Spam from accounts that look real
Spam from an old account with a photo and friends is caught by what the comment says, not by what the account looks like.
On this page
Before you start
- The rules read the comment, not the account.
- Moderation Assist can stay on beside them.
- Mute an author who keeps coming back.
Switch on the rules that read the comment.
Left sidebarPagesyour pageModerationthe rules with the AI mark
You will see Each rule shows examples of what it catches, whoever wrote it.
Add the exact patterns that spam always carries.
URLsEmails and phone numbersKeywords
You will see Every link, number or listed word is caught, whoever posted it.
Catch fake support by the name and the comment together.
The Impersonators ruleits switchDelete
You will see Fake support comments go, and the Bin keeps the record.
Mute the ones that keep coming back.
Inboxthe mute icon under the commentMute
You will see Their next comments arrive hidden, with a Muted author chip.
If something is off
- A real customer was hidden
- Unhide, whitelist the words that must survive, and write the Negativity note.
- The same person is back under the next ad from a new account
- The rules catch the pattern, not the account; mute the ones you recognise.
- Moderation Assist hid something first
- It shows as hidden outside the app; keep both on or turn one off.
- The spam is on an old post, not a new comment
- Rules are for new comments; Post cleanup handles an old post.
Common questions
Does CommentWatchDog look at the commenter's account age, profile picture or friend count?
No. Every rule judges the comment: the AI rules read what it means, in any language, and the Impersonators rule reads the author's name and the comment together. The text rules match links, contact details and the words you add. An old account with a photo and friends is caught on what it posts, which is what Facebook's account-based criteria let through.
Can I keep Facebook Moderation Assist on at the same time?
Yes. Moderation Assist runs inside Facebook and CommentWatchDog runs on the comments that reach your Page through Meta's API. A comment Moderation Assist hides first arrives in the inbox marked as hidden outside the app; the rest is judged by your rules.
What happens when the spammer changes the wording?
The AI rules read meaning, so a reworded scam matches the same topic. Only the Keywords rule needs the exact words. When a new pattern appears, describe it once under Custom topics and test it in the Playground.
Which plans include this?
The preset AI moderation, Profanity and Negativity, and the text rules run on every plan, Mini and Lite included. Custom topics, the Negativity note and the Impersonators check start on Starter.
For agents: plain text of this guide · all guides
For agents
Guide: Spam from accounts that look real
This page: https://commentwatchdog.com/help/moderation-rules/spam-from-accounts-that-look-real
Topic: Moderation rules
Time: about 3 minutes
Who can do it: Owner, Administrator or Manager
Author: https://anantgupta.xyz (Anant Gupta, founder of CommentWatchDog)
Start in the app: https://app.commentwatchdog.com/en/inbox
Result when done: Spam from an old account with a photo and friends is caught by what the comment says, not by what the account looks like.
Context: Facebook's Moderation Assist judges the person commenting: a new profile, no profile picture, no friends or followers, a record of hidden comments. Spam from an old account with a photo, friends and years of posts passes every one of those tests, and that is where most of the spam under ads comes from now, because the accounts are bought or taken over. CommentWatchDog judges the comment instead: what it says, in any language, and whether the name and the words together pose as your support. An account that looks real and posts a scam is caught on the scam.
Before you start
1. The rules read the comment text and the author name. The age of the account, its photo and its friend count play no part.
2. Moderation Assist can stay on; a comment it hides first arrives in the inbox marked as hidden outside the app.
3. For the ones that keep coming back, Mute hides everything an author writes from then on.
Steps
Step 1 of 4: Switch on the rules that read the comment.
Where: Left sidebar, Pages, the page, then the Moderation tab. Profanity, Negativity, Custom topics and Impersonators carry the AI mark and judge what the comment means. Under Custom topics, describe the spam you see in plain words, for example people offering to recover a hacked account, or anyone selling followers.
Link: https://app.commentwatchdog.com/en/inbox (Open the app)
Screenshot: https://commentwatchdog.com/media/shots/moderation.en.webp (The rules that judge meaning, with the AI mark.)
Screenshot: https://commentwatchdog.com/media/shots/moderation-phone.en.webp (The same on a phone.)
You should see: Each rule shows examples of what it catches, whoever wrote it.
