Shoptet shows one order count, Google Analytics a lower one, and the purchases reported by Google Ads, Meta and Sklik often add up to more orders than the store actually received. That rarely means one of the systems is lying. Each answers a different question, sees a different share of customers and credits a purchase by its own rules. The right response is not to hunt for one true number, but to name the question first and only then choose the source.
Each system answers a different question
Shoptet is the order book. Its statistics are built from orders actually received in the admin, and they can be filtered by order status and shown with or without VAT. Whether the customer accepted cookies makes no difference. Shoptet alone, however, cannot tell which channel earned the sale.
Google Analytics 4 measures behaviour on the website. Its revenue data depends on a tracking tag that has to load on the store's pages. Its strength is that it judges paid and organic channels under one setting, so they can be compared with each other. An order entered manually in the admin or through the point of sale is invisible to it.
Ad platforms measure mainly to steer their own campaigns. By default, Google Ads gives credit only to Google's paid channels. Meta credits a website purchase to its ad if it happened within a set number of days after someone saw or clicked it. Sklik counts a conversion for anyone who arrived from one of its ads in the previous 30 days. None of them considers what the other platforms did in the meantime.
Why the numbers drift apart
Not every purchase is measured
Since 1 January 2022, Czech websites may store cookies that are not strictly necessary only with the visitor's active consent. Shoptet's help centre states that when a customer declines analytics and marketing cookies, the tracking tag does not fire. Ad blockers, or a page closed before the tag loads, have the same effect.
Google's and Seznam's systems fill part of the gap with modelling. GA4 models the behaviour of users who decline consent only when the site uses advanced consent mode and has enough data – among other things, at least 1,000 events a day with consent denied for seven days. The Conversions column in Google Ads may include modelled conversions, and cross-device conversions are partly estimated. Since April 2022, Sklik has shown observed and modelled conversions together. An ad platform can therefore report more purchases than GA4 actually captured without either number being wrong.
The same purchase lands on a different day
Google Ads reports a conversion on the date of the click; GA4 reports it on the date of the purchase. Someone who clicks on Monday and buys on Wednesday counts towards Monday in Google Ads and Wednesday in GA4. To compare with store data, Google Ads offers columns by conversion time. Sklik counts a conversion on the day it happens.
Day boundaries and data freshness add to the gap. GA4 uses the property's time zone and Google Ads the account's time zone; if they differ, daily figures differ too. GA4 processing takes 24 to 48 hours and figures change in the meantime. Comparing yesterday often just measures what has not finished processing.
Each system claims credit differently
By default, GA4 looks back up to 90 days before a purchase for a channel touchpoint. Google Ads uses a conversion window of 1 to 90 days. Meta counts a purchase within 1 or 7 days of a link click, within one day of merely seeing an ad and, since March 2026, separately within one day of another interaction such as a share or save. Sklik recognises a visitor for 30 days after a visit from its ad.
A customer who saw an Instagram ad, came from Sklik a week earlier and finally bought after clicking a Google ad can therefore be one order in Shoptet, one in GA4 and a conversion in three ad accounts at once. Summing conversions across platforms does not describe the number of orders.
The amount does not contain the same things
For VAT-registered stores, Shoptet's built-in integration sends GA4 the order value excluding VAT, shipping and payment fees; shipping and tax are sent separately. Shoptet's own statistics can be shown with or without VAT. The conversion value in an ad platform is whatever the tracking tag sends; Sklik, for example, recommends passing the order's revenue. Before you compare revenue or calculate cost as a share of it, check that both sources use the same basis.
Cancellations and repeat conversions
In Shoptet, a cancellation is an order status. The ad platform, however, counted the purchase when the order was placed. Google Ads removes conversions for cancelled orders only if the store sends them on a regular schedule, and Shoptet has a dedicated setting for that. Google Ads can also count every conversion after a click or only one, while GA4 may record every occurrence of the event.
Choose the source by the question
| Question | Default source | Keep in mind |
|---|---|---|
| How much did we actually sell and earn? | Shoptet with the chosen statuses and VAT basis | It does not see which channel the purchase came from |
| Which channels bring purchases, compared with each other? | GA4 with one attribution setting | It misses visits without consent and purchases outside the website |
| How should budget and bids be managed within one platform? | That ad platform | Do not add its numbers to other platforms' |
| Is advertising paying off as a whole? | Spend across all platforms against revenue from Shoptet | It does not assign credit to individual campaigns |
| Is measurement working? | The ratio of GA4 and ad platforms to Shoptet web orders over time | The change matters, not the absolute level |
One question, one default source. The other systems act as a check, and their gap needs to be explained, not eliminated.
Watch the gap, not a single number
Korzaro compares Shoptet with GA4 and your ad accounts and flags when measurement drifts from your usual level.
When a gap is a real fault
A gap that looks similar week after week is usually a property of measurement. A sudden change is what deserves suspicion: GA4 suddenly captures a noticeably smaller share of web orders than usual, or an ad platform stops reporting purchases while orders in Shoptet keep coming in.
Breaks like these tend to have a concrete cause. Shoptet names a tracking tag installed more than once – for example through the admin, the HTML editor and Google Tag Manager at the same time – as the most common reason for missing purchases in GA4. Google Ads needs the site's redirects to keep the click identifier intact. A new cookie banner can cause a similar jump.
For the comparison to mean anything, compare like with like: website orders only, the same VAT basis, closed and fully processed days, and Google Ads figures by conversion time. Every store's setup, range and share of customers who decline consent is different. That is why there is no generally correct size of gap – the benchmark is your own usual level.
What to do now
- Name the questions. For each report you read regularly, write down what it is supposed to answer.
- Assign a source and a definition. Give each question one default source and note which orders it includes, whether it counts VAT and which day it uses.
- Compare on the same basis. Website orders from Shoptet, the same VAT treatment, closed days and, in Google Ads, columns by conversion time.
- Do not add up platform conversions. For the overall picture, set spend across all platforms against revenue from Shoptet.
- Track the ratio over time. Compare the share of purchases captured by GA4 and by ad platforms week by week and investigate any sudden change.
Korzaro places Shoptet orders next to GA4 and ad-account data and alerts you when the share of measured website orders moves away from the store's usual level. It does not decide which source to trust for you, but it shows when the gap has changed enough to be worth checking.