Home / Blog / Conversions vs Store Orders
How-to · 2026

Why Your Conversions Do Not Match Shopify

Add up three platforms and you have more purchases than your store recorded. Nobody is lying, the gap is structural, and the tool that quietly reports one blended total is the one to distrust.

By · Published August 20, 2026 · 8 min read

Add up the conversions your platforms report and you will usually find they claim more purchases than your store recorded. This is not a tracking bug, it is not a tool to be fixed, and the number of purchases has not changed. Every platform is counting something slightly different and calling it the same word.

Each platform counts what it thinks it caused

A conversion figure is a claim: this purchase happened, and we contributed. That framing explains almost everything that follows.

A customer sees a LinkedIn ad on Tuesday, clicks a Google ad on Thursday, buys on Friday. LinkedIn claims it. Google claims it. Your store recorded one order. Nobody is wrong within their own frame, and summing the three claims produces a total that describes nothing that happened.

Overlap is not a data-quality issue to be cleaned. It is the arithmetic consequence of asking three parties which purchases they caused, and it does not go away with better tracking.

The four causes, in order of size

1. View-through conversions

A purchase credited to an ad that was displayed and never clicked. The argument is that exposure contributed, which is sometimes true and cannot be checked from the data. It is enabled by default in some places, it inflates the claim substantially, and turning it off is usually the single largest step in closing a large gap.

2. Attribution windows

How long after seeing or clicking an ad a purchase still counts. Windows differ by platform and by setting — a longer window claims more purchases, straightforwardly. Two platforms with different windows will disagree about the same customer, and both will be behaving as configured.

This is worth checking before any cross-platform comparison, because it is a configuration difference that presents as a performance difference.

3. The date a conversion is filed under

Click date or conversion date. A click on 29 January leading to a purchase on 2 February lands in January under one convention and February under the other, and Google publishes both as separate report variants. Monthly totals built from mixed conventions will not reconcile against anything, and the reason is not visible in the file.

4. What counts as an order on your side

Pending payments, cancellations, test orders, orders that were later refunded. Your store total includes or excludes these according to how you built it, and the platform counted whatever fired the conversion event — usually at checkout, before any of that was known.

This one is entirely within your control, and it is worth settling first because it is the only one of the four that you can make exact.

From being on both sides of this conversation

The gap gets discovered in the same way every time: someone adds up the platform numbers for a board pack, notices the total exceeds actual orders, and concludes the tracking is broken. It is usually not broken. It is doing exactly what it was configured to do, and nobody had told the person asking what a conversion figure is a claim about.

What ends the conversation productively is putting the store total and each platform's claim side by side in one table, with the windows and the view-through setting written underneath. It looks like an admission and it functions as the opposite: it is the version people can reason about, and it moves the discussion from “which number is right” to “what are we going to decide”.

Which number to use

For anything about the business, use the store. It is the record of what was actually bought and paid for. Revenue, orders, margin and growth all come from there.

For comparing campaigns within one platform, use that platform's conversions. Both sides of the comparison are counted by the same rules, so the comparison is valid even though the absolute figure is a claim. This is what platform conversion data is genuinely good for.

For deciding where budget goes across platforms, be careful and be explicit. The overlap is unknown and not symmetric — a platform that runs broad awareness campaigns will overclaim more than one running branded search. Comparing them on claimed conversions systematically favours whichever platform claims most freely.

Getting the gap smaller, honestly

Three things genuinely help:

  1. Turn off view-through where the decision is a budget decision. It removes the least falsifiable part of the claim.
  2. Align attribution windows across platforms, so at least the comparison is between like settings.
  3. Record the source on the order. A UTM or click id captured at checkout gives you your own attribution from your own record, which is the only version you control. It is not perfect — last-click attribution has well-known flaws — but it is consistent, and it makes joining store and ad exports possible at all.

What does not help is any amount of processing on top of the exports. The information needed to deduplicate is not in them.

What we compute, and what we refuse to

Quiriz for paid media computes spend, impressions, clicks, conversions, conversion value and every rate built from them, per campaign, platform, objective or account. Those are your export's figures, aggregated correctly.

We do not deduplicate conversions across platforms, and we will not present a blended conversion total as though the overlap had been resolved. It cannot be resolved from exports, and a single tidy number would be the most misleading thing we could show you. Per-platform figures, side by side, with your store total from the store data — that is the honest table, and it is the one that supports a decision.

The same principle governs the store side: Quiriz for e-commerce reports ad spend that matches no order as an explicit Unallocated row rather than spreading it to make the table look complete.

If your gap is large, check view-through first. It is one setting, it is on by default in more places than people expect, and it usually accounts for most of the difference in a single step.

See what each platform is actually claiming

Upload your ad exports and ask for conversions and spend by platform, side by side. Free to start.

Try Quiriz free →

Frequently asked questions

Why do my Google Ads conversions not match my Shopify orders?
Several reasons at once, and none of them is a bug. The platform counts a purchase it believes it caused within its attribution window, which may include people who saw the ad and did not click. It may file the conversion under the click date rather than the purchase date. And your store total may include or exclude pending, cancelled or test orders differently from what the platform counted. Any one of these produces a gap on its own.
Can I add conversions from Google, Meta and LinkedIn together?
No. Each platform reports the conversions it believes it caused, and a customer who saw a LinkedIn ad, clicked a Google ad and bought is a single purchase that two or three platforms will each claim. The sum is larger than the number of purchases that happened, by an amount that is not in any export. Compare platforms against each other; do not add them.
What is a view-through conversion?
A purchase credited to an ad that was displayed but never clicked. The reasoning is that exposure contributed, which is sometimes true and unfalsifiable from the data. It is the largest single source of overclaim, it is on by default in some places, and switching it off usually explains most of a large gap in one step.
Which number should I actually use?
The store total, for anything about the business. It is the record of what was actually bought. Platform conversions are useful for comparing campaigns within one platform, where the same counting rules apply on both sides of the comparison, and are not a substitute for revenue. If a decision depends on total sales, take it from the store.
Can a reporting tool reconcile the gap for me?
Not from exports. Deduplicating conversions across platforms requires knowing that two claimed conversions are the same purchase, and no export contains an identifier that would establish it. A tool that presents one blended conversion total is not resolving the problem, it is hiding it — the honest treatment is per-platform figures and the store total shown side by side.