Step 2 of 4: Add the exact patterns that spam always carries.
Where: Further down the same tab. URLs catches every link, with your own domains allowed. Emails and phone numbers catches contact details and WhatsApp numbers. Keywords catches the words you add. These match exactly what is in the text, so an account with a thousand friends and a link is caught on the link.
Screenshot: https://commentwatchdog.com/media/shots/moderation-text.en.webp (The rules that match what is in the text, from URLs down.)
Screenshot: https://commentwatchdog.com/media/shots/moderation-text-phone.en.webp (The same on a phone.)
You should see: Every comment with a link, a number or one of your words is hidden or deleted, whoever posted it.
Step 3 of 4: Catch fake support by the name and the comment together.
Where: The Impersonators rule, on the same tab. It reads the profile name and the comment as one thing and catches accounts posing as your page or its support, including names that only resemble yours. Delete is the usual action for it.
Screenshot: https://commentwatchdog.com/media/shots/moderation.en.webp (The Impersonators switch.)
Screenshot: https://commentwatchdog.com/media/shots/moderation-imp-phone.en.webp (The same on a phone.)
You should see: Fake support comments are removed and kept in the Bin with the Impersonator label.
Step 4 of 4: Mute the ones that keep coming back.
Where: In the inbox, the block or mute icon under a comment. Mute hides everything that author writes from then on, quietly; on Facebook, Block stops them reaching the Page at all. Select several comments and use Bulk actions to do it for many authors at once.
Link: https://app.commentwatchdog.com/en/inbox (Open the inbox)
Screenshot: https://commentwatchdog.com/media/shots/inbox-muted.en.webp (A muted author: their comments arrive hidden, with the chip.)
Screenshot: https://commentwatchdog.com/media/shots/inbox-muted-phone.en.webp (The same on a phone.)
You should see: Their next comments arrive hidden and carry a Muted author chip, so the team knows why.
If something is off
Problem: A real customer was hidden
Fix: Their comment matched a rule, not their account. Unhide it, add the words that must survive to the Whitelist, and for Negativity write the note so it knows what a fair complaint looks like for you.
Problem: The same person is back under the next ad from a new account
Fix: The rules catch the pattern, not the account, so the new account is caught on the same words. Mute or block the ones you recognise.
Problem: Moderation Assist hid something first
Fix: It arrives in the inbox marked hidden outside the app. Keep both on, or turn Moderation Assist off in Facebook and let the rules decide alone.
Problem: The spam is on an old post, not a new comment
Fix: Rules run on comments that arrive from now on. Post cleanup fetches every comment on a post or ad link and checks them against the same rules.
Common questions
Question: Does CommentWatchDog look at the commenter's account age, profile picture or friend count?
Answer: No. Every rule judges the comment: the AI rules read what it means, in any language, and the Impersonators rule reads the author's name and the comment together. The text rules match links, contact details and the words you add. An old account with a photo and friends is caught on what it posts, which is what Facebook's account-based criteria let through.
Question: Can I keep Facebook Moderation Assist on at the same time?
Answer: Yes. Moderation Assist runs inside Facebook and CommentWatchDog runs on the comments that reach your Page through Meta's API. A comment Moderation Assist hides first arrives in the inbox marked as hidden outside the app; the rest is judged by your rules.
Question: What happens when the spammer changes the wording?
Answer: The AI rules read meaning, so a reworded scam matches the same topic. Only the Keywords rule needs the exact words. When a new pattern appears, describe it once under Custom topics and test it in the Playground.
Question: Which plans include this?
Answer: The preset AI moderation, Profanity and Negativity, and the text rules run on every plan, Mini and Lite included. Custom topics, the Negativity note and the Impersonators check start on Starter.
Related guides
Catch impersonators: https://commentwatchdog.com/help/moderation-rules/catch-impersonators
Custom topics in plain words: https://commentwatchdog.com/help/moderation-rules/custom-topics-in-plain-words
Mute or block an author: https://commentwatchdog.com/help/inbox/mute-or-block-an-author
Plain text of this guide: https://commentwatchdog.com/help/moderation-rules/spam-from-accounts-that-look-real.md
All guides for assistants: https://commentwatchdog.com/help/llms.txt