Why ad platform purchases don't match your Shopify orders

By the HonestTag team · Published October 27, 2026

Meta says one number. Google says another. Shopify says a third. Here's why.

Two different questions

Shopify answers "how many orders did the store take?" An ad platform answers "how many conversions does this platform credit to its ads?" Meta's help page on event counts says Ads Manager shows "actions taken by people who interact with your business as a result of being shown your ad." Google's page on conversion data says the Conversions column counts conversions across your primary conversion actions. Neither count starts from your order list. How purchases reach each platform in the first place is covered in the Meta Conversions API and Google Ads conversion tracking.

Why the platform count can be higher

• Each platform credits its own ads. Meta's count covers people shown a Meta ad. Google's count covers conversions after an interaction with a Google ad. A shopper who clicked a Meta ad on Monday and a Google ad on Thursday, then bought once, can be counted by both. Add the two platforms together and one order shows up twice.
• A view can count. Meta's attribution settings include view-through, which "Counts events that occurred within 1-day after an impression of your ad." Meta also says that if a person sees an ad and does not click, "but later browses your website and decides to purchase," Meta may attribute that purchase to the ad they saw. Google Ads has a view-through window too, 1 day by default according to Google's conversion window page. Google counts a view-through conversion when customers "view, but don't interact with your ad, and later complete a conversion on your site," and shows those only in its "View-through conversions" and "All conversions" columns, not in "Conversions."
• Modeled conversions. Google says its Conversions column "may include modeled conversions in cases where you aren't able to observe all conversions," and that cross-device conversions combine "observed and modeled conversions." A modeled conversion is an estimate, not an order you can open.
• Duplicate pixel fires. When Ads Reporting shows more conversions than your other tools, Meta's troubleshooting page says to check for duplicate pixel fires, "which can sometimes be caused by redirects." Two tools sending the same order is a separate case with its own page: duplicate purchase events.

Why it can be lower, or land on another day

• Ad blockers. Meta says "The Meta Pixel may not fire if someone is using an ad blocker while browsing your site. This can cause Ads Reporting to report fewer conversions as compared to your internal data."
• Windows. A sale outside the window is not counted. Meta's click-through setting counts events "within 1-day or 7-day after a link click on your ad." Google's click-through window is 30 days unless you change it, and Google says that with a 7-day window "any conversion that happens more than 7 days after the ad interaction won't be recorded." How long your own buyers take is their click-to-order lag.
• The reporting date. Google's primary conversion columns "are calculated based on the time of the click, not the time of the conversion," so a sale on Thursday from a Monday click is reported on Monday. Google offers "by conv. time" columns and says they help "when comparing your Google Ads data with third-party analytics like Google Analytics or Shopify." Meta's troubleshooting page says Ads Reporting lets you report "by time of conversion or time of exposure when the impression was served."
• Time zones. Meta's checklist asks you to "Check that both of your reporting tools are set to the same time zone." An ad account on one zone and a store on another put late-evening orders on different days.

Worked example

These numbers are made up to show the arithmetic. They are not from a store.

Over 14 days a store takes 100 orders in Shopify. Meta Ads Manager reports 60 purchases. Google Ads reports 50 purchase conversions. Added together, the platforms claim 110 purchases, 10 more than the store took. Nothing has to be broken for this to happen:

• 25 orders are counted by both platforms: the buyer clicked a Meta ad and a Google ad, and both windows include the purchase. No other order is counted by both.
• So the two platforms credit 60 + 50 − 25 = 85 different orders. The other 15 orders are ones neither platform credits.
• 6 of Meta's 60 are view-through: the buyer saw a Meta ad, did not click it, and bought within a day. Those 6 are not in Google's 50, so they sit inside Meta's 60 and outside the 25.

Every one of those 110 claims can follow the platform's own rules, and the store still took 100 orders.

In HonestTag's Mirror plan, Meta's row would put its 60 beside the new customers whose first recorded ad click was Meta, say 24. The other 36 of Meta's 60 are not invented sales: 6 are those view-throughs, 10 are orders whose first recorded ad click was Google, and 20 are returning buyers whose first recorded ad click was Meta. 6 + 10 + 20 = 36, and 24 + 20 + 10 + 6 = 60. What the row measures is the distance between what Meta reports and the new customers whose first recorded ad click was Meta.

Compare the two counts fairly

1. Use the same dates and the same time zone in the ad account and in Shopify.
2. Report by conversion date. In Google Ads, add "All conv. (by conv. time)." Google says that column may give a better comparison with third-party platforms or your own systems because it reports conversions by the time they occurred, rather than the time of the ad click. "Conversions (by conv. time)" leaves view-through out. Google also says differences "may still exist due to varying attribution models, lookback windows, and how each platform processes data." Give the newest days 48 hours to settle: Google says "reporting lags can take up to 24-48 hours." In Meta, report by time of conversion.
3. Know your attribution setting. Write down the click and view windows each platform used before you compare.
4. Never add platforms together and call it sales. The sum of every platform's claims is not an order count.
5. Rule out a second purchase source. A second tool sending the same order is a separate case. See duplicate purchase events.

What HonestTag compares

• HonestTag puts what Google Ads, Meta, Microsoft Advertising, TikTok and Criteo report beside the new customers whose first recorded ad click or ad-tagged link was for that platform, counted from your store's own orders, over the last 14 days (for Criteo, that count reads 0 for now). When the platform reported a conversion count, the row reads Above first-click, Undercounted or Matched, decided by a tolerance band you can inspect. While the days being compared start before that platform's click window has passed since install, the row reads Too early to judge instead, because the platform can credit sales to clicks from before HonestTag could record them.
• The Gap, on paid plans and the trial, also adds up every connected platform's claims and sets the total against your store's orders for the same days.
• Clicks only. HonestTag never learns which shopper saw an ad, only the daily impression totals the platforms report, so it gives no view-through credit. Each order is credited to one platform at most: the one behind its first recorded ad click or ad-tagged link. First-click attribution explains the rule, and attribution models sets it beside the others.

The honest trade-offs

• The Mirror plan counts new customers, not every purchase. A platform that mostly reaches returning buyers will read Above first-click even if every sale it claims was real. That is the question the Mirror plan asks: which ads bring new customers.
• HonestTag sees only the clicks it recorded. A click from before install is not on record, because HonestTag starts measuring on the day you install and never invents history. An explicit refusal blocks tracking, so that shopper's clicks are not collected. An order with no recorded click has no first ad click in HonestTag.
• No credit for views. If an ad works by being seen and the buyer arrives later by typing your address, HonestTag will not credit that ad. A platform's view-through count may be the only place that sale shows up. That count is credit for a view, and HonestTag cannot match it to an order because it never learns which shopper saw an ad.
• Neither number is the truth. Attribution is a rule for handing out credit. A holdout test is the way to measure what an ad caused.

Frequently asked questions