The situation
The account looked like it was working. Reported return on ad spend was comfortably above target, spend had been increased twice on the strength of it, and the monthly reporting pack was reassuring throughout.
The finance team disagreed. Revenue attributed to paid search was roughly double what the order system recorded for the same channel over the same period.
What we found
A tag manager migration two years earlier had never been completed. The original hard-coded conversion tag was still present on the confirmation template alongside the new container. Both fired on every order.
Because the inflation was consistent, nothing looked anomalous internally. Every campaign was overstated by the same factor, so relative comparisons still made sense and nobody had reason to question the absolute numbers.
What we changed
- Removed the legacy tag and consolidated collection into one container
- Rebuilt conversion actions with correct counting and values
- Fed refunds back so reported revenue reflected retained revenue
- Reset bid strategy targets to match the corrected, lower baseline
- Rebuilt reporting to reconcile against the order table monthly
What happened next
Reported performance fell sharply on the day of the fix, which needs saying out loud before it happens. Nothing about the business changed — only the measurement did. Targets were reset against the true baseline and budget decisions from that point were made on numbers the finance team recognised.
Add the approved figures here. Keep them specific, dated, and tied to a defined period, so anyone reading can tell exactly what is being claimed.