Why Your Analytics and Your Bank Statement Disagree

Analytics & Growth Intelligence · Sample Article

It is one of the most common conversations we have. The ad platforms report one number, the analytics property reports another, and the finance team reports a third. Everyone assumes someone else is wrong, and the meeting stalls.

Usually nobody is lying. The three systems are answering three different questions.

They count different events

An ad platform counts a conversion it believes it caused. An analytics property counts a session in which an event fired. An accounting system counts money received. A trial that never converts, an order that is later refunded, and a lead that never becomes a customer will each be treated differently by the three. Before reconciling totals, agree what each system is supposed to be counting.

They attribute differently

Platform reporting attributes to itself, within its own attribution window, using its own model. Two platforms running at once will each claim the same conversion, which is why the sum of platform-reported conversions frequently exceeds the actual order count. This is not a bug; it is what the reports are designed to do.

Events fire more than once, or not at all

Duplicate tags, a thank-you page that can be refreshed, a tag firing on page load rather than on form submission, consent banners blocking the first pageview — each of these produces a specific, recognisable distortion. They are ordinary implementation faults and they are fixable.

The tracking is client-side only

Browser-based tracking is subject to blockers, privacy settings and network conditions. Server-side event tracking closes part of that gap by sending the event from your own infrastructure rather than the visitor’s browser. It is not a cure-all, but it materially improves the reliability of what you record.

What to do about it

Pick one system as the source of truth for money — it is almost always the one closest to the bank. Use the platforms for optimisation signal, not for financial reporting. Then document the event taxonomy so everyone knows what each number means, and check it on a schedule rather than only when something looks wrong.

That documentation step is the one most often skipped and the one that saves the most time later.

OajEx builds analytics and tracking that can be audited, and we reconcile it against what actually happened before we optimise anything on top of it. If your numbers are not agreeing, book a strategy call.