Your WhatsApp campaign gets messages but no paying customers because, by default, a click-to-WhatsApp ad is optimized to start conversations, and that is exactly what it delivers. Meta finds people who are likely to tap and type "hi", not people who are likely to pay. Unless the offer filters people before the chat, someone answers fast and qualifies them, and the sale is reported back to the campaign, the system keeps buying cheap conversations and has no way to learn which of them were worth anything.
The fix is rarely one setting. It is usually four or five small gaps between the ad and the sale, and each one leaks. Below is how we diagnose them, an illustrative calculation of where the money goes, and a checklist you can run this week.
What the campaign is actually optimizing for
A campaign with a messaging objective and a "maximize conversations" goal tells Meta that a started chat is success. The delivery system does what it is told: it looks for people who start chats cheaply. Some of them are buyers. Many are curious, price-checking, outside your delivery area, or tapping by accident.
This is not a fault in the platform. It is a mismatch between the event you optimize for and the result you want. The same pattern shows up with lead forms, where cheap leads are not the same as qualified ones; we cover that side in leads vs. qualified leads.
The way out is to give the campaign a better signal, and to make the conversation itself do the filtering. Both require work outside Ads Manager.
The diagnosis, step by step
1. Is the offer clear before the chat starts?
If the ad shows no price, no product and no condition, people message to ask "how much?". Many leave as soon as they hear the answer. That is a conversation you paid for that the ad could have answered for free.
Show the price, or a price range, and the main condition (delivery area, minimum order, booking lead time) in the ad itself. Your cost per conversation will usually rise. That is fine: you are paying for fewer, better chats.
2. How fast does someone reply, and who is it?
A person who messages a business expects an answer in minutes, not hours. When the first human reply arrives the next morning, the buyer has often already ordered somewhere else.
Check three things: the median time to first reply during working hours, what happens outside working hours, and who answers. An automated greeting buys time but does not sell. The person who answers needs to know prices, stock and delivery terms without asking a colleague.
3. Is there a script with qualification questions?
Without a script, every chat is improvised. Write down the two or three questions that separate buyers from browsers in your business: what they need, where they are, when they want it, roughly what budget. Ask them early and politely. Then give a clear next step: a link to order, a booking time, a payment method.
4. Is there a follow-up?
Many chats stop after the price is sent. A single polite follow-up, the next day, with a reason to come back (a delivery slot, an answer to an objection) recovers some of them. Two or three unanswered follow-ups is pestering. One is service. Follow WhatsApp's own rules on when and how businesses may message people after a conversation ends.
5. Can you trace a chat back to the ad, and forward to a sale?
This is where most campaigns go blind. If a chat becomes an order by phone, at the counter or through a website, nobody records which ad started it. The campaign report shows conversations; your till shows sales; nothing connects them.
There are two levels of fixing this:
- Manual, immediately. Label chats in WhatsApp Business by stage (new, qualified, ordered) and record the campaign or ad they came from in a simple CRM or spreadsheet. Even rough labels let you compare ads by chats that became sales, not by chat count.
- Automated, with the API. When you use the WhatsApp Business Platform, a chat that starts from a click-to-WhatsApp ad carries a click identifier, ctwa_clid, in the data of the incoming message. Stored against the conversation, it lets you send events such as a qualified lead or a purchase back to Meta through the Conversions API for business messaging. That is how the campaign learns which conversations mattered. The plain WhatsApp Business phone app does not give you this; it needs the API, either directly or through a business solution provider.
6. Are you optimizing on messages when you could optimize on outcomes?
Once real outcome events flow back, you can move the campaign from optimizing for conversations toward optimizing for those events, where the options in your account allow it. Until then, you can still judge ads on your own outcome data and shift budget manually toward the ads that produce sales.
7. Is the audience sending junk?
Look at where the chats come from. Wide geography, placements that attract accidental taps, or an audience outside your delivery or service area will fill the inbox with people who cannot buy. Tighten the location to where you actually sell, and check the language and placement breakdowns for chats that never answer back.
A worked illustration of where the money goes
The figures below are illustrative, chosen to show the mechanics. They are not results from any client account.
