Most bans don't happen because you "sent too much." They happen because you crossed the line on three specific metrics - and the system reacted automatically. Meta doesn't warn you, doesn't send notices, doesn't explain. The account just disappears.
Meta publishes no official thresholds for regular (non-WABA) accounts. Official docs say only this: an account that receives complaints "may be temporarily or permanently blocked." That's it.
The figures in this article come from practice: forums, Reddit, operator case studies, observations across hundreds of accounts. This is not Meta policy - these are practical benchmarks. Where sources disagree, I'll say so directly.
This is the only metric that can kill an account instantly, regardless of everything else.
Critical threshold: 3–5 complaints out of 100 messages, collected within 1–2 hours.
The system looks not at percentage but at complaint velocity. Three unique users tap "Report and Block" within an hour - the algorithm treats it as a spam burst and triggers a ban. The fact that the other 97 recipients didn't complain doesn't matter.
Two recipient actions matter here:
| Action | System signal | Consequence |
|---|---|---|
| "Block" | Soft negative | Lowers account trust |
| "Report and Block" | Direct anti-spam signal | Instant ban trigger |
The difference is fundamental. Block is a personal choice. A report is a signal into Meta's moderation system.
What changes the picture - volume context. In practice, 3 complaints on 500 messages from a warmed account carry roughly 40–50% ban risk, not a certainty. 3 complaints on 50 messages from a new account - almost guaranteed ban. The system weighs proportion - but only if the account has history.
If the same 3–5 complaints are spread over 8–12 hours instead of one or two - the chance of a "soft" response (24-hour restriction instead of full ban) is noticeably higher. Complaint concentration in time is the key trigger.
This metric doesn't kill immediately. It slowly erodes internal account trust - and makes the account vulnerable to any next negative signal.
Critical threshold: more than 70–80% unread out of messages sent.
If fewer than 20–30 out of 100 messages were opened within a few hours - the system reads it as carpet-bomb outreach with irrelevant content. Not spam in a legal sense, but a behavior pattern indistinguishable from spam.
The mistake almost everyone makes: delivered = read. Two gray checkmarks mean delivery to the phone only. Until the chat is opened, the message counts as unread and works against the account.
The mechanics: high unread rate → falling internal trust score → on the next complaint, the system reacts harder than it would on a "healthy" account.
Critical threshold: more than 15–20% with a single gray checkmark within an hour of sending.
One gray checkmark means the message never reached the recipient's phone - invalid number, deleted or banned account. For anti-spam, the conclusion is clear: the list was built automatically (parser, generator, leak dump), not gathered organically.
A normal person can't message 20 non-existent numbers in a day - simply because they don't know they're non-existent. A bulk sender can, and the system knows it.
In practice: if 20+ out of 100 sends fail - that's not just lost reach, it's a direct signal about list quality.
| Metric | Safe zone | Danger zone | Critical point |
|---|---|---|---|
| Complaints (Report and Block) | 0–1 per 100 | 2 per 100 | 3–5 within 1–2 hours |
| Unread | up to 40% | 40–70% | above 75–80% |
| Undelivered | up to 5% | 5–15% | above 15–20% |
These three metrics don't work in isolation - they multiply risk. An account with 60% unread and 10% undelivered that gets 2 complaints is in a fundamentally different position than one with all three in the green zone.
Some sources say "3 complaints = ban," others say "3 complaints on 500 messages is fine." Both are right - they're talking about different accounts.
| Scenario | Complaints | Sent | Ban risk |
|---|---|---|---|
| New account (< 2 weeks), low volume | 1–2 | < 100 | 80–100% |
| New account, medium volume | 2–3 | 100–300 | ~90% |
| Warmed (> 1 month), medium volume | 1–2 | 100–300 | ~70% |
| Warmed, high volume | 3 | 300–600 | ~60% |
| Warmed, very high volume | 3 | 600–1000 | ~40% |
Risk drops another ~20% if complaints arrive evenly over 24 hours instead of clustering in the first hours after launch.
Important caveat: all figures are practical benchmarks, not official Meta thresholds. The algorithm weighs dozens more signals: whether recipients save your number, reply, what content type you send. The table is a guide, not a guarantee.
1. List validation. Run through a WhatsApp checker before the first send. Goal - remove dead accounts and push undelivered toward zero. The only way to control the third metric.
2. Ramp-up warmup. A new account doesn't start with bulk sends. First 7–10 days - real conversations, replies, groups. Day 1–2: 5–10 contacts. Day 3–4: up to 30. Day 5–7: up to 100, tracking replies. More on warmup and account protection in a separate breakdown.
3. First message that gets the chat opened. Not a pitch, not a link - just an open. Once the recipient opens and reads, the unread marker clears. The formula for a first message that gets replies is its own topic - but it directly affects metric 2.
4. Stop at two complaints per hour. Don't wait for a third. Pause at least 24 hours, then resume with lower volume and tighter list segmentation. If you've already been banned - review the technical reasons before repeating the same playbook.
If you're planning your first campaign - start with list validation and account warmup. If you've already launched and got bans - look at your numbers through these three metrics: the problem is usually obvious immediately.
Practical rule:
WhatsApp doesn't ban for volume - it bans for complaint speed and list quality. Three complaints in an hour from a warmed account with a clean list is not the same as three complaints in an hour from a new account with a scraped database.
