Build a Comment Moderation Word List That Works

To hide comments odn a Facebook post before anyone else sees them, add a blocked word list in your Page settings so any comment containing a flagged term is hidden the moment it lands. It is the single fastest control you have, it costs nothing, and it runs 24/7 without you refreshing a feed. The catch is that a lazy list either misses the spam you care about or nukes real customers by accident.

This guide walks through building a word list that pulls its weight: which categories to block, a copy-paste starter set for Facebook and Instagram, and where the list stops working so you know when to bring in AI.

Why a Word List Is the Fastest Way to Hide Comments on a Facebook Post

Spam is not a rounding error in your comments. In a 2025 analysis of 168.8 million comments across 7.6 million posts, roughly one in five comments contained spam, bot activity, or abuse, according to Respondology's 2026 Business of Comments Report. The same research found that spam and bot comments accounted for 41.8% of all comments brands chose to hide, which tells you most moderation work is not about angry customers. It is about junk.

That junk has a price. When it lands on paid posts, the same report measured a 14.7% drop in conversion on ads carrying spam and bot comments. So a scam link under your best-performing ad is not just ugly. It is quietly taxing every dollar you spend to put that ad in front of people.

A word list is your cheapest defense because it acts before a human ever looks. You define the terms once, and Facebook hides any matching comment automatically, keeping it visible only to its author. That last detail matters: hiding avoids the confrontation and engagement loss that comes with deleting, which is why it is the preferred move for ads. If you want the full rundown on hiding versus deleting, see our starter guide to social media comment moderation.

Categories to Block: Profanity, Scams, Links, and Numbers

A good list is organized by intent, not thrown together as one giant blob. Four categories cover the vast majority of what shows up under ads and organic posts.

Profanity and slurs

The obvious one. Facebook ships a built-in profanity filter you can turn on, but its list is unpublished and conservative, so it misses creative spellings and leaves gaps. A custom list lets you add the specific words your audience actually uses to attack, plus the leetspeak and spaced-out variants (like "s p a m") that slip past defaults.

Scam and fraud phrases

This is where the money leaks. Scammers impersonate your brand in replies, promising refunds, giveaways, or "account verification." Block the phrases they lean on: "message this number," "claim your prize," "you have been selected," "verify your account," "DM to claim," and "customer support" paired with a handle that is not yours. Impersonation scams are one of the top reasons brands need active moderation on ads.

Competitor and off-platform links

Comment sections on your ads are prime real estate for competitors and affiliate spammers dropping links. Blocking common URL fragments (like "http," "www," ".shop," "bit.ly," and "wa.me") catches most of them, which is often the highest-value category on paid posts.

Phone numbers and contact bait

Scammers post phone numbers to move victims into WhatsApp or SMS where you cannot see the conversation. Static word lists struggle with raw digits, but you can block common framing like "call now," "whatsapp me," "text," and "+1." This category is where you will feel the limits of keywords most, and it is a good preview of why AI matters later.

Copy-Paste Starter List for Facebook and Instagram

Drop these into your blocked words settings and adjust for your niche. Keep each on its own line in your Page tools. This is a starting point, not a finished list, so expect to add and prune over your first few weeks.

Scam and fraud

• claim your prize
• you have been selected
• verify your account
• message this number
• dm to claim
• congratulations you won
• limited slots
• investment opportunity
• guaranteed returns
• act now

Links and off-platform pulls

• http
• www
• bit.ly
• wa.me

-.shop
-.xyz

• link in bio
• check my page
• visit my profile

Contact bait

• call now
• text me
• whatsapp
• +1
• inbox me
• reach out at

Profanity and abuse

• add your platform's slur list here
• common spellings and spaced variants (f r e e, s c a m)
• refund scam, fake, fraud (tune these carefully to avoid catching genuine complaints)

One caution on that last group: words like "refund," "fake," and "scam" show up in legitimate customer questions too. Blocking them outright can hide a real complaint you needed to answer. Use them only if your team reviews the hidden queue, or leave sentiment calls to AI. On Instagram, the same blocked-words settings live in your account tools, so the starter list ports over with minor tweaks.

The Limits of Word Lists, and Where AI Picks Up the Slack

Static lists break in predictable ways. Spammers know your filter reads exact strings, so they write "fr€e," "sc4m," or "�888" with unicode lookalikes. They swap words into images and emoji-only comments that carry no text to match. They rephrase every message so no two are identical. A keyword list cannot read tone either, so it cannot tell an angry-but-fair complaint from a coordinated troll pile-on.

This is the gap AI closes. Instead of matching strings, a real-time AI filter reads meaning: it recognizes a scam pattern it has never seen the exact wording of, flags a comment as harassment based on intent, and handles 50-plus languages without you maintaining a list for each one. Sweep Inbox layers this on top of your keyword rules, hiding spam, scams, trolls, and hateful comments within a few seconds of them posting, using Meta's official Graph API and webhooks rather than scraping. The word list handles the obvious, the AI handles the clever, and together they cover far more than either alone.

Speed is part of why this matters on ads. A scam comment that sits for an hour under a high-traffic ad reaches thousands of people. Catching it in seconds keeps your comment section clean while the ad is at its most active.

Per-Page Tuning So Each Brand Gets Its Own Rules

A supplement brand gets different spam than a SaaS company or a local restaurant. One fights "miracle cure" claims and fake reviews, another fights crypto bots, the third fights competitor menus and off-topic links. A single global list forces all of them into the same rules and guarantees both misses and false positives.

Per-Page rules solve this. Agencies feel it most: managing ten client Pages with one shared blocklist means a term that protects Client A creates false positives for Client B. Sweep Inbox supports per-Page rules under one plan, so each brand runs its own keyword set and AI sensitivity while your team works from a single unified inbox. If you handle multiple accounts, our comparison of social media moderation tools breaks down what to look for.

Practical tuning tips:

• Review your hidden queue weekly for the first month and move any false positives off the list.
• Add new terms the moment you spot a fresh spam wave, since these tend to run in coordinated bursts.
• Keep profanity and abuse rules tighter on ads (where reach is bought) than on niche organic posts.
• Log which terms fire most so you can retire dead weight and focus the list.

Where to Go From Here

Start today by opening your Facebook Page settings and pasting in the starter list above, then split it into the four categories so you can tune each one without rewriting the whole thing. That alone will hide the bulk of scam links and junk under your ads within minutes.

Then close the gap the keywords leave. Layer real-time AI moderation over your word list so misspellings, emoji spam, and multilingual trolls get caught the second they post, not the next time you check your phone. With 47% of consumers holding brands responsible for the toxic comments on their pages, a clean comment section is brand protection, not housekeeping. Sweep Inbox pairs your keyword rules with real-time AI filtering across every connected Page, so try layering the two and let your ads run without a scam link riding along.