The WhatsApp Broadcast Rules Nobody Explains (Why Your Free Campaign Gets Banned in 24 Hours)

On 14 March 2026 I lost a WhatsApp Business number I had been using for eight months. On 2 May 2026 I lost the replacement number in under 24 hours. Two bans, two numbers gone, one very confused founder staring at a red "This account can no longer use WhatsApp Business" screen with zero appeal option that actually worked.

I was not running a scam. I was not sending "GET RICH QUICK" links. I was a founder sending onboarding messages to people who had literally typed their number into my signup form the week before. And WhatsApp still killed the number. Twice.

The reason nobody explains the actual broadcast rules is that the people who write blog posts about "WhatsApp marketing tips" have never had a number banned. They read Meta's public docs, rephrased them, and moved on. Meta's public docs are technically correct and practically useless — they tell you what the rules are named, not where the tripwires actually sit.

This post is what I wish someone had handed me in February 2026. Every number in here is a real number I have hit with my own account, or the account of an OT1-Pro customer I have watched get sandboxed.

What actually gets a number banned (it is not "spam")

Meta does not read your messages and decide if they are spammy. That is a myth. The system is dumber and more brutal than that. It watches three signals:

1. Block ratio — how many recipients tap the three-dot menu and hit "Block" within 48 hours of receiving your message. If more than roughly 2% of your daily recipients block you, your quality rating drops within hours.
2. Report ratio — same mechanic, but for "Report" instead of "Block". This one is weighted harder. A single "Report" is worth roughly 5 "Blocks" in the internal scoring I have reverse-engineered from watching my dashboard.
3. Non-response ratio — how many business-initiated messages you send that never get a reply. Meta does not publish this threshold. From my two bans I estimate it is around 70% non-response over a rolling 24-hour window on new numbers.

Notice what is not on the list: content, keywords, links, emoji count. You can send a message that reads "Hey, your order shipped" and get banned if 3% of recipients block you. You can send a message full of dollar signs and it will pass if people reply.

My first ban: the "warm list" that was not warm

March 2026. I had 847 email subscribers from a lead magnet. I exported their phone numbers (they had entered them at signup, tickbox and everything) and imported into WhatsApp Business App. I sent a broadcast that said: "Hi {name}, this is Omar from OT1-Pro. You signed up for my WhatsApp automation guide last month — do you want me to send the follow-up?"

Polite. Named. Referenced the signup. Consented in the loosest legal sense.

Banned in 11 hours.

Post-mortem: those 847 people had signed up two to eight months earlier. Most did not remember me. Around 4% blocked me on the spot. Another 1% reported me. My non-response rate hit 82% because most people just ignored it. All three signals red-lined at once. Meta did not care that they had technically opted in.

The real WhatsApp broadcast rules (rate limits and tiers)

WhatsApp has four messaging tiers on the Cloud API. Nobody puts them in one clear table so here is mine:

Two things people always miss reading this. First, "conversations" not "messages" — one 24-hour window with the same person counts as one conversation no matter how many messages fly. Second, the counter is unique users, not total sends. Sending 10,000 messages to the same 500 people does not unlock Tier 2.

The 256 number you have probably heard — that is the broadcast list cap in the free WhatsApp Business App, not the Cloud API. Broadcast lists on the free app can hold up to 256 contacts, and every recipient has to have saved your number in their phone contacts or the message silently does not deliver. That last part is the ambush. You think you sent 256 messages. Meta thinks you sent 40 that landed and 216 to strangers.

Quality rating: green, yellow, red

Every Cloud API number has a live quality rating you can see in Meta Business Manager under WhatsApp Accounts, Phone Numbers. It updates every few hours.

My second number went from Green to Red in a single 6-hour broadcast. Once you hit Red, you have roughly 24-48 hours to see zero new blocks and reports before the ban lands. In practice you cannot fix it in that window because you cannot un-send.

Opt-in language: what Meta actually accepts

Meta reviews templates by hand and their reviewers are inconsistent, but I have submitted around 60 templates across three accounts and the pattern is now clear. This table is my working cheat sheet:

The one that catches most founders is the "free guide" line. If your signup form does not literally use the word WhatsApp and mention ongoing messages, you do not have valid opt-in under Meta's rules, no matter what your privacy policy says.

Template messages vs session messages (the 24h window)

Every WhatsApp message you send falls into one of two buckets and mixing them up is what gets founders banned.

Session messages are free-form messages sent inside the 24-hour customer service window. That window opens the moment a user sends you a message and closes exactly 24 hours after their last inbound. Inside it you can send anything — text, images, PDFs, follow-up questions — and it costs nothing.

Template messages are pre-approved messages you send outside the 24h window or to start a conversation. Every template has to be submitted to Meta, categorised (marketing, utility, authentication), and manually approved. Marketing templates cost the most per conversation and get the harshest quality scoring. Utility templates (order updates, appointment reminders) cost less and are scored more forgivingly.

The trap: if a user messages you at 09:00 Monday, you have until 09:00 Tuesday to reply in free-form. At 09:01 Tuesday you can no longer send them a plain text. You must send an approved template or nothing at all. Founders who reply late, using the same text they used yesterday, wonder why the message shows one grey tick forever.

