Why Your Ad Platform, Email Platform, and Store Report Different Revenue

Different marketing systems can report different revenue without any of them measuring the same thing. A store records transactions. An advertising platform assigns credit under its rules. An email platform assigns credit under another set of rules.

The first step is to identify the question each report answers. Do not add attributed revenue across platforms and label the sum total sales.

Follow one hypothetical journey

A customer sees a video, clicks an ad, joins an email list, clicks a welcome message, and returns directly two days later to buy. Depending on the tools and settings, several systems may claim some or all of that order.

The order still happened once. Attribution is a model for allocating credit, not a second transaction ledger and not automatic proof of causation.

Compare the definitions

Check date range, timezone, currency, order status, taxes, shipping, discounts, refunds, and test orders. One report may recognize a sale on the order date while another organizes results around an interaction date.

Then inspect attribution windows and eligible interactions. Click-through, view-through, and open-based credit can produce different totals. Klaviyo documents configurable attribution windows and interaction exclusions, illustrating why settings matter. [Source: Klaviyo attribution guidance.]

Tracking consent, device changes, browser restrictions, and missing integrations can also affect observable journeys. A missing attribution signal does not mean the customer had no prior exposure.

Separate three reporting layers

Transaction reporting answers what was ordered, paid, refunded, and fulfilled. Use store and payment records, reconciled appropriately with accounting.

Channel reporting helps compare delivery, interaction, and credited outcomes within a defined model. Use it to diagnose campaigns without treating the credit as unique total revenue.

Business reporting combines actual revenue, costs, new customers, contribution, and cash. It helps determine whether the overall activity supports the business.

These layers should inform one another while retaining their definitions.

Reconcile a small sample

Choose a manageable set of orders and compare their identifiers, timestamps, values, refunds, and available marketing data. Look for duplicated purchase events, currency mistakes, missing cancellations, or multiple integrations reporting the same order.

Document the settings and the explanation for major differences. If a platform changes its reporting model, mark the date and avoid comparing old and new exports without accounting for the change.

A useful reconciliation table includes store order ID, transaction value, recognized net revenue, channel claims, and known limitations. Keep personal data out of shared analysis unless it is necessary and appropriately protected.

Use incrementality for the causal question

If the question is “Would these sales have happened without the campaign?”, attribution alone cannot answer it. A well-designed randomized test or another appropriate causal design can help, provided volume and assumptions support it.

For smaller stores, use consistent measurement and cautious interpretation. Do not claim that a before-and-after change isolates one channel when price, seasonality, inventory, and other campaigns changed too.

Choose one stable business dashboard and keep channel models as supporting views. The goal is a decision you can explain: what happened in the store, what the platforms credit, and what remains uncertain about the cause.