Managing comments on confession posts

Short answer. Tell readers not to name, tag or guess who a post is about, delete any comment that does, and add the name to the keyword filter so that repeats are hidden automatically. Look at each new post's comments within the first 15 to 30 minutes, and close the thread when attempts at identification keep coming.

Admins spend most of their care on the post. They take out the name, soften the detail, think about whether anyone could be identified — and then publish beneath it an open box in which any reader can type "this is obviously about Sarah from 2B" and tag her.

A confession page moderates submissions before they are published and comments only afterwards, if it moderates them at all. That gap is where much of the real harm occurs. This guide is about closing it.

Why the comments are your responsibility

It is tempting to regard comments as belonging to the people who write them. As a matter of law and of platform policy, that is not a safe assumption.

The Evidence Act 1950, in section 114A, presumes that the administrator of a page published what appears on it. In 2021, Malaysia's Federal Court held the operator of a news portal liable for contempt over comments posted by its readers, reasoning that it had provided the platform and had the means to control what appeared there. Platforms, for their part, take action against pages on the strength of what is in their comment sections as well as in their posts.

Quite apart from liability, there is the plain point that you built the room in which this is happening and you hold the keys to it.

What goes wrong in comments

Identification. Guessing or stating who a post is about. Tagging the person. Posting their handle "so they see this". This is the commonest problem and the one that does most damage, because it turns an anonymous post into a public accusation with a name attached.

Pile-ons. Dozens of people directing contempt at one person — the subject of the post, another commenter, or the supposed author.

Guessing the author. "This sounds like something X would write." Just as corrosive, and often wrong.

Escalation. Two commenters arguing until it becomes personal, with threats, slurs or private details appearing.

Spam and scams. Fraudulent links, false giveaways, and comments from accounts imitating your page ("you've won — message us on Telegram").

Harmful replies to vulnerable posts. Mockery or encouragement beneath a post about despair. See When a submission mentions self-harm.

Set the rule and state it

Add one line to your page's rules and repeat it wherever you can:

In the comments: don't name, tag or guess who a post is about or who wrote it. Comments that do are removed, and repeat offenders are banned.

Then enforce it consistently. A rule that readers see being enforced is one they begin to follow, and after a while to enforce on one another.

The tools you have on Facebook

The labels and locations of settings change. Look in your page's settings and in Meta Business Suite for the current versions.

Hide a comment. The comment remains visible to the person who wrote it and to their friends, and is hidden from everyone else. The author is not notified. This is the quietest tool and usually the right first step, since it removes the harm without starting an argument.

Delete a comment. Removes it altogether. Use this for anything that names a person, since a hidden comment is still seen by some people.

Keyword blocking. You can supply a list of words and phrases, and comments containing them are hidden automatically. This is the most useful proactive tool for the present purpose: when a name begins to appear beneath a post, add it — and its obvious variants — to the list, and further comments containing it are hidden before anyone sees them.

Profanity filter. A setting that hides comments containing commonly reported offensive words.

Automated comment moderation. Meta provides rule-based moderation for pages that can hide comments automatically according to criteria you set — comments containing links, for example, or comments from very new accounts or from accounts that do not follow the page. It is particularly effective against spam and scams.

Limit who can comment on a post. For an individual post, you can restrict commenting to a narrower group.

Turn off commenting on a post. Where the option is available for your page and post type, this closes the thread. Where it is not, restricting who can comment as tightly as possible, or hiding the post, achieves much the same.

Ban or restrict a person from the page. They can no longer comment on it or interact with it.

The tools you have on Instagram

Hidden Words. Filters comments and message requests containing offensive terms, with a custom list for words of your own. The same approach applies as on Facebook: add a name as soon as it starts to circulate.

Delete and report individual comments.

Restrict. A restricted person's comments are visible only to that person unless you approve them. They are not told. This is useful for someone who is a persistent problem without warranting a ban.

Block. Removes the person's ability to interact with your account.

Limits. Temporarily hides comments and messages from accounts that do not follow you or that followed only recently. It was designed for moments when an account is suddenly flooded, which is exactly what a brigade looks like.

Pin comments. You can pin a small number of comments to the top of a post. Pin the ones that set the tone you want.

