How Do You Know Which Ad Brought a WhatsApp Conversation? A Practical Tracking Guide
Most sales in the Saudi market close on WhatsApp, and WhatsApp is a closed box to measurement tools. Here is what can actually be measured, how to set it up, and where tracking stops and manual logging starts.
The short answer
No tool reads WhatsApp conversations and attributes them to ads automatically. The answer is layers: track every WhatsApp button click on your site through Google Tag Manager, give each source its own ready message that reveals it in the first line, use UTMs on ad links, then log every conversation and its outcome in one sheet.
A business owner pays Google, Snapchat and Meta, and customers get in touch on WhatsApp. At the end of the month they ask: which platform brought the customers? The usual answer is a guess, because the conversation started outside every tool that measures.
The problem isn't only technical. When you don't know the source, you set budgets by feel. You might pause a campaign that was bringing your best customers because its dashboard says "zero conversions", while its customers were messaging you on WhatsApp all along.
Why are WhatsApp conversations hard to track?
Because WhatsApp is an app separate from your site and from the ad platforms, and its conversations are encrypted and out of reach of analytics tools. What can be measured is what happens before WhatsApp opens (the button click and the visit's source) and what the customer writes in the first message. What comes after, whether they bought and for how much, is only known if you record it.
So there's no single fix. There are layers, each covering part of the picture, and together they give you something reliable enough to base budget decisions on.
How do I track clicks on the WhatsApp button on my site?
The first layer is the easiest and the most important: every click on a WhatsApp link on your site is recorded as a GA4 event through Google Tag Manager. The link usually starts with wa.me or api.whatsapp.com, which is what the trigger picks up.
- 1Give each WhatsApp button an attribute that identifies it in the code, such as data-source="home-hero" for the top homepage button and data-source="service-bottom" for the button at the bottom of a service page.
- 2In GTM, create a Just Links click trigger that fires when the click URL contains wa.me or api.whatsapp.com.
- 3Create a variable that reads the data-source attribute from the clicked element.
- 4Create a GA4 event tag with a fixed name such as whatsapp_click, carrying the source and page URL as parameters.
- 5Test in preview mode, then mark the event as a key event in GA4, and import it as a conversion in Google Ads if you advertise there.
This is exactly what I use on my own site: every WhatsApp button carries its source, so I know which page and which spot on the page drives contact. And because GA4 knows where the visit came from, every click has both an ad source and a position on the page.
Why do I need a ready WhatsApp message for each source?
The click tells you someone opened WhatsApp. The ready message tells you where they came from when they send it. A WhatsApp link accepts pre-written text that appears in the message box, and customers usually send it as is.
| Source | Ready message | What the first line tells you |
|---|---|---|
| Google search ad to a service page | Hello, I saw your ad on Google and want to ask about the service | Channel and service |
| Snapchat campaign landing page | Hello, I saw your ad on Snapchat and want to know the details | Channel and campaign |
| Homepage | Hello, I'd like a consultation | Direct visit or regular search |
The rule: a different message for each source that matters to your budget decisions, worded naturally rather than like a code. Don't put tokens like "ref-123" in the message; customers may delete them or find them odd. An ordinary sentence works as long as it differs between sources.
And remember the customer may edit the message or delete it and type their own question. That's normal, which is why the message is one layer, not the whole solution.
How do UTM parameters help connect WhatsApp to a campaign?
UTM parameters are additions to the end of an ad link that tell GA4 the source, medium and campaign. When a visitor lands on your page through a link carrying utm_source=snapchat and utm_campaign=ramadan, for example, and then clicks the WhatsApp button, the click is attributed to that session and its source.
- Google Ads adds a click ID automatically when auto-tagging is on, so you usually don't need manual UTMs there.
- Snapchat, TikTok and Meta: add UTMs manually or with each platform's dynamic URL templates.
- Keep naming consistent: snapchat, not Snap one time and snapchat the next, or the data splits across reports.
- Never put UTMs on internal links between your own pages; that breaks the session and credits the wrong source.
What about click-to-WhatsApp ads on Meta?
With click-to-WhatsApp ads, the button in the ad opens the conversation directly without passing through your site. That removes a step for the customer, but it means GA4 and GTM see nothing. Measurement happens inside Meta itself: Ads Manager shows how many conversations each ad started.
A conversation started isn't a customer. Many are a price question followed by silence. So you need to finish the measurement yourself: how many conversations from this campaign became an order? And if your number is connected to a business account and the WhatsApp Business platform, you'll sometimes see which ad a conversation came from, which makes logging easier.
One project I worked on was a Meta account review for a massage centre aimed at lowering cost per conversation: CPM fell by 27% and click-through rate reached 7% in that project, and results vary by business and market. But those are ad numbers, not booking numbers. What happens after the conversation starts has to be recorded somewhere else before you know its value.
How do I log WhatsApp leads?
The last layer is manual, and there's no substitute for it. One sheet, simple enough that whoever answers WhatsApp keeps filling it in every day. These columns are enough for most businesses:
| Column | Example | Why |
|---|---|---|
| Date | 2026-10-13 | To compare periods |
| Source from first message | Google / Snapchat / direct / unknown | Links the conversation to a channel |
| Service or product | AC cleaning | Shows which offer brings orders |
| Status | Question only / quote sent / confirmed order | Separates conversations from customers |
| Value | Order value in SAR if confirmed | Shows the real return per channel |
After a month of steady logging, you can compare: how much you spent on each channel, how many confirmed orders came from it, and their value. That number matters more than anything in the platform dashboards. Later on, confirmed orders can be uploaded to the platforms as offline conversions so they learn from them, but that needs organised data first.
What can't be measured no matter what you do?
- A button click doesn't mean a message was sent. Some people open WhatsApp and back out.
- A customer who saved your number from an old ad and messages you two weeks later won't show up with a source.
- Someone who forwards your number to a friend, who then messages you directly, arrives with no source.
- Someone who sees the ad, then searches your name and messages you, will usually be credited to search, not to the ad that introduced you.
These limits don't cancel measurement. They just mean reading the numbers as a direction rather than absolute truth, and letting the "unknown" row in your sheet keep its natural size instead of spreading it across channels by guesswork.
When don't you need all this?
If you advertise on one platform only and get no conversations from other sources, that platform's own measurement plus a simple sheet may be enough. And if conversation volume is very low, asking the customer "how did you find us?" at the start of the chat is simpler than any technical setup.
If you have a site with WhatsApp buttons and don't know which campaign brings customers, message me on WhatsApp. I'll review what's currently installed and tell you which layer is missing.
Sources and references
Common questions
Can I tell whether the customer actually sent the message after clicking?
Not from your site. Tracking tools only see the button click. Comparing clicks with actual WhatsApp conversations over the same period gives you an idea of how many back out.
Should I use a different WhatsApp number for each platform?
You can, but it's costly to manage and splits your conversations. A different ready message per source gets most of the benefit with one number.
Should a WhatsApp click count as a conversion in Google Ads?
If WhatsApp is your main sales channel, yes as a starting point, because it's the closest signal available. But watch conversation quality; if clicks rise and orders don't, the campaign is learning from a weak signal and will need more precise conversions later.
Is a click-to-WhatsApp ad enough without a website?
It's enough to start conversations, and it suits some businesses. But you lose site measurement and a page that explains the offer, and all measurement then sits inside Meta and your manual sheet.
Want this applied to your business?
This is exactly what the Tracking & analytics service covers: written scope, your accounts in your name.