Summarise this article with:
On this page
- What WhatsApp quality rating measures — and what it gates
- Template quality and the 3-hour, 6-hour, disabled ladder
- The sending patterns that turn a number yellow fastest
- RED: the first hour, then the 7-day recovery runbook
- Audience hygiene: the system that keeps you GREEN
- What most teams get wrong about quality rating
- Monitoring: webhooks, thresholds and who gets paged
- Get started with InfiQ
A broadcast goes out at 7pm on a Friday. By Saturday morning the number reads YELLOW in WhatsApp Manager, two templates are paused, and the weekend campaign is dead. Nobody changed anything — same list, same template, same platform as last month. What changed is what recipients did with the message.
That gap is why WhatsApp quality rating catches teams out. Delivery dashboards look healthy — sent, delivered, read — while the score that governs how many people you may reach is written by signals no campaign report shows: the people who tapped Block or Report instead of reading.
Key points
- WhatsApp quality rating has four states:
GREEN(high),YELLOW(medium — "may soon be paused or disabled"),RED(low — "in danger of being paused or disabled soon"), andUNKNOWN(pending, no feedback yet). - Template pausing follows a fixed ladder: a first low-quality instance pauses the template for 3 hours, a second for 6 hours, and a third disables it.
- Messaging limits are set at the business portfolio level and shared across every phone number in that portfolio, per Meta's messaging limits documentation.
- Automatic scaling up the ladder — 2,000 to 10,000 to 100,000 to Unlimited — depends on sustained quality, so a number stuck at YELLOW stops climbing.
What WhatsApp quality rating measures — and what it gates
A WhatsApp quality rating is a score Meta assigns to an individual business phone number based on how recent recipients responded to messages from that number, shown as GREEN, YELLOW, RED or UNKNOWN. Meta's messaging limits documentation ties the rating to user feedback — blocks and the Report action inside a chat — rather than to delivery receipts, latency or API errors.
That distinction is the whole ballgame. A campaign can deliver to 100% of a list and still shred your rating, because the minority who blocked you count for more than the majority who read you. Teams that treat this as a delivery problem spend a week fixing phone-number formats while the real cause — frequency, or a list that never really opted in — goes untouched.
The four states
| Rating | Meta's label | Meta's wording | Action |
|---|---|---|---|
GREEN |
High quality | — | Hold the cadence; green is not headroom |
YELLOW |
Medium quality | "may soon be paused or disabled" | Freeze non-essential marketing today |
RED |
Low quality | "in danger of being paused or disabled soon" | Stop all marketing; run first-hour triage |
UNKNOWN |
Pending | No feedback received yet | New or low-volume number; ramp slowly |
Meta publishes neither a block-rate threshold nor its lookback window, so any article quoting a precise percentage is guessing.
What the rating gates
A messaging limit is the maximum number of unique users a business can message outside an open 24-hour Customer Service Window in a rolling 24-hour period. New portfolios start at 250. The 2,000 tier is reached through business verification, partner verification, or delivering 2,000 high-quality messages in 30 days, after which automatic scaling raises the limit to 10,000, then 100,000, then Unlimited, based on sustained quality and utilisation. That word sustained is where quality does its work: a number oscillating between GREEN and YELLOW is quietly frozen rather than downgraded, which is why some teams never leave 10,000. Our WhatsApp messaging limits and tiers guide covers the full ladder and the deprecated messaging_limit_tier field.
Quality gates send speed too. Throughput can scale to 1,000 messages per second, but only for numbers with a Medium or higher quality rating that also meet the eligibility conditions — an unlimited messaging limit plus 100,000 or more unique users per 24 hours, or unlimited initiation capability. A RED number is not eligible.
A third ceiling is invisible in most dashboards. WhatsApp dynamically caps how many marketing templates any single user receives across all businesses, based on that user's recent read-rate and inbox load. Per Meta's per-user marketing limits documentation it applies in India, though not in the EEA, UK, Japan or South Korea, and marketing templates are not delivered to +1 US numbers at all. The error code is 131049; retrying inside 24 hours risks further failures.
Template quality and the 3-hour, 6-hour, disabled ladder
Template pausing is Meta's automatic enforcement action that temporarily blocks a specific template from being sent after it accumulates low-quality user feedback. It runs on its own scoreboard, separate from the number-level rating, and it is usually what a team notices first.
| Template status | Your move |
|---|---|
| Active – Quality pending | Send in small waves, not the full list |
| Active – High quality | Scale within your messaging limit |
| Active – Medium quality | Rewrite or cut frequency now |
| Active – Low quality | Stop sending it; a pause is next |
| Paused | Change something before it unpauses |
| Disabled | Rebuild the message, do not clone it |
| Rejected | Fix the cause, appeal with a sample |
| Appeal Requested | Appeals are reviewed within 24 hours |
Per Meta's template pausing documentation, the first low-quality instance pauses a template for 3 hours, the second for 6 hours, and the third disables it. Paused templates unpause automatically or can be handled via the API and WhatsApp Manager.