Turn off commenting on an individual post, from the post's menu.

Watch the first hour

Most comments arrive soon after a post is published, and the earliest comments set the tone for the rest. If the first three are guesses at a name, the next thirty will be as well. If the first three are sympathetic, the thread tends to continue in that way.

So:

• an admin should look at each new post's comments within the first 15 to 30 minutes;
• sensitive posts should be published only when someone is able to watch them — see How often should a confession page post?;
• remove the first identifying comment at once, before others build on it;
• consider adding a first comment from the page itself — "Reminder: no names or guesses in the comments" — on posts that are likely to attract them.

When to close comments

Close the thread, or hide the post, when:

• attempts at identification keep coming despite removals;
• a pile-on has started and is growing faster than you can moderate it;
• the subject of the post has asked for it to be removed — remove the whole post, and not only the comments;
• no admin will be available for a long period and the post is a risky one;
• the discussion has become a quarrel between commenters that has nothing to do with the post.

Closing comments is not a failure. Some posts are worth publishing and are not worth discussing.

Hide, delete or ban?

You do not owe anyone a warning before removing a comment that identifies a person. Remove it first. If you wish to explain, do so in a private message and not in the thread.

Replying as the page

When the page itself comments, it speaks with the authority of the admins. Use that sparingly, and well.

• Do restate the rule when necessary, briefly and without sarcasm.
• Do answer genuine questions about how the page works.
• Do add sources of help beneath posts about distress.
• Do not take sides in arguments between commenters.
• Do not comment on the truth of a submission, or hint at who wrote it or whom it concerns.
• Do not ridicule a commenter from the page's account, however much they may deserve it. Your audience is watching how you behave when you hold the power.

Impersonators and scams

Popular pages attract accounts that copy their name and picture and then reply to commenters — "Congratulations, you've been selected, message us" — in order to defraud them. Your followers may not notice the difference.

• Report impersonating accounts to the platform when you see them.
• Tell your audience from time to time that the page never asks for payment, passwords or codes through comments or direct messages.
• Use automated moderation rules to hide comments containing links, and comments from very new accounts.

Keep evidence when it is serious

Before you delete a comment that contains a threat, discloses someone's private information, or forms part of a sustained campaign against a person, take a screenshot showing the comment, the account and the date. The person targeted may need it for a report to the platform, to MCMC or to the police. Then delete it. See What to do if an anonymous page posts about you for what they can do with it.

A daily routine

• Check new posts' comments within half an hour of publication
• Remove identifying comments on sight, and add the name to the keyword filter
• Go through the day's posts once more in the evening
• Review the list of hidden comments weekly, since automated filters hide innocent comments too
• Each term, go back over the banned list and the keyword list and remove what is no longer needed

The comment section is where your community meets itself. If you leave it alone, it tends towards the level of its worst participant. With a little regular attention, the same space is where a person who wrote something hard and sent it anonymously finds out that others have felt the same.

Common questions

Is a page admin responsible for the comments on its posts?

It is not safe to assume otherwise. Section 114A of the Evidence Act 1950 presumes that the administrator of a page published what appears on it, and in 2021 Malaysia's Federal Court held the operator of a news portal liable for contempt over comments posted by its readers. Platforms also act against pages on the strength of what is in their comment sections.

How do I stop people naming someone in the comments?

State the rule — no naming, tagging or guessing — and enforce it every time. Delete the first identifying comment at once, then add the name and its obvious variants to Facebook's keyword blocking or Instagram's Hidden Words so that later comments containing it are hidden before anyone sees them.

Should I hide or delete a comment?

Hide a comment that is rude but harmless: it stays visible to its author and their friends and is hidden from everyone else. Delete anything that names, tags or identifies a person, because a hidden comment is still seen by some people.

When should comments be turned off on a post?

When attempts at identification keep coming despite removals, when a pile-on is growing faster than you can moderate it, when no admin will be available for a long period and the post is a risky one, or when the thread has become a quarrel that has nothing to do with the post. If the subject of the post asks for removal, remove the whole post and not only the comments.

How soon should an admin check the comments on a new post?

Within the first 15 to 30 minutes. Most comments arrive soon after publication and the earliest ones set the tone: if the first three are guesses at a name, the next thirty will be as well.