Moderation tools and community defences

A community is one of a creator's most valuable assets, and also one of the main places harassment happens. Moderation is how you keep the space safe enough for your audience and for yourself. This lesson covers community guidelines, the built-in tools platforms provide, keyword filters and how to measure them, human moderators and how to look after them, and the wider community defences that make abuse less likely to take hold in the first place.

Start with community guidelines

Moderation is easier and fairer when everyone knows the rules. Community guidelines are a short public statement of what is welcome and what is not. They do not replace platform rules, which apply everywhere, but they let you set a higher standard for your own space. Good guidelines are short, specific and written in your own voice. They usually cover respect for others, no hate or harassment, no sharing of personal information, no spam or scams, no self-promotion without permission and what happens when rules are broken.

Publish the guidelines where people will see them: in a pinned comment, a channel description, a community tab or the rules section of a chat server. Refer to them when you remove content or ban someone. Consistent enforcement builds trust: audiences accept moderation more readily when they can see it follows clear rules rather than personal whims.

Platform tools

Major platforms provide moderation tools, and they change often, so check each platform's help centre for current options. Common tools include:

• Blocking, which stops an account from seeing or interacting with you.
• Muting or restricting, which hides an account's comments from others without telling them, reducing the chance of escalation.
• Keyword filters, which hide or hold comments containing words or phrases you choose.
• Automatic filters for offensive language, provided by the platform.
• Comment controls, such as limiting comments to followers, followers of a certain age or approved accounts, or turning comments off for a post.
• Held-for-review queues, where comments wait until a moderator approves them.
• Slow mode and follower-only modes in live chat, which limit how often people can post.
• Moderator roles, which let trusted people remove comments and ban users on your behalf.
• Limits on direct messages, such as allowing messages only from people you follow.
• Hidden word lists for direct message requests.

During a pile-on, temporary restrictions can make a big difference. Limiting comments to long-standing followers for a few days, or turning off comments on a particular post, stops a flood of new accounts from reaching you while the attention passes.

Keyword filters

Keyword filters are a creator's first line of defence. A good filter list includes slurs and common insults, threats, words linked to doxxing such as "address" or the name of your town, spam phrases and common misspellings or substitutions harassers use to get around filters. Lists need regular updates, because harassers adapt. Filters are not perfect: they miss abuse that uses ordinary words, and they sometimes catch innocent comments. Treat them as a way to reduce volume, not a complete solution.

Two measures help you judge a filter. Recall, sometimes called the catch rate, is the share of abusive comments the filter catches. Precision is the share of caught comments that are actually abusive. A filter with high recall but low precision catches most abuse but also hides many innocent comments. A filter with high precision but low recall rarely hides innocent comments but lets much abuse through. In practice, creators usually aim for high recall with a review queue, so that a human can rescue innocent comments.

Human moderators

As a channel grows, one person cannot read everything. Human moderators, whether trusted fans, friends or paid staff, can remove abuse, answer questions, enforce guidelines and spot problems early. Moderators should be chosen carefully: they need good judgement, discretion and loyalty to the community's values, not just enthusiasm. Give them clear guidelines, examples of what to remove and what to leave, a way to escalate serious cases such as threats or doxxing to you or your manager, and limited permissions that match their role.

Volunteer moderators do valuable work, and creators should treat them fairly. Thank them, include them in decisions about community rules, and consider paying regular moderators or offering other compensation. Remove permissions promptly when someone steps down.

Looking after moderators

Moderation exposes people to large volumes of abuse, sometimes including threats, hate and disturbing images. Research on professional content moderators has found high levels of stress and symptoms of trauma among people who review harmful content for long periods. Volunteer and part-time moderators for creators face a smaller version of the same risk.

Protect moderators by rotating shifts so no one moderates for too long, encouraging breaks, letting them step back from particularly upsetting content, using filters to reduce what they see, blurring images by default where tools allow and checking in after difficult incidents. Make it clear that moderators can say no to tasks that are too much. If someone is struggling, encourage them to talk to someone they trust or a professional.

Community defences

The best moderation is supported by the community itself. Communities with clear norms, active and respected moderators and a culture of kindness tend to self-correct: members discourage bad behaviour, report problems and welcome newcomers. Creators can encourage this by modelling the behaviour they want, highlighting positive comments, thanking members who help others and explaining why rules exist.

Some creators also use structure to slow down abuse. Membership tiers, verification steps or waiting periods before new members can post in private communities make it harder for brigades to arrive all at once. Separate spaces for different purposes, such as a public comment section and a members-only chat, let creators keep their most personal interactions in safer places.

Reporting to platforms

Platform reporting tools let you flag content that breaks platform rules, such as hate, harassment, threats, impersonation, private information and non-consensual intimate images. Report through the specific category that fits, include links and context, and keep a record of reports and responses in your evidence log. Many platforms have dedicated processes for creators, such as partner support teams or creator help lines, which can be faster than standard reporting. If you work with a multichannel network, agency or manager, they may have direct contacts at platforms.

Platforms do not always act quickly or consistently. If a report is rejected, check whether an appeal is possible, try a more specific category, and escalate through creator support channels. For threats, stalking, image abuse and extortion, do not rely on platform reporting alone; Lesson 9 covers legal options.

Moderating live streams

Live streams need faster moderation than recorded videos, because messages appear in real time and harmful content can be seen by many viewers before anyone acts. Useful practices include having at least one moderator present for every stream, turning on slow mode or follower-only chat during busy or contentious streams, using automatic filters for slurs and links, and giving moderators the power to time out users quickly. Agree in advance on what moderators should do if a raid of hostile accounts arrives, such as switching chat to subscriber-only mode for ten minutes.

Creators should also decide how they will react on camera. Reading abuse aloud can amplify it and encourage more. A calm, brief acknowledgement, such as "Mods are handling it; let's keep going", keeps the focus on the content. If a stream becomes unsafe or overwhelming, it is fine to end it early.

Transparency and appeals

Communities trust moderation more when it is transparent. Explain your guidelines, let people know when and why content was removed where practical and offer a simple way to appeal a ban, such as an email address or form. Most appeals will be from people who genuinely misunderstood a rule or were caught by a filter by mistake. Handling them fairly keeps good members in the community and shows that moderation is about safety, not silencing disagreement. Keep notes on recurring issues so that you can clarify guidelines or adjust filters.

Moderation for small creators

Creators with small audiences often think moderation is only for big channels. In fact, a small community is easier to shape early. Setting guidelines, turning on basic filters and responding consistently from the start establishes norms before problems arise. A small creator can usually moderate alone using reply windows and filters, but it is worth identifying one trusted person who can step in if a video unexpectedly attracts a large, hostile audience. Viral attention can arrive without warning, and having a plan ready is much easier than building one in the middle of a pile-on.

As your community grows, revisit your moderation setup every few months: check filter recall and precision, ask moderators how their workload feels and update guidelines to reflect new kinds of problems.