Key takeaways
- GA4 can record a click to a WhatsApp URL; it cannot infer whether a message was sent or a sale closed.
- Keep clicks, conversations, qualified leads and sales as separate stages.
- Test one outbound-click implementation before marking any event as a key event to avoid double counts.
A click on a WhatsApp button is a useful signal, but it is not a lead or a sale. Someone may open WhatsApp, close it, message the wrong number or ask a question that never turns into business. If every button tap is reported as revenue, marketing decisions become unreliable.
Here is a simple measurement setup for a Nepali business site using Google Analytics 4 (GA4) and a manual lead log. It extends the WhatsApp marketing guide with an exact reporting distinction.
Table of contents
Define the four stages before changing GA4
| Stage | What happened | Where to count it |
|---|---|---|
| WhatsApp click | A visitor clicked a link to wa.me or api.whatsapp.com | GA4 outbound click |
| Conversation | A message reached your business account | WhatsApp inbox or CRM |
| Qualified enquiry | The person has a real need, location and timeframe | CRM or lead worksheet |
| Sale | The customer paid or the booking was confirmed | Order or booking records |
Write down who owns each stage and how duplicate enquiries will be handled. For example, a customer may click the button twice before sending one message. A click-to-conversation rate is meaningful only if the two systems cover the same dates and you can identify the source with reasonable confidence.
Check whether GA4 already collects outbound clicks
GA4's enhanced measurement can collect outbound click events for links that take people away from your domain. Inspect your web data stream settings and use DebugView or Realtime while clicking a test https://wa.me/... link on your own site. The event should carry a destination such as link_url; check that it points to WhatsApp.
If a Google Tag Manager trigger already sends a custom WhatsApp event, inspect that too. Do not fire an enhanced-measurement click and an identical custom event, then add both together in a report. Choose one primary click definition and document it.
Build a focused click report
In GA4, create an Exploration filtered to the outbound click event and WhatsApp destinations, such as wa.me and any other URL form your buttons actually use. Break it down by landing page, device and source/medium. Check a single date range and the site's time zone.
If the interface does not expose the destination as a usable dimension, configure the required event parameter or register a custom dimension according to your GA4 setup. Validate with a controlled click from a tagged test visit before relying on the report. Treat whatsapp_click, if you create it, as a micro-conversion. Mark it as a key event only if your team understands that it means a click, not a qualified enquiry.
Connect the inbox to a lead log
For each genuine conversation, record the first contact date, customer reference, topic, location, source if known, qualification status and outcome. A small spreadsheet is enough at the start. Ask the team to use one definition of “qualified” and one way to mark a sale. Avoid copying private message contents into analytics tools.
Illustrative example, not a client result: a monthly report might show 70 WhatsApp link clicks in GA4, 30 new inbox conversations, 12 qualified enquiries and 4 paid jobs. The four numbers answer different questions. They should not be called 70 sales, and they cannot be divided into precise attribution rates unless the records are matched reliably.
Review the SEO measurement guide for a wider dashboard. When you change a button, location or campaign, note the date and compare the same stages over a complete later period.
Run a controlled test before reporting
Use a private browser window, open a tagged test landing page and click the WhatsApp link once. Check that exactly one chosen click event appears in GA4 with the expected page and destination. Repeat on mobile, where an app handoff may behave differently from desktop. Then send a real test message from a separate phone and confirm the inbox receives it. The two observations validate different stages.
For the first month, reconcile the click report with a simple daily inbox count. Note missing messages, repeat customers and contacts who came from a saved phone number rather than the website. If attribution is unknown, label it unknown instead of assigning it to SEO. A useful report can show uncertainty openly: “42 website link clicks, 19 new WhatsApp conversations, source known for 11.” That tells the team what is measured and what still needs work.
Sources and further reading
Want a lead report that separates interest from actual sales?
Get in touch