Blog / Tools & automation /

Checking GA4 key event setup after a form redesign

Imagine this. The new form is live. It looks great. You click submit, a little green tick appears, and everyone’s happy.

Then you open GA4 a week later and enquiries have dropped to zero.

Did people stop enquiring? Probably not. More likely the measurement broke underneath the redesign. Maybe the form now shows a confirmation message instead of loading a thank-you page. Maybe the button has a new identifier. If the event setup still expects the old behaviour, the report quietly stops counting the one action you care about.

In GA4, the language is key events. The practical job is to confirm that the intended user action is recorded, that the event means what the team thinks it means, and that the setup survives the new form journey.

Write down the expected journey first

Before inspecting reports, describe how a person completes the form. Which page contains it? What fields are required? What happens after submission: a thank-you page, an inline message, a modal or a redirect? Is there a validation step or a second confirmation? This sequence is the reference for testing.

Decide which actions matter. A successful submission may be the key event, while opening the form or clicking its button may be supporting events. Don’t treat every interaction as a business outcome. I’d rather track one honest completion than five clicks that might mean anything. A click can occur without a valid submission, and a page view can happen without a person reaching the confirmation state.

Record how the old implementation was measured, if known. A redesign may replace a browser navigation with an asynchronous response, change a form provider or alter the URL. Those details help explain why the existing trigger may no longer fire. If nobody can confirm the old design, label the assumption and test the new behaviour directly.

Inspect the implementation and event path

Check the form’s actual success behaviour in a browser. Submit valid test data in a safe environment, observe the network or page state as appropriate, and confirm what unique signal marks success. Avoid using a real customer’s personal information for a test. If the form sends data to an external service, understand whether a test mode is available and follow the organisation’s data-handling rules.

Review the analytics tag or event configuration. Identify its trigger, any conditions, the event name and parameters, and whether the key event is marked appropriately in the GA4 property. A trigger based only on a button click is fragile when the success state matters. When I check a form after a redesign, the success state is the first thing I look for. A trigger based on a URL may fail if the form now updates the page in place. The signal should correspond as closely as possible to the completed action.

Check for duplicate paths. If both a form provider and a site tag send a submission event, a single lead might be counted twice. If the new design leaves an old hidden form on the page, a broad selector may attach to both. Inspect the event stream while testing, rather than assuming the configuration is correct because an event with the expected name appears once.

Test positive and negative cases

Test a successful submission and confirm the event arrives with the intended name and relevant non-sensitive context. Then test a validation error and make sure that failed submission is not counted as success. If the form has multiple steps, check what happens when someone abandons midway. These cases reveal whether the event represents completion or merely interaction.

Use a debug or preview workflow before publishing changes to the live configuration. Confirm that consent settings and tag firing behaviour are respected. Do not add personal details such as names, email addresses or free-text answers as analytics parameters. Measurement should provide useful context without transmitting information the analytics setup should not collect.

After publication, make a controlled test on the live journey if appropriate, then check the real-time or debugging view and the processed reports after their normal delay. A missing event in a standard report immediately after testing does not necessarily indicate failure; processing and reporting views differ. Keep a note of the release time and test conditions so the team can interpret what it sees.

Document what changed

Record the event name, what qualifies as success, the trigger or implementation location, the key-event setting, and who owns future changes. Include the test steps and the date of verification. This makes the next redesign less dependent on someone remembering why a particular condition exists.

Review the event in the reports where the team actually uses it. Check that filters or comparisons have not excluded the new path and that the event name remains consistent with historical reporting. If the definition has materially changed, annotate that break in the reporting context rather than presenting the before-and-after series as directly comparable.

The redesign isn’t measured just because the form submits. Can you complete a test submission today and verify, from the success state through to the GA4 key event, that one real completion produces the intended measurement?

Finally, make the success definition visible to anyone changing the form. A short note in the implementation or handover document can prevent a future redesign from treating a button click as a completed enquiry. Include the test steps and the owner of the analytics configuration.

Keep a test submission out of normal reporting where possible, and record how it was identified. This lets analysts distinguish verification from genuine demand. If no test mode exists, agree on a safe procedure with the analytics owner before sending test data to the live property.

Make ownership explicit

A form can be owned by marketing while its analytics tag is managed by a developer or an agency. Put the event definition somewhere all of them can find it. Include the user action that counts as success, the event name, the implementation location and the person who approves changes. That short record helps prevent a later redesign from quietly changing the meaning of a report.

Also decide how to handle a form that has more than one successful outcome. A request for a callback and a newsletter subscription may share a visual component but represent different intentions. Keep the measurement specific enough that the team can distinguish those actions, and do not add personal form answers as event parameters. If there is no safe way to test a live conversion, ask the analytics owner to agree a test procedure first.

After the release, compare the new event path with the previous definition. If the event name or success condition changed, note the break in continuity so a chart does not imply a direct comparison that the implementation cannot support.

So here’s a question worth asking your team this week: who owns the definition of a successful form submission? If nobody answers straight away, start there.