How to check if your ad tracking is working

Compare a closed week of paid orders with Events Manager and Google Ads. See where to start and when the gap is still a tracking problem.

To check whether ad tracking is working, compare a closed week of paid store orders with what Meta Events Manager received and with the "Conversions (by conv. time)" column in Google Ads. The goal is to explain most of the gap, not to zero it out.

TL;DR

  • Use a closed week and the same time zone across dashboards. Avoid a "yesterday" window, because older conversions are still arriving.
  • In the store, start from paid orders. Set pending, cancelled and refunded orders aside.
  • In Meta, check Events Manager first (what arrived) and only then Ads Manager (what was credited to an ad).
  • In Google Ads, compare against "Conversions (by conv. time)" when the target is the order date.

Why not compare "yesterday"?

Recent days still change. Google Ads help asks you to wait 24 to 48 hours before comparing results with Analytics, and notes that after significant changes, conversion data can take up to 7 days to appear in full. See Data discrepancies: Factors and troubleshooting.

In Meta, Ads Manager's "yesterday" and the store's "yesterday" are rarely the same slice: the ads report uses the impression date, not necessarily the day of payment, as How Conversions Are Attributed in Commerce Manager describes. A week that has already closed reduces that noise.

Note the store's time zone, the Meta account's time zone and the Google Ads account's time zone. If one is in one zone and another is elsewhere, the same order lands on different days.

What should I check in the store?

The store is the record of what arrived at checkout. Before you open the ad dashboards, decide which slice you use for cash: paid orders, or the net report after cancellations and refunds.

A delayed payment can create the order before the money arrives. If you compare "created orders" with Purchase or with a conversion, the gap already starts there. Separate created, awaiting payment and paid. Note cancellations and refunds separately.

That store total is the reference. The ad dashboards do not replace this list.

What should I check in Meta Events Manager?

Events Manager shows the events Meta received, both attributed and unattributed. Ads Manager shows the events credited to someone who was shown your ad, after deduplicating and applying rules. Meta itself explains this in Differences Between Event Counts in Meta Ads Manager, Ads Reporting and Events Manager.

Start with Events Manager, for the same period and time zone as the store. If the number of purchases Meta received is already far from paid orders, the problem starts with the event being sent: the tag does not fire, a blocker stops the pixel, the thank-you page does not load, or the same order arrives more than once. If Events Manager is close to the store and Ads Manager is far from it, the gap is attribution: the window, view-through credit, the impression date. Why Meta Ads sales don't match your store covers those rules.

Meta also says conversion counts in Ads Reporting can differ from third-party tools, because the tools detect, define and calculate conversions in different ways. See About Conversion Count Differences Between Meta Ads Reporting and Third-Party Reporting Tools.

What should I check in Google Ads?

In the default report, the conversion is dated to the ad interaction. The discrepancies page already cited talks about the impression date; Understand conversions by time with new columns describes the "Conversions (by conv. time)" column, which records the conversion at the moment it happened. When the target is the order date in the store, use that column.

Note the window, the attribution model and the counting setting (every conversion or one per click). Check that the tag sends a unique transaction ID per order, as described in Use a transaction ID to minimize duplicate conversions.

Do not treat modeled conversions as paid orders. In About consent mode modeling, Google Ads describes this modeling as an estimate of conversions from people who declined cookies. It can appear in the conversions column and be used in bidding. The wider picture with Meta, GA4 and the store is in Why GA4, Meta, Google Ads and the store don't match.

A UTM helps label the visit, but it does not make the dashboards start counting the same thing. How to Use UTM Parameters for Meta Ads covers the labelling part. Server-side tracking can increase the events a platform receives when the browser fails; that does not turn the dashboard number into the order total. What Is Server-Side Tracking? explains the reach and the limits.

When is the gap still a tracking problem?

After you align the period, time zone, order status and attribution rule, a hole can still remain that you cannot explain: a Purchase event that does not fire, a blocked tag, a repeated transaction ID, a pixel that is missing from checkout.

That is different from an expected gap. Meta and Google Ads can credit the same order. Each one only sees its own ads. View-through conversions and modeled conversions also pull the dashboard away from the list of paid orders. In those cases, tracking can be working and the numbers still do not line up. What Is Ad Tracking? describes what tracking does and what it does not fix.

A useful check ends with two separate decisions: for cash, paid store orders and total media spend; to optimize inside a platform, that platform's metrics, always with the same setup across the campaigns you compare.

When does it help to see the numbers side by side?

When the question is "what did the store record, and what did each platform report in the same period?", opening several dashboards makes it easier to mix up the window, the time zone and the status.

Independent attribution tools bring those numbers into one place. Clakta is one of them: it shows, side by side, what Meta and Google Ads report and the orders the store confirmed. Revenue comes from those same store orders, with status; each order enters that total once; a pending delayed payment is not revenue yet; refunds and cancellations update the same order. The pixel and the Google tag stay in place: Clakta runs alongside them, does not send events to the platforms, does not change their reports and does not replace GA4 (What is Clakta?, How attribution works, Verify your setup). For stores, see the ecommerce page.

The store, Meta and Google Ads remain different numbers. In Clakta, they appear on the same screen.

Frequently asked questions

What is the first number I should check?

Paid store orders, in a closed week, with the time zone written down. Only then open the ad dashboards.

Why start with Events Manager, not Ads Manager?

Because Events Manager shows what Meta received. If that number is already far from the store, the problem starts with the event being sent. Ads Manager shows what was credited to an ad, with a window and attribution rules.

Can I add Meta and Google to see if tracking is working?

No. Both platforms can credit the same order. The sum usually exceeds the total of paid orders and does not prove that tracking failed.

Are UTMs enough to check tracking?

No. A UTM labels the visit. You still need the order, the payment status and a way to tie the two together.

Does Clakta replace this check in the Meta and Google dashboards?

No. Clakta brings together the numbers the platforms report and the store's orders, side by side, so you can compare them in one place. It does not send events to Meta or Google, does not change those reports and does not replace GA4.