Read that ladder as a timer with three lives. A three-hour pause landing at 6:30pm removes the whole evening send window in IST, which for a D2C brand is the window that matters. And the counter does not reset because you waited: unpause, resend the same copy, pause again for six hours, and the next strike is not a longer pause.
Categorisation feeds into this. Meta's auto-recategorisation process has run since 1 July 2024, typically moving mis-categorised templates from utility to marketing, and repeat offenders get no advance notice as of 16 April 2025. Our WhatsApp template categories guide covers classification and the misuse penalty ladder; the template rejection reasons post covers approval failures, a different problem.
The sending patterns that turn a number yellow fastest
The ordering reflects how quickly we see each pattern cause damage, not a published Meta ranking.
| Pattern | Why recipients react badly | Speed |
|---|---|---|
| First broadcast to a purchased or scraped list | No memory of your brand, so blocking is rational | Severe, same send |
| Marketing template to transactional-only opt-ins | Consent mismatch reads as bait-and-switch | Severe, same send |
| Re-adding contacts who opted out | Your most negative cohort, and they all block | Severe, cumulative |
| Same template to the same list more than weekly | Repetition reads as spam regardless of copy | High, within days |
| Sends outside 9am–8pm IST | An 11pm marketing ping is an intrusion | High, immediate |
| No visible opt-out in a marketing message | Users who cannot leave politely block instead | Moderate, cumulative |
The most effective fix is the frequency cap, because it works even when segmentation is mediocre. A platform-level rule — no more than two marketing templates per contact in any seven-day window — protects you from the calendar every growth team builds, where four teams each schedule "just one send" into the same base.
Worked example: frequency maths for a Kota coaching institute
Take a hypothetical coaching institute in Kota with 40,000 opted-in parent contacts, sitting at the 10,000 messaging tier. In counselling season it runs three marketing broadcasts a week: batch announcement, scholarship reminder, deadline push. Weekly sends: 40,000 × 3 = 120,000, so every parent gets three marketing templates in seven days.
Assume — purely as an arithmetic placeholder, not a benchmark — that 4 in every 1,000 recipients block or report. That is 40,000 × 0.004 = 160 negative signals per broadcast, or 480 a week, and the rate on the surviving audience climbs as blockers leave the reachable base while volume stays flat.
There is a second constraint nobody noticed. At the 10,000 tier the institute can message 10,000 unique users outside an open Customer Service Window per rolling 24 hours, so a 40,000-contact broadcast takes at least four days. Three a week is arithmetically impossible without queueing, which is why Monday's campaign is still landing on Thursday.
Restructure into four segments of 10,000 by exam year and stage, one relevant broadcast each per week. Weekly sends fall to 4 × 10,000 = 40,000, a two-thirds cut; each parent receives one template instead of three; negative signals fall to 10,000 × 0.004 × 4 = 160 a week rather than 480; and each send clears inside one 24-hour window. At the current India marketing rate per message — ₹0.94 today (ex-GST), call it R — weekly spend drops from 120,000 × R to 40,000 × R. Pull per-category figures from Meta's rate card or InfiQ's WhatsApp cost breakdown; Meta may only change rates on 1 January, 1 April, 1 July or 1 October. The aggressive plan sent three times the volume, cost three times as much, and generated three times the negative feedback.
RED: the first hour, then the 7-day recovery runbook
A RED rating means enforcement is close, so the first hour is triage, not diagnosis. Kill every scheduled marketing send across the whole portfolio — limits are portfolio-level, so a second number still pushing volume gives Meta a consistent negative signal from the same business. Pause trigger-based automations too: drip sequences keep firing long after someone clicks "pause campaigns". Note where each Paused or Low quality template sits on the ladder, since that tells you how many mistakes you have left. Then find the largest audience that received a marketing template in the last seven days and check its consent scope.
Do not rotate to a fresh number: it starts at UNKNOWN with no history, sits under the same portfolio limit, and — pointed at the same list and copy — will be YELLOW within a week. And do not ask Meta to reset the rating; it comes from recipient behaviour, so there is no reset lever, only new behaviour.
The runbook
- Day 1 — Freeze and audit. Stop all marketing templates portfolio-wide, but keep utility and service messaging running: order updates, reminders and replies are messages recipients want, and they generate positive signal. Export 30 days of sends and flag anything above 5,000 recipients, or any contact who got more than two marketing templates in a week.
- Day 2 — Rebuild the suppression list. Permanently suppress everyone who opted out, replied negatively, or went silent across three consecutive sends. If a CSV import can re-add a suppressed contact, fix that first — it is the most common cause of a repeat RED.
- Day 3 — Re-derive consent scope. Split the base into explicit marketing consent with a date and source, transactional consent only, and unknown provenance. The third group is a liability, not a segment. India's DPDP regime asks for valid, informed consent — also a workable test for whether someone will block you. Our WhatsApp opt-in and DPDP guide covers collection mechanics; confirm the regulatory position with your own counsel.
