Avoid a WhatsApp Business Account ban in India

A suspended WhatsApp Business Account is one of the few failure modes that can take a messaging programme to zero overnight — and unlike a rejected template, it is slow and painful to reverse. The good news: WABA bans in India are almost never random. They trace to three causes, every one of which is preventable. Here is how Meta's quality system actually decides, and the playbook to stay green.

How Meta's quality system works

Every WABA phone number carries two signals that move together:

Quality rating — Green (high), Yellow (medium), Red (low). It is computed from recent user feedback: blocks, "report" taps, and how many of your sends are read vs ignored over a rolling window.
Messaging limit tier — the number of unique users you can message in 24 hours: 250 → 1K → 10K → 100K → unlimited. You climb tiers by sending quality traffic; you get knocked down (or frozen) when quality drops.

A ban is the end state of a quality spiral: Red rating → rate-limited → flagged → restricted → disabled. The system is reactive, so the damage is usually done before you see the Red badge. Prevention beats remediation by a wide margin.

The spiral, one rung at a time

Green · highSending quality trafficTiers climb: 250 → 1K → 10K → 100K → unlimited unique users in 24 hours.
Yellow · mediumBlocks and reports are accumulatingThe early warning, and the last rung where pausing to diagnose is enough.
Red · lowComputed from recent feedback over a rolling windowBy the time the badge shows, the traffic that caused it has already gone out.
Rate-limited, then flaggedSending is throttledYour tier is knocked down or frozen, whatever your list size.
Restricted, then disabledThe account is downSlow and painful to reverse — unlike a rejected template, which you can fix and resubmit the same day.

Nothing on this ladder happens quickly, which is the one piece of good news in it: every rung is entered days before the next one.

Cause 1 — marketing-categorised bulk sends

The single most common trigger. Teams push transactional-natured messages (order updates, payment alerts, reminders) through the MARKETING template category — either by mistake or because the copy reads promotional — then blast them at volume. Marketing templates require opt-in, are quality-scored aggressively, and generate the block/report feedback that tanks your rating fastest.

The fix is categorisation discipline: route every message that genuinely is utility (the user expects it, it's tied to an action they took) through the UTILITY category, and write the copy so Meta agrees. If you find your rating dropping on "transactional" sends, check the actual category on Meta — it is probably marketing. See our template categorisation playbook for the rewrite rules.

Cause 2 — low-quality template copy

Even within the right category, copy quality drives the block rate. The patterns that generate complaints:

Messages the user didn't expect. If a recipient can't immediately tell why you're messaging them, they block. Lead with the context ("Your Acme order #1234 has shipped"), not a greeting.
Too-frequent sends. Three notifications about the same order in an hour reads as spam regardless of category.
Stale or scraped lists. Messaging numbers that never opted in or haven't engaged in months produces a spike of blocks that the quality model reads as a spam signal.

Rule of thumb: every template should map to a real event the user would expect a notification for. If you can't name that event, don't send it.

Cause 3 — naive flat-rate blasting

New or recently-warmed numbers that suddenly send tens of thousands of messages in a flat burst look exactly like a compromised account to Meta's systems. The pattern, not just the volume, is the flag.

The fix is rate shaping: ramp volume gradually so it tracks your tier, spread sends across the day rather than one spike, and let early engagement (reads, replies) earn the next volume step before you take it. This is the same logic as email IP warming, applied to WhatsApp.

The stay-green checklist

Categorise honestly. Utility content in the utility lane; only true promotions in marketing, with opt-in.
Review copy before submission. Lead with context, cut promotional language, one clear purpose per template.
Warm new numbers. Ramp volume over days, not in one burst; let the tier climb earn the next step.
Clean your lists. Only message users who opted in or transacted; drop long-dormant numbers.
Watch the rating daily. A drop from Green to Yellow is your early warning — pause and diagnose before it hits Red.
Cap frequency. De-duplicate and rate-limit per user so nobody gets three messages about one event.

If you're already restricted

If your number is rate-limited or flagged: stop all non-essential sends immediately, let the rating recover over the rolling window (days, not hours), and open a Meta Business support case if you believe the action was wrong. Do not spin up a fresh number and blast the same content — Meta links numbers under a business and the pattern follows you.

How QuickAuth keeps accounts green

Avoiding a ban is mostly process, and process is the first thing that gets skipped under launch pressure. So the protections below are not advice we give you — they are how a send behaves by default, whether or not anyone remembers the checklist.

A 1% canary goes before the list. Every campaign sends to a small probe first, bounded between 50 and 1,000 recipients — campaigns under 1,000 use a flat 20-message probe, and a send of 20 or fewer skips it. A template that is going to draw reports draws them from the probe, not from your whole audience.
Health is re-checked between every wave. After each batch, WABA and template quality are read again, and the campaign halts rather than continuing if the template goes RED or every number in the pool does. This is the control that turns a bad hour into a stopped campaign instead of a disabled account.
Opt-out is enforced at send time, and never billed. A contact who replies with a stop keyword is skipped on the next campaign and is not charged for. Reports from people who asked you to stop are among the strongest ban signals there is.
Pacing adapts to the size of the send. Concurrency is set from the campaign's own size rather than a flat rate, so volume ramps in a shape Meta's quality systems tolerate instead of arriving as a spike.
Load is spread across a pool of senders. No single number carries a whole campaign, so one number's rating cannot take your entire programme down with it.

Templates are also categorised on submission, so utility content is submitted as utility rather than being reclassified later — the quiet cause behind a large share of the ratings drops described above.

What none of this does is guarantee you will never be restricted. Meta decides, the criteria are not published, and anyone promising immunity is selling something. What it does is make the three common causes hard to do by accident, and stop a send early when the signals turn — which is the difference between a paused campaign and an appeal with a one-in-ten chance. If your account is already down, the recovery path is written up separately in what actually works to get a banned WABA back.

A ban is preventable. Categorise honestly, write for the user, ramp gradually — and watch the rating the way you would watch error rates on any other production system.