Case study · Audit & measurement · Cosmetics e-commerce
The account was not underperforming. It was under-reporting.
Reported conversions didn't match backend orders — the campaigns were optimising toward a signal that reflected roughly two thirds of actual revenue. Conversion capture went from 65% to approximately 90% of backend orders.
I took over an account where reported conversions did not match backend orders, which made the performance data unusable for decisions. The campaigns were optimising toward a signal that reflected roughly two thirds of actual revenue.
The instinct in this situation is to restructure campaigns. That treats the symptom. I audited the GA4 and platform-level setup first, found where the events broke — consent handling, duplicate tags, a purchase event firing on the wrong trigger — and rebuilt the measurement layer before touching campaign structure.
Conversion capture went from 65% to approximately 90% of backend orders. The interesting part came after: several campaigns the team had been ready to kill turned out to be profitable, and one 'top performer' was revealed to be claiming credit for demand it never created. When the numbers are real, the decisions make themselves.