- Day 4 — Rewrite the offending templates. A paused template that unpauses unchanged pauses again one rung higher. Name the reason this contact is receiving the message, put the value in the first line, keep one call to action, add a plain opt-out. Submit on Day 4, because approvals and rejections take up to 24 hours.
- Day 5 — Restart with a 5% canary. Send one rewritten template to your warmest segment — recent purchasers, enquirers, anyone who replied in the last 30 days — capped near 5% of the base, then watch quality and replies for 24 hours. If a template pauses here, the problem is your audience, not your copy.
- Day 6 — Widen to 20% and staff the inbox. Send to a second warm segment inside the 10am–7pm IST window, with humans on the shared inbox for two hours afterwards. Replies open a 24-hour Customer Service Window, and messages inside it do not count toward your messaging limit.
- Day 7 — Set guardrails before resuming. Configure a cap of two marketing templates per contact per seven days, a 9am–8pm IST window, segment size limits, and an alert that pages a named person on any move off GREEN. Resume at half your previous weekly volume for two weeks. Teams tell us this step gets skipped, and that is why the same number goes RED next cycle.
If your stack cannot enforce a frequency cap, a permanent suppression list and a scheduling window without someone remembering to, that is a tooling problem — InfiQ's broadcast and campaign controls exist because these guardrails have to be structural.
Audience hygiene: the system that keeps you GREEN
Recovery is an event; hygiene is what stops you needing another one.
| Control | The rule that actually holds |
|---|---|
| Opt-in recency | Route contacts with 6–9 months of no interaction into a re-permission flow. Consent decays, and strangers block |
| Suppression | One permanent list of opt-outs, blocks, complaints and long-term non-engagers. If an upload can override it, it is a suggestion |
| Frequency caps | Cap at contact level across every team and automation, marketing templates only |
| Sending windows | 9am–8pm IST for marketing. Fee-due notices land differently at 9pm than at 11am |
Hygiene also has a cost dimension this year. Service messages have been free since 1 November 2024, and utility templates and non-template messages inside an open Customer Service Window since the July 2025 pricing change — but per Meta's non-template messages pricing documentation, Meta resumes charging for both from 1 October 2026. The inbox volume that repairs a damaged rating becomes a line item. Our post on the Meta Business Agent Platform and the October 2026 pricing change works through the budgeting.
What most teams get wrong about quality rating
They treat it as a delivery metric. Delivery says the message arrived; the rating says how people felt about it arriving. A team at 99% delivery with a rising block rate is in worse shape than one at 92% with none.
They read YELLOW as a warning to ride out. YELLOW is Meta saying the number "may soon be paused or disabled". Running the calendar through a YELLOW week bets that the next send lands better than the last, with no evidence for it.
They confuse rejection with quality failure. A rejected template failed review — usually parameter formatting such as mismatched {{1}} braces, special characters like # or %, non-sequential variable numbering, or duplicate content matching an existing template's body and footer. A low-quality template passed review and then failed with users. Meta's template review documentation covers the first; only your sending behaviour fixes the second.
They let retention flows run unsupervised. Each automation looks reasonable alone, but a contact in three flows can receive five marketing templates in a week with no campaign scheduled. Our abandoned cart recovery guide sets out a three-touch sequence that stays inside a frequency budget.
Monitoring: webhooks, thresholds and who gets paged
WhatsApp Manager is the authoritative surface for rating and template status per number, but the goal is a rating change reaching a named human within minutes rather than at the next weekly review. If you consume webhooks directly, subscribe to the account-level update events carrying phone-number and template quality changes. The InfiQ developer documentation covers subscription and payload handling, and our WhatsApp webhooks guide covers the patterns that stop you missing an event.
| Signal | Threshold | Alert | Response |
|---|---|---|---|
| Rating moves off GREEN | Any change | Growth lead, on-call ops | Freeze marketing within the hour |
| Rating reaches RED | Immediate | Growth lead, support head, owner | Run first-hour triage in full |
| Template hits Medium quality or Paused | First occurrence | Template owner, growth lead | Note ladder position; fix the copy |
| Error 131049 returned | Above your error floor | Engineering on-call | Stop the send; no retry inside 24h |
The 131049 trigger is one you set for yourself, not a Meta-published limit, and you should baseline your opt-out rate before you need it: the useful signal is deviation from your own history.
Get started with InfiQ
Quality rating problems are rarely single-campaign problems. They come from a stack that lets several teams schedule into the same base, lets an import resurrect opted-out contacts, and reports delivery rates instead of the signals Meta scores. The fix is structural: frequency caps and suppression enforced at the platform layer, send windows that hold in IST, and alerts that reach a named person the minute a number moves off GREEN. InfiQ onboards you as an official Meta Business Solution Provider, with the broadcast controls, template library, shared inbox and analytics that make those guardrails the default.
Ready to keep your number at high quality while you scale sends? Start your 7-day free trial — InfiQ gets you live on the official WhatsApp Business API in about 2 hours, with broadcast frequency controls and a shared team inbox that turns post-send replies into your best quality signal. Or book a walkthrough if you want to see it on your own use case first.