My second ban: the "reactivation" campaign

May 2026. New number, freshly warmed, quality rating High. I got clever. I sent a marketing template to 4,000 people who had messaged us in the past 90 days but never bought: "Hey {name}, we launched a Starter plan at $29/mo — want a walkthrough?"

Template was approved by Meta the day before. All 4,000 had messaged us first at some point. I thought this was airtight.

Banned in 19 hours. Post-mortem: of the 4,000, around 2,600 had messaged us more than 30 days ago and did not remember the conversation. Block rate hit 3.1%, report rate 0.4%. Red rating in six hours, ban in nineteen.

Lesson from ban number two: template approval is not permission. Meta approves the wording; the quality rating still judges the audience. A perfect template sent to a cold list will kill you exactly as fast as a bad template sent to a warm one.

The free path from Business App to Cloud API

Here is the part almost nobody explains clearly: Cloud API is free. Meta hosts it, you do not pay per message for the API itself, you only pay for conversations at the rates Meta publishes per country. In the US around $0.025 per marketing conversation, in India around $0.0135, in Egypt around $0.0384. The first 1,000 service conversations per month are free.

You do not need Twilio ($0.005 per message on top of Meta's rate), you do not need 360Dialog ($49/mo minimum), and you absolutely do not need WATI ($49/mo starter) to access the same API they are all reselling. The BSP (Business Solution Provider) model exists because setting up Cloud API used to be genuinely hard. In 2026 you can do it in about 40 minutes if you already have a Meta Business account.

The catch: you need a verified Meta Business Portfolio before Meta will approve your WhatsApp Business Account for anything above the sandbox tier. I wrote up the whole verification path in my meta app verification founder guide — same verification unlocks Facebook Login, Instagram Messaging, and WhatsApp Cloud API in one go.

Why WATI at $49/mo does not save you from bans

I get this question every week: "If I use WATI or Respond.io, will they protect my number?"

No. They physically cannot. WATI ($49/mo Starter), Respond.io ($79/mo Growth), 360Dialog ($49/mo minimum), and OT1-Pro Starter ($29/mo) are all sending through the same Meta Cloud API pipes. The quality rating, the block ratio, the tier limits — those live on Meta's side, not the BSP's side. Your account will get banned at exactly the same block ratio whether you paid $29 or $79 a month.

What a good tool can do is stop you from sending to cold contacts in the first place. That is the actual value. Not "protection", but "guardrails". I built OT1-Pro's broadcast module after my second ban with three hard-coded rules: no broadcasts to contacts older than 30 days since last inbound, no marketing template sends above 20% of your Tier limit per hour, and automatic pause if quality drops to Medium. See how those compare to WATI's rules on the OT1-Pro vs WATI page.

The rules I follow now (that have kept my third number alive for 4 months)

1. Never broadcast to a contact who has not messaged me in the last 30 days. Ever. Use email for reactivation.
2. Warm every new number for 14 days: send under 50 messages per day, only to people who message us first.
3. Marketing templates get sent to segments of maximum 500 people per hour, spread across the day, never more than 3,000 per day on Tier 2.
4. Every broadcast has a reply-worthy first line — a question, a specific offer, something that pulls a reply. Non-response rate is the silent killer.
5. Watch quality rating twice a day. If it drops to Medium, pause all outbound for 48 hours.
6. Never import a list from anywhere. Not from a CSV a partner gave me, not from an old CRM, nothing. Only contacts collected on my own forms with explicit WhatsApp opt-in wording.

What to do if you are already banned

Meta has an appeals form at business.facebook.com/business/help. In my experience appeals succeed roughly 1 in 5 times, and only if you can show the account was genuinely mistaken for spam (screenshots of your opt-in flow, sample of your messages, proof of business registration). If your ban was earned — real bad block ratio, real cold list — the appeal will be denied and reapplying with the same number is impossible.

Your realistic path is: get a fresh phone number (a new SIM, not a VoIP number — Meta rejects most VoIP ranges), start a new WhatsApp Business Account on that number, and rebuild slowly. The BSP you used before does not matter. What matters is the audience you send to.

The one-line summary I give every founder who asks

WhatsApp does not ban you for what you send. It bans you for who you send it to. Fix the audience question — real recent opt-ins, warm segments, reply-worthy messages — and you can send freely at $29 a month. Ignore it and no amount of WATI money will save you.

If you want the broadcast guardrails I built after ban number two baked into your account from day one, OT1-Pro Starter is $29/mo and includes the 30-day recency rule, hourly rate caps, and quality-rating auto-pause out of the box. Or just create a free account and connect your existing WhatsApp Cloud API number — the guardrails work on the free plan too.

Run WhatsApp, Instagram, Messenger, Telegram, and email from one inbox

OT1-Pro is the unified inbox I built after losing too many deals to slow, siloed replies. One AI agent, five platforms, one conversation per human — replies in your brand voice, respects Meta's 24-hour window, and escalates to you only on real deals. Free forever plan, no credit card, founder-accessible on WhatsApp.

Start free → · Pricing from $8/mo · Why we beat WATI · The Meta verification guide founders actually need · Talk to me on WhatsApp