Meta, Google, GA4, Shopify, and Finance are not designed to report the same number. Each has a different record of exposure, clicks, orders, revenue, timing, and business context. The goal is not forced agreement; it is an explainable difference and a usable next decision.
Why the numbers diverge
Ad platforms report activity they can associate with advertising exposure or clicks under their own attribution settings. GA4 records sessions and tracked events on the site. Shopify records orders in the commerce system. Finance records recognized business revenue under accounting rules. None of those systems is automatically wrong when the totals differ; they are answering different questions and observing different parts of the customer journey.
Timing is one common difference. A platform may credit an ad exposure or click within its attribution window. GA4 may record the visit under a different session rule. Finance may recognize revenue after a return window, fulfillment event, subscription renewal, tax adjustment, or accounting close. If two reports use different dates, revenue definitions, or order status, comparing their headline totals produces noise rather than insight.
Credit is another difference. Customers can see a Meta ad, search on Google, return directly, and purchase. Both platforms may have a valid record of their own interaction. Their reported conversions are not designed to sum to Shopify orders. They are evidence about a customer path, not mutually exclusive claims on the cash register.
A practical reconciliation sequence
1. Align the spend taxonomy
Make the channels mean the same thing everywhere. Separate Google Search, PMax, YouTube, and Demand Gen where the business needs that distinction; split Meta awareness and conversion campaigns if they support different decisions. The objective is not the most detailed naming convention. It is a stable, shared map of spend that works across the warehouse, platform reports, attribution, and MMM.
2. Compare revenue scope
Document what each system includes: web orders, phone sales, Amazon, retail, refunds, discounts, shipping, tax, subscription rebills, and canceled orders. An MMM may include more of the total business than MTA or GA4. That can be appropriate, but it means the two reports are not apples to apples.
3. Put promotions and seasonality on the timeline
A major sale can make click-based ROAS look spectacular because many people who were already considering the brand choose to buy. MMM can model a promotion or seasonal baseline explicitly. GA4 and platform reporting are still useful during the event, but they do not automatically separate demand created by the promotion from demand created by the channel that received the last click.
4. Treat subscriptions and returning customers carefully
February media might be associated with a March rebill in a time-series model, while a click-based system sees only the original February visit. If the aim is a direct comparison, filter recurring auto-bills or define them consistently. If new-customer acquisition is the decision, confirm that every system uses reliable new-versus-returning flags.
Which gaps are expected—and which deserve attention
It is normal for platform-reported conversions to exceed business orders when several channels claim overlapping paths. It is normal for GA4 to differ from Shopify because browser controls, consent choices, ad blockers, cross-device behavior, and checkout tracking affect what analytics can observe. It is normal for Finance to use a different revenue basis from a marketing report.
A gap deserves attention when it changes a material decision or signals a broken foundation: a channel is missing from the spend taxonomy; orders disappear at checkout; a promotion is not documented; a subscription revenue definition changed; or a large source of sales is included in one system but silently absent from another. Reconciliation is not a one-time cleanup. It is part of the measurement operating system.
Turn the remaining gap into a decision
After the definitions are aligned, do not expect the reports to collapse into one number. Ask what each system contributes. Platform and attribution data can offer fast clues about customer behavior. MMM can estimate broader portfolio contribution across business context. A lift study or geo test can check whether a particular intervention actually caused an incremental outcome.
For example, if a lower-funnel channel receives large last-click credit but appears less incremental in the broader evidence, the business may set a higher operating ROAS hurdle while a test is planned. If an upper-funnel channel shows weak click credit but a plausible halo, shift budget carefully and define what evidence would validate or reverse that move. The reports become useful when they help the team explain what it knows, what it infers, and what it will learn next.
FAQs
Which number should be the source of truth?
Use Finance or the commerce system for the business outcome you are trying to explain, with definitions documented. Use channel and analytics reports for the evidence they uniquely provide. A single report should not be asked to answer every question.
Why do Meta and Google conversions add up to more than Shopify orders?
They can both credit interactions on the same customer journey under their respective attribution rules. Their totals are not intended to be additive.
Can GA4 replace platform reporting?
No. GA4 provides an important site-level view, but it has its own coverage limitations and does not capture every platform exposure, customer interaction, or offsite transaction.