Why Meta, Google and Shopify revenue differs
Why the numbers never match, and how to tell an attribution gap from broken tracking.
By Mark Tran · Updated
Different totals do not automatically mean broken tracking
Meta, Google Ads and Shopify answer different measurement questions. Your store records orders; Meta and Google attribute eligible conversions to interactions with their own ads. Their totals can differ because of attribution rules, timing, data collection and revenue definitions. A tracking fault is another possibility, not the only explanation.
I reconcile the definitions before judging the difference. Requiring the Meta and Google totals to equal store orders can lead to changing tracking that was working as configured.
One sale can appear in two advertising reports
Consider an illustrative A$100 order. A shopper interacts with a Meta ad, later clicks a Google ad and then buys. Depending on each account's attribution settings, both may credit themselves with the purchase.
| Report | Revenue shown |
|---|---|
| Store order | A$100 |
| Meta attributed revenue | A$100 |
| Google attributed revenue | A$100 |
| Sum of Meta and Google claims | A$200 |
The business still received one A$100 order. This example illustrates why adding Meta and Google claims together is not a reliable measure of deduplicated revenue. It does not show that either Meta or Google had no influence on the purchase.
What I check before changing the tracking
- Reporting period and time zone. Compare the same interval. Check whether the report assigns a conversion to an interaction date or a conversion date.
- Attribution settings. Record the actual click, view and other eligible interaction windows in each account. Do not assume a default or compare different windows without saying so.
- Revenue definition. Check currency, tax, shipping, discounts, refunds and cancelled orders. A purchase value sent at checkout may differ from a later net-sales report.
- Event setup. Check that the purchase fires at the intended point, carries the right value and is not counted twice. Review browser and server-event deduplication where both are used.
- Collection gaps and delays. Consent choices, blockers and processing time can change the available data. Allow reports to settle before judging a recent period.
Shopify documents reasons analytics systems can disagree, while Google explains its reporting delays. Use those definitions alongside the settings actually present in your account.
Keep three views in the report
Store results: revenue on a defined basis, orders, refunds, customer mix and the costs needed to understand contribution. This anchors the discussion in what the business recorded.
Channel attribution: what Meta and Google each report for conversions and revenue, with the window and definitions attached. This can support campaign decisions without presenting the sum as additional store revenue.
Incrementality evidence: what a suitable experiment or other measurement design can tell you about sales that would not otherwise have happened. This is a different claim from attribution. Meta itself describes incrementality experiments as a stronger way to assess that question.
What a short account audit can establish
A short review can identify suspicious values, inconsistent settings, missing context and the next investigation. It cannot turn an attribution report into proof of incremental sales. Where the question needs a tracking investigation or an experiment, I say so.
My bedding case study's measurement notes distinguish the published Meta and Google figures and reporting windows. Once your own revenue definition is clear, use the break-even ROAS calculator to test the economics.