Meta Ad Comments: Reply Fast, Hide Spam, Reuse the Post

Look below your own ad the way a stranger does. The first comment is a complaint about a refund that took three weeks. The second is a phone number from a man asking for fares. The third says "scam" and nothing else. Those three lines are now part of the ad, because everyone it reaches reads them before deciding whether to message you. Meta's help page on ad comments puts it plainly: for most advertisers, anyone who can see the ad can comment on it.

The ad creative article argued that the feed rewards posts that read like a person wrote them. The comment thread is where that gets tested in public, at the speed you reply. This piece is the mechanics: hide versus delete, the filters that work overnight, how one post keeps its proof across campaigns, and what to say to the three comments every service business gets.

Hidden or deleted: who sees what

Meta's help page for Pages says a hidden comment stays visible to the person who wrote it and their friends, and is hidden from everyone else; replies under it go with it. Meta adds that this can help prevent more unwanted comments, a polite way of saying the person often doesn't notice and doesn't come back angrier. Delete removes the comment from the post, and the commenter can see that it's gone.

So hide is the default for spam, abuse and the vague "fraud" with nothing behind it. Delete is for the comment that should not exist even for the commenter's friends: a phone number or document details posted in public, anything unlawful, and accounts pretending to be your business.

Who can do it matters in a small team. Meta's pages say hiding, sorting or deleting comments on Page posts needs Facebook access to the Page (you switch into the Page to do it); task access alone can't. In Meta Business Suite, answering comments needs community activity permission. Ads Manager also has a comment view for each ad where you can reply, like, hide and report. Whichever screen you use, the reply carries the Page's name, so settle who answers and in what tone before the ad goes live.

Filters that work while you sleep

Two Page-level settings cover ads as well as organic posts, and Meta says so: the profanity filter and the blocked-words list extend to ads from the Page. The list takes up to 1,000 keywords in any language, and Meta hides common variations automatically (misspellings, plurals, numbers in place of letters, the hashtag version), so one entry does the work of ten. Add the abuse you actually receive, in the spellings your market uses, Hinglish included. Emojis can be blocked too.

Moderation Assist, the newer tool that hides comments from week-old accounts, accounts with no profile picture, or comments carrying links and images, is useful for organic posts. Read the fine print: Meta's page says it does not hide comments on any ads you run. For ads, blocked words and the profanity filter are the automation you have.

Filters hide, they don't judge. "Is this a scam or genuine?" trips a list containing "scam", and that is the comment you most want to answer in public. Once a week, open the hidden comments on your running ad and unhide the genuine ones.

Why a new ad starts at zero, and how to keep the proof

Reactions and comments belong to a post, and every ad you build fresh in Ads Manager creates a new post. That is why a copy of last month's winner arrives in a new ad set with nothing under it while the original sits on a hundred reactions and forty answered questions. Meta gives you two ways to carry the thread. When you duplicate an ad, a tick box shows the existing reactions, comments and shares on the new ads. When you build a new ad, choose the option to use an existing post and pick the same one (Meta suggests selecting it by the post ID shown in the Page posts list rather than copying an ID out of a URL). Do that across ad sets and campaigns and one post carries its proof everywhere it runs.

The choice has costs, and Meta's help pages list most of them:

• Call-to-action buttons can't be edited on ads that use an existing post, and changing details on the ad, such as the phone number, changes the post itself.
• A boosted post can't have its text, image or video edited once it has been reviewed and published as an ad.
• Delete the post from the Page and you can no longer use it in new ads or edit the ads that include it.
• If you customised the creative per placement, the engagement can be carried over only once when duplicating.
• Our addition: one post means one thread. A troll's comment appears under every ad that uses it, and so does the question you left unanswered last Tuesday.

Our advice: test new angles as fresh ads, where you can still edit. Once an ad has earned real questions and your replies, stop duplicating and run that post ID wherever it fits. If the offer changes, make a new post rather than editing the old one around its comments.