Say a campaign spends $300 a month at $1.50 per conversation:
- Conversations started: $300 ÷ $1.50 = 200
- Share who answer the business's first reply: 60%, so 120
- Share of those who pass basic qualification (right area, real need): 25%, so 30
- Share of qualified chats that buy: 20%, so 6 sales
Cost per sale: $300 ÷ 6 = $50. The campaign report shows a cheap $1.50 per conversation and looks healthy. The business sees six customers and a $50 acquisition cost.
Now change the offer and the handling, not the budget. Price shown in the ad, a replying person within minutes, a short script:
- Cost per conversation rises to $2.50, so the same $300 buys 120 conversations
- Reply rate improves to 75%, so 90
- Qualification rate improves to 45%, because people already know the price, so about 40
- Close rate improves to 30% with a script and one follow-up, so 12 sales
Cost per sale: $300 ÷ 12 = $25. Fewer conversations, a higher cost per conversation, and half the cost per customer. If the campaign only reports conversations, the second version looks worse. That is why the outcome has to be measured.
Use this as a template with your own numbers. The step with the biggest drop is the one to fix first.
What we have learned running messaging campaigns
Two practices we apply on accounts where chats are the main channel:
- Every chat is connected to the campaign it came from. That lets us compare ads by chats that become visits or sales, rather than by chat volume. Ads that produce many conversations and few customers are cut even when their cost per conversation looks attractive.
- Only qualified leads go back to Meta as a signal. We never send every chat as a conversion. Reporting junk as success teaches the delivery system to find more junk. A small number of honest events is a better signal than a large number of noisy ones.
The same principle applies to lead generation in other forms; see why cheap solar leads often do not convert for the lead-form version of this problem.
Checklist
- The ad shows the price or price range and the main condition.
- Location targeting matches where you actually sell or deliver.
- Median first reply in working hours is minutes, not hours.
- Out-of-hours chats get an honest automated reply with a time when a person will answer.
- The person replying can quote prices, stock and terms without escalating.
- A written script includes two or three qualification questions and a clear next step.
- Non-responders get one polite follow-up within the rules.
- Every chat is labelled by stage and by the ad or campaign it came from.
- Sales are matched to chats, whether they close in WhatsApp, by phone or on the website.
- If you use the WhatsApp Business Platform, ctwa_clid is stored and qualified leads or purchases are sent back via the Conversions API for business messaging.
- Ads are compared on cost per qualified chat and cost per sale, not cost per conversation.
- Only qualified outcomes are reported back to Meta.
If the chat leads to a website to order, the page has to do its part too; our notes on landing page anatomy cover that step.
How we work
We start by looking at the full path: the ad, the first reply, the script, where sales close, and what is reported back to Meta. Usually the first changes are outside the ad account: the offer in the creative, response handling and tracking from chat to sale. Then the campaign is restructured around outcomes, as described on our Meta Ads service page. We do not promise a cost per sale before seeing how your chats convert today.
If your inbox is full and your sales are not, apply here with your monthly spend, rough chat volume and how sales are recorded. We reply in writing with what we would change first.
Frequently Asked Questions
Why are my click-to-WhatsApp conversations so cheap but nobody buys?
Because the campaign is optimized to start conversations, and cheap conversations often come from people who are curious or cannot buy. Showing the price in the ad, tightening location and replying quickly usually raise the cost per conversation and lower the cost per customer.
What is ctwa_clid?
It is a click identifier attached to a conversation that starts from a click-to-WhatsApp ad when you use the WhatsApp Business Platform. Storing it lets you match later outcomes, such as a qualified lead or a purchase, to the ad and send them back to Meta through the Conversions API for business messaging.
Can I track WhatsApp sales without the API?
Roughly, yes. Label each chat with its stage and the ad it came from, and record sales against those chats in a CRM or spreadsheet. It does not feed Meta's optimization automatically, but it lets you compare ads by sales instead of by chat count.
Should I send every WhatsApp conversation to Meta as a lead?
No. Reporting every chat as a conversion teaches the delivery system to find more of the same, including people who never buy. Send only qualified leads or purchases, even if the numbers are small.
How quickly should a business reply to WhatsApp ad messages?
Within minutes during working hours. Slow replies are one of the most common reasons paid chats go cold, because the person often contacts a competitor in the meantime.