anyone from an account used in an official capacity. Everywhere else, assume the block will be screenshotted.
What do we do about comments on a rival's page?
Nothing organized. Organising supporters to flood an opponent's comments is a platform and reputational risk: if it runs through fake or duplicate accounts it falls under Meta's Inauthentic Behavior policy, and a coordinated pile-on aimed at a named person can breach the bullying and harassment rules. It is also the easiest kind of campaign story to prove.
How long should we keep the moderation records?
At least through the post-election period when complaints and press questions arrive, and longer if your national electoral rules set a retention period. Keep enough to reconstruct a decision: the comment, the action, the reason, the time and the person or system that acted.