Three comments you will get, and what to say

The price or availability question. Answer in public with the real number or range, then one next step. "Yes, the Saturday 11am slot is open. The first consultation is Rs 500, adjusted against treatment. Send your name and preferred time on WhatsApp and we'll confirm within the hour." One reply, fifty people answered. If the honest answer is "it depends", say what it depends on ("fares change daily; send your travel dates by message and we'll quote two options") rather than "DM for price", which reads as hiding.

The complaint. Reply within the hour if you can, name the problem, take the detail private, then fix it. "Sorry, three weeks for a refund is too long. Please message us the booking reference and I'll call you today with the status." No excuses in public, no asking them to remove the comment. When it's fixed, a short follow-up ("Resolved by phone, thanks for your patience") finishes the story for everyone watching. Leave fair criticism up: a thread with one handled complaint reads as a real business, and a thread scrubbed clean reads as a business that scrubs.

The troll. Abuse, slurs and the account that posts "fraud" under every ad: hide, don't answer. Ban repeat offenders and report comments that break Meta's Community Standards. The exception is the one-word accusation a real customer might read as a warning. Give it one calm line ("If you've had a problem with us, message us the details and we'll sort it out") and leave it. Never argue, never post their profile, never reply three times.

Comments and the lead path: one next step

The public reply explains; the private channel sells. Point each reply to one place, WhatsApp or a call, never "WhatsApp, call, email or the website". Business Suite's Inbox lets you message a commenter straight from the comment (on Instagram, Meta says only within seven days of it), so use that for haggling, personal details and anything needing a document. Price negotiation never belongs in the thread, because the discount you give one person in public becomes the price for everyone who reads it.

Don't bait. Asking people to react, to comment with a specific word or number, or to pass the post to friends is what Meta's publisher guidelines call engagement bait, and Meta says such posts get less distribution. A plain question in the copy ("Travelling with parents over sixty? Ask us about wheelchair assistance") invites a real conversation. Real questions are also the kind of Page engagement the retargeting article's custom audiences are built from; a thread of "yes" is not.

What Meta's rules say about bought and seeded comments

Buying comments, asking staff and cousins to post as customers, or seeding praise from a second account are cheaper than answering real people, and all sit on the wrong side of Meta's written rules. The Spam section of the Community Standards says Meta does not allow "selling, buying or exchanging for engagement, such as likes, shares, views, follows, clicks". The Inauthentic Behaviour section covers fake accounts and Pages used to deceive people about who an audience really is. Meta's publisher guidelines add that accounts engaging in fake engagement see reduced distribution. None of those pages promises a specific penalty for a small Page, and you shouldn't need one: a bought "best service" under an unanswered complaint convinces nobody and hands the next critic a target.

A routine you can keep

Pick two fixed times a day, say 10am and 6pm, and keep them for the life of the campaign. At each one: open the comments on every running ad from Inbox or Ads Manager, answer every question with the answer and one next step, move price, personal details and disputes to messages, hide spam and abuse, and leave fair criticism up with a calm reply. Once a week, review the hidden comments and unhide the genuine ones, add new spellings of abuse to the blocked-words list, and check that the post you are reusing still says something you can honour.

Our recommendation: treat the thread as part of the ad and budget time for it the way you budget money for the media. Turn on the profanity filter and a blocked-words list today, reply inside working hours so no question sits overnight, and once an ad has earned a real thread, stop duplicating it and run that post ID instead. Nothing in the ad copy is read as closely as your reply to the first complaint.

FAQ

Should I switch comments off on my ads?

On Facebook you moderate rather than switch off; Meta's page on ad comments says anyone who sees the ad can comment. Instagram lets you turn comments off on a post, but an ad with no thread looks like one nobody trusted enough to ask a question. Keep them on, filter the abuse, answer the rest.

A customer posted their phone number under the ad. Hide or delete?

Delete, then message them. A hidden comment stays visible to their friends, and a number in public under a travel or clinic ad invites accounts pretending to be you to message them first. Reply in the thread that you've messaged them, so the next person uses WhatsApp instead.

Does hiding or deleting comments hurt delivery?

Meta's help pages treat hiding, blocking words and the profanity filter as ordinary moderation and attach no penalty to them. What Meta says it reduces distribution for is engagement bait and fake engagement. Moderate as much as you need to; what you shouldn't do is manufacture the thread.