<aside> 🗣

FYI: I wrote this for a broader audience, since we might publish this as a technical article. Feel free to skip over anything you already know!

This article is in 2 parts:

1. What we learned about WhatsApp quality ratings & limits
2. How we set up conversion events and ads </aside>

At Kaya Guides, we operate a guided self-help course for depression over WhatsApp. One of the pillars of our strategy is that, unlike other orgs in this space who typically recruit through partnerships, 100% of our users come from digital advertising. We’re huge proponents of it for any suitable intervention, and it’s cheap, scalable, and highly effective—when set up properly.

The problems, however, came as early as 6 months into delivery. WhatsApp were flagging us, saying we were spamming our users. We also began to see some templates failed to send to users. Naturally, we were confused—although the occasional user got stuck in the bot, we didn’t actively reach out to people without their consent, and had easy ways of opting out of the bot. What gives?

What we learned about WhatsApp quality ratings & limits

There’s a lot of apocryphal advice out there about quality ratings and limits, without much evidence. I’ll try to stick to specifically things from Meta, and include anything we’ve directly observed on the margins, without mixing in too many guesses.

Quality Rating and Messaging Limits - WhatsApp Business Platform - Documentation - Meta for Developers

For the basics, I’ll assume the reader has read Meta’s support article on the topic. When talking about anything in the Meta ecosystem, I find it helpful to acknowledge the objects we’re talking about, and the properties they have. So, here’s our cast of characters:

• Chat: A series of interactions between a user and a business. Every message ever sent between a user and a business is a part of their shared chat.
• Conversation: A back-and-forth between a user and a business, defined by a 24-hour inactivity window. In other words, a conversation lasts from its first message, to 24 hours after its last message. Any messages after that window start a new conversation.
• Template: A pre-written message that can be sent to a WhatsApp user. Businesses can only send messages to users after the 24-hour inactivity window if they send them a template.
  • Type: Templates can be categorised as Marketing, Utility, or Authentication. Restrictions are applied differently depending on the category.
  • Quality Rating: A traffic-light rating about whether this template is useful to users. May be Low, Medium, High, or Unknown (in the APIs, these are RED, YELLOW, GREEN, and UNKNOWN, respectively).
• Phone Number: An individual phone number registered with WhatsApp.
• WhatsApp Business Account (WABA): An account that may hold multiple phone numbers. Meta calculates the below properties per phone number, but all phone numbers on the account share the highest of all limits for the purposes of figuring out whether they can send messages to users. This means you can make progress up the messaging limits for one number, and new numbers you add will start at that limit, rather than from scratch.
  • Messaging Limit: The number of new conversations any of the accounts’ phone numbers can initiate in a 24-hour period, by sending a template to an inactive user. Can be 250, 2,000, 10,000, or Unlimited (in the API, there are additional tiers that might exist for legacy reasons; I couldn’t find any evidence that they’re used).
  • Quality Rating: A traffic-light rating about whether the account’s phone numbers is behaving well toward its users. May be Low, Medium, High, or Unknown.
  • Status: A sort of catch-all that indicates how the phone number is going. In the API, it may be PENDING, DELETED, MIGRATED, BANNED, RATE_LIMITED, CONNECTED, DISCONNECTED, UNKNOWN, or UNVERIFIED. This used to be important, but we don’t need to think about it anymore.

The ways WhatsApp restrict messaging

As of October 7, 2025, WhatsApp have 4 major ways they handle anti-spam:

1. Accounts have a quality rating and messaging limit, that determines whether they can consistently increase their 24-hour limits for initiating new conversations. It’s unclear whether this rating is used for other enforcements.
   1. Before October 7, 2025, a consistently low quality rating would reduce your messaging limit, but this enforcement no longer applies.
2. Templates have a quality rating, that’s used to determine whether that specific template can be sent out to users.
3. Templates are paced, meaning that under certain conditions, Meta restrict how often you can send them and more carefully monitor for feedback. We have never seen template pacing in practice, so don’t have a sense of how it works or how to avoid it.
4. Users have a Marketing messaging limit, that’s used to determine whether that individual user can receive new Marketing templates across all of their accounts. It seems to also depend on factors related to their engagement with that particular business.

In 2025, WhatsApp seem to have been shifting to much more granular enforcement of anti-spam, to the level of individual users and templates. On one level, this is much less punitive if you mess up; on the other, it can be more nuanced to figure out how to appropriately send messages to users. Spammy businesses will be caught at the template level before they can cause widespread harm, and many users won’t receive spammy messages at all if they haven’t engaged with the business.