Note · Product analytics
Measure the decision, not every tap
The short answer
I prefer a small set of trustworthy signals to a dashboard full of unexplained activity. Before adding an event, write the decision it should inform. If nobody would change their mind after seeing the result, the measurement may be decoration. Analytics becomes useful when it connects something a person tried to do with whether they actually managed to do it.
A practical example
Suppose people open a booking form but rarely finish. Measure the start, relevant errors and confirmed booking. More button clicks could mean interest, confusion or repeated attempts. Watching someone use the form helps distinguish those explanations; the count alone cannot tell you which is true.
Verify what arrives
GA4 supports recommended and custom events. Choose meaningful names and verify their arrival with its debugging tools. Keep personal details out of event payloads. A click on Submit and a successfully stored booking should remain different events because they describe different outcomes.
Make one testable change
Try clarifying an ambiguous field, then compare completion alongside errors and support requests. Record when the change happened and whether traffic changed too. A before-and-after improvement is a useful clue, not automatic proof of causation. The next decision should reflect that uncertainty rather than hide it.
Sources to check
Keep reading