A ban doesn't land out of nowhere. It's the end of accumulated negativity that built for weeks while metrics still looked "fine." There's no official death date - Meta doesn't publish a decay formula. But the signs show up before a block, and that's where you start.
In official WhatsApp Business API (WABA), number quality is measured by Quality Rating. The window is the last 7 days of interactions. Main negative signals: user reports, contact blocks, other explicit negative feedback.
Quality statuses:
When Flagged, Meta gives 7 days to fix the causes. If quality doesn't improve - limits drop, scaling locks.
Meta doesn't publish exact thresholds or the formula. Anything like "0.1% reports guaranteed triggers decay" or "Trust Score drops X points per report" - operator observations, not official data.
The main trap is system inertia. Reports, silent blocks, and negative signals don't trigger instant reactions. They pile up in the background while the number keeps working.
Operators observe accumulated negativity can surface weeks later. A blast that didn't get banned doesn't mean no consequences. The next campaign starts from a worse reputation base.
That's why judging a number only by "not banned yet" is wrong. Watch metric trends.
Official data and field practice diverge - worth stating plainly.
Operator observations (not Meta's official position):
One documented operator-community case: a three-year account with thousands of active chats - Response Rate fell from 25% to 4% over three months of aggressive weekly blasts, then on day 95 a routine message to a regular client triggered a permanent ban.
Important caveat: no formula exists. "2–3 months" is a practical benchmark, not an official number lifespan. Speed depends on list quality, blast frequency, volume, and audience reaction.
| Account type | Warm-up | Risk on aggressive start |
|---|---|---|
| New, no history | No | Very high, fast |
| New, 10–14 day warm-up | Yes | Reduced |
| Old warmed, good list | Yes | Low, but decay accumulates |
| Old warmed, cold list | N/A | Decay in 60–90 days |
Most practical section. A block is the finale. Before it, the system sends signals.
Signal 1: falling reply rate. If reply percentage drops week over week at stable send volume - not random list noise. Audience attitude toward the number is worsening.
Signal 2: rising undelivered. Messages stuck on one gray check (sent, not delivered) on a list that used to deliver fine. Falling Delivered Rate on a verified segment - practical red flag, though Meta publishes no official threshold.
Signal 3: Quality Rating turns yellow. Observable directly in WABA. Yellow means stop, not accelerate.
At first signs: pause blasts, check template relevance, prune inactive contacts, shift temporarily to inbound and service chats. Continuing blasts while Quality Rating degrades - straight path to Flagged and lower limits.
One confirmed fact: an inactive WhatsApp account can be deleted after 120 days without connecting to the network.
From operator practice: a number with good history idle for a long time, then launched straight into mass outreach - risky. A sharp activity spike after dormancy triggers antispam filters.
Calling an "idle penalty" an official Meta mechanism - you can't. It's observation, not a confirmed algorithm.
If the number sat idle - ramp up. Infrastructure operators recommend starting at 20–50 messages per day, doubling volume every 3 days, reaching working volume by day 10–14. Practice, not official Meta guidance. More on warm-up patterns.
"No ban after the blast - we're fine" Negativity accumulates in the background. No ban after one send doesn't mean no hit to the next.
"Old numbers are unkillable" Age softens risk but doesn't cancel decay under systematic negativity. A ban on day 95 hits like day one.
"Delivery is what matters" Delivered ≠ good. Meta scores reaction to delivered messages, not delivery alone.
"WABA session fees protect reputation" No. Session charges still run on a degrading number. Quality Rating affects limits and status regardless of spend.
Minimum operator checklist:
If two metrics fall together for several weeks - not random list noise.
Look at Response Rate trend over the last three weeks. If replies fall at stable volume - stop the blast and fix list quality before the next wave.
Practical rule:
Number reputation doesn't die from one blast - it dies from the habit of ignoring the symptoms.
