Why Google Ads reports more conversions than your store

Google Ads says it drove far more conversions than my store recorded. Is it inflating, or am I reading the wrong number?

When Meta reports more purchases than your store, the gap is usually tens of percent and the explanation is a difference of clocks and windows. When Google reports more, the gap is often a multiple, and the explanation is usually simpler and less interesting than it looks: you are reading a number that was never a purchase count.

The Conversions column is a sum, not a count

Google Ads reports conversions as a total across every conversion action the account has configured and marked as primary. If the account tracks purchases, add-to-cart, newsletter signups, phone calls and a PDF download, all five arrive inside the same column. Nothing about the interface warns you, because from Google's point of view that is the correct behaviour: those are the outcomes the advertiser said they cared about, and Smart Bidding optimises toward their sum.

So the first question is never whether Google is inflating. It is which actions are inside that number. Segment the report by conversion action and the multiple usually resolves itself in one screen.

In a measurement published on this blog, one store's Google account reported 17,484 conversions in a 90-day window against 7,712 real orders, or 227%. That was deliberately excluded from the comparison rather than presented as over-counting, for exactly this reason. Publishing it as a finding would have been the biggest available error in that piece.

Google reports on the click date, your store on the order date

By default, a conversion in Google Ads is recorded against the day of the ad interaction that earned the credit, not the day the conversion happened. Search campaigns carry a 30-day click window by default, so a report for last week can contain sales that will not be placed for another three weeks, and cannot yet contain sales that will eventually be credited to it.

This has a consequence people rarely account for: a recent window is always incomplete on Google's side, and it keeps filling in for weeks. Comparing yesterday against yesterday, or even last week against last week, compares a settled number against one that is still moving. Google offers a conversion-date view in some reports, and switching to it is the only way to align the two on a daily basis.

Data-driven attribution splits one sale into pieces

Google's default model for most accounts distributes credit across the interactions in a path rather than giving all of it to the last click. One order can therefore appear as 0.42 of a conversion on a brand search campaign and 0.58 on a Shopping campaign. The decimals are the model working, not a bug.

The practical effect is that a per-campaign conversion count cannot be checked against a per-campaign order count, because the platform is not claiming a campaign produced whole orders. The account total is still comparable; the campaign rows are shares of it.

Modeled conversions are estimates, and Google labels them

Where consent was declined, or a journey crossed devices in a way Google could not observe directly, Google estimates the conversions it believes occurred and includes them. This is not hidden: the documentation describes it, and the behaviour is a deliberate response to the loss of observable signal.

An estimate can be accurate in aggregate and still be unauditable for any single sale, and both halves of that matter. You can reasonably budget against a modeled total. You cannot pull one order out of it and trace it, which means a modeled conversion can never be reconciled against a specific row in your store.

Impressions can earn credit too

Video campaigns can record engaged-view conversions, where someone watched a meaningful portion of an ad without clicking and converted within a few days. Display impressions can receive credit under data-driven attribution. Neither produces a visit, so neither will ever appear in your own analytics, and their absence there is not a tracking failure.

What to do instead

Filter to the one conversion action that fires on a completed purchase. Compare a whole window rather than individual days, and if you need day-level agreement, switch Google to the conversion date. Include or exclude refunds consistently on both sides, remembering that Google keeps a conversion unless someone uploads an adjustment. Then look at what is left.

What is left is the honest gap, and it is usually made of things that are real: view-through and engaged-view credit for people who never visited, modeled conversions for people you cannot observe, and cross-device journeys your own pixel could not stitch. None of those is a defect in Google's reporting, and none of them is verifiable from Google's reporting either.

The number that settles it

There is one comparison that does not depend on agreeing about models, windows or actions. A platform can legitimately resolve a buyer you never saw, but no platform can resolve an order that does not exist. Your store's settled order count for a window is a hard ceiling on what every platform combined can truthfully claim about it.

Total what each connected platform claims for the same window, on the purchase action only, and compare it against that ceiling. Under the ceiling, nothing is proven either way and the honest answer is that part of the claim is unverifiable. Over it, the excess is double-counting, and it is provable without knowing which platform caused it.