Click-to-WhatsApp ads: why Meta shows conversions you never received

There is a pitch going around at the moment. Email is dead, it only gets twenty percent open rates, move your budget to WhatsApp where people actually read messages. It usually arrives with a screenshot of a Meta dashboard showing a satisfying column of conversions at a very low cost.

I want to take that seriously rather than dismiss it, because click-to-WhatsApp advertising is a legitimate channel that works well for certain businesses. But two things in that pitch need examining before anyone moves budget: what Meta is actually counting when it reports a WhatsApp conversion, and whether the open rate comparison means what it appears to mean.

Neither holds up the way the pitch implies. Here is the detail.

What Meta counts when it reports a WhatsApp conversion

Start with the mechanic, because everything else follows from it.

In many campaign setups, particularly those using the basic WhatsApp Business App, Meta records a conversion at the moment someone taps your ad and WhatsApp opens on their device. Not when they send a message. The reason is technical rather than sinister: WhatsApp cannot tell Meta that a message was sent until after it has been sent, but Meta can see the tap that opened the app. So the tap is what gets counted.

Picture the actual sequence. Someone sees your ad while scrolling. They tap. WhatsApp opens with a pre-filled message waiting. They read it, hesitate, get distracted, or simply change their mind. They close the app. Meta's dashboard records a conversion. Your inbox records nothing.

There is a second distortion sitting underneath this one that almost nobody mentions. The metric Meta reports is called messaging conversations started, and it deduplicates on a rolling seven day window. If the same person messages you on Monday and again on Wednesday, that is one conversion, not two. So the number is simultaneously inflated by taps that never became messages and deflated by repeat contact from genuinely interested people. It is not a clean count of anything.

Why this costs you more than a misleading report

If the only consequence were an inflated number in a slide, this would be a reporting annoyance. It is worse than that, because the same signal that feeds your report also feeds the algorithm.

Meta optimises against the events you send it. If the only event it ever receives is "a conversation started," then it will get progressively better at finding people who start conversations. That is a different population from people who buy. Cheap taps from low-intent users look like success to a system that has never been told what success actually is.

So the cost compounds. You get a report that overstates performance, a budget allocated on that report, and an algorithm that is actively learning to find the wrong people. This is the same failure I described in the vanity metrics post, except here the bad metric is not just describing the campaign, it is steering it.

How to actually close the gap

The fix exists and it is well documented, which is exactly why its absence is telling when someone pitches you this channel.

When a user taps a click-to-WhatsApp ad, Meta generates a click identifier called ctwa_clid. When that user sends their first message, Meta includes the identifier in the webhook payload sent to your messaging platform. Your system stores it against the contact, and attaches it later when something meaningful happens: a qualified lead, a booking, a purchase. That event goes back to Meta through the Conversions API for business messaging, and the loop closes. Even if the purchase happens three days after the chat, Meta can tie it to the ad that started it.

1. Move to the WhatsApp Business Platform. The Cloud API is the technical foundation for any of this. The free Business App cannot support proper attribution or, as covered below, proper compliance.
2. Capture ctwa_clid from the webhook. No customer action is required, which is the advantage over the older workaround of matching on pre-filled reference codes. That method broke whenever someone deleted the pre-filled text, which happened constantly.
3. Define what a real conversion is for your business. Not "chat started." Qualified lead, appointment booked, appointment attended, quote sent, purchase completed. Pick the events that represent money.
4. Fire those events back through the Conversions API with the business messaging action source, passing value and currency where a sale is involved so Meta can optimise on revenue rather than volume.
5. Report on the real events, not the reported conversations. Once the loop is closed, the conversation count becomes a diagnostic for creative performance rather than a headline business metric.

The compliance part nobody includes in the pitch

This is where the risk stops being about wasted budget and starts being about legal exposure, particularly for anyone with contacts in Europe.

The short version: WhatsApp can be used lawfully for business messaging, but only through the Business API with real consent infrastructure behind it. The free Business App and the personal app fall short on the fundamentals, including double opt-in, handling data subject access requests, and controlled data hosting.

The trap in the middle of this

Meta's own platform policy has accepted a general marketing opt-in since late 2024, meaning a consent that does not specifically name WhatsApp can satisfy Meta's rules. GDPR does not work that way. It requires consent to be specific and informed, which means your privacy notice needs to disclose that you communicate over WhatsApp and that messages route through Meta's infrastructure.

The practical consequence: you can be fully compliant with Meta's platform policy and still be non-compliant under GDPR. Those are two separate tests and passing one tells you nothing about the other. If your consent flow never mentions WhatsApp, a generic opt-in may keep your account in good standing while failing as valid consent for anyone in the EEA.

And the clearest line of all: uploading phone numbers of people who have never interacted with your business, so they can be messaged on WhatsApp, has no lawful basis under any of the grounds available. There is no clever structure that makes that work.

Now the "email is dead" claim

This is the part that made me want to write the post, because the argument undermines itself from both ends.

The email side of the comparison is measuring the wrong thing. Reported open rates are inflated by privacy features that pre-load tracking pixels whether or not a message was actually opened. Roughly half of reported opens are affected. So when someone quotes twenty percent as proof email is finished, they are quoting a number that does not mean what they think, and which I would argue against using as a headline metric even in email's favour. I went into this in more depth in the deliverability post.

The WhatsApp side is borrowing credibility it did not earn. WhatsApp read rates are genuinely high, and the reason is that it is a personal, permissioned channel that historically carried very little commercial messaging. People read WhatsApp messages because WhatsApp messages have been worth reading. That is an attribute of the channel's current state, not a permanent property, and every business that treats it as a broadcast list spends a little of it.

We have seen this cycle before. Email open rates were extraordinary before commercial volume caught up. SMS the same. The read rate advantage of a new channel is a function of scarcity, and the pitch inviting you in is itself the mechanism that ends it.

The more useful framing is that they do different jobs, and one difference matters more than the rest:

Anyone advising you to abandon an owned channel for a rented one, on the strength of a metric comparison that does not hold, is not giving you strategic advice. They are selling what they happen to sell.

When click-to-WhatsApp genuinely works

Having spent most of this post on the caveats, the honest other side: this channel is very good in specific conditions, and I would recommend it for the right business.

• Conversation is the shortest path to the sale. High-consideration purchases where the buyer has questions that a landing page cannot answer well.
• The market already lives on WhatsApp. In much of Latin America, Southern Europe, India, and the Middle East it is the default business channel. Forcing those buyers into a form adds friction rather than removing it.
• The product needs configuring or quoting. Anything where price depends on specifics benefits enormously from a back-and-forth.
• Somebody is actually there to reply. Same principle as any lead source. Speed of response decides the outcome, and an unstaffed WhatsApp inbox is no better than an unworked CRM.
• You have built the measurement. With CAPI configured and real events flowing back, the channel becomes accountable and can be scaled with confidence.

It works poorly as a cheap substitute for a landing page, as a way to make cost per conversion look lower than it is, or as a replacement for a channel you own.

Five questions to ask anyone selling you WhatsApp conversions

Six mistakes I see repeatedly

The bigger point

Every few years a channel arrives with better engagement numbers than the incumbent, and the pitch is always that the old thing is dead. The engagement numbers are usually real. What gets left out is that they are temporary, that the measurement is immature, and that the reported metric is often counting something adjacent to the thing you care about.

Click-to-WhatsApp deserves a place in plenty of media plans. It deserves it after you have checked what the dashboard is counting, built the events that tell Meta what a customer looks like, and confirmed your consent actually covers the channel. Do that and it can be excellent.

Skip it and you are buying taps at a very attractive cost per tap.