Monday morning, weekly report: Meta Ads Manager shows 48 conversions, Google Ads 41, your shop backend or CRM 52. Added up, Meta and Google would have delivered 89 conversions, although only 52 happened in total. Which number is right?
The honest answer: each number answers a different question. Part of the gap is normal and cannot be calculated away. Another part is a real tracking error that makes your campaigns optimise on wrong data. This article helps you tell the two apart.
The short version
- Your shop or CRM counts what actually happened: orders, enquiries, leads. Regardless of where they came from.
- Meta and Google Ads count what they credit to themselves. Each platform by its own rules, windows and estimates.
- The sum of all platforms is almost always higher than reality, because both can claim the same conversion.
- If one platform is far below your shop, even though it drives most of your traffic, that often points to signal loss or broken tracking.
Normal: reasons that lie in the measurement logic
These differences are by design. They don't go away even with perfect tracking.
1. Every platform credits the conversion to itself
A customer sees an ad on Instagram, searches for your brand on Google two days later, clicks the search ad and buys. Meta counts the conversion because the customer saw or clicked its ad. Google counts it too because the customer clicked the search ad. Your shop counts one order.
Neither platform knows what the customer did on the other one. That is why the sum of all platforms is not a meaningful number. Which channel deserves which share of the result is a question of attribution, and no ad platform answers it neutrally.
2. Different attribution windows
How long after a click a conversion is still credited to the ad differs by platform:
| Meta | Google Ads | |
|---|---|---|
| Default after click | 7 days | 30 days |
| Default after view without click | 1 day | 1 day (only certain campaign types, e.g. Display and Video) |
| Adjustable | Yes, in the ad set settings | Yes, per conversion action |
A purchase nine days after the click shows up in Google Ads, but not in Meta with default settings. Check which windows are actually set in your accounts before comparing numbers.
3. View-through conversions
By default, Meta also counts conversions from people who only saw an ad without clicking, if they convert within a day. Someone who was going to buy anyway and scrolled past your ad on Instagram creates a Meta conversion this way. In your shop, that purchase looks like a direct visit.
4. The date the conversion is counted on
Your shop counts an order on the day it comes in. In their standard reports, Meta and Google Ads assign the conversion to the day of the ad interaction instead, the day of the click or view. A purchase on Monday after a click on Friday appears on Monday in the shop and on Friday in the platform. In daily or weekly comparisons this creates apparent gaps that even out over longer periods.
Then there are time zones: if your ad account uses a different time zone than your shop, conversions around midnight move to the next or previous day.
5. Modelled and fractional conversions
Platforms don't only show measured conversions. Google Ads uses Consent Mode v2 to add conversions it could not measure directly because consent was missing, through modelling. Meta also estimates missing data statistically in some cases. These conversions don't exist in your shop as individual orders, they are a projection.
Google Ads also uses data-driven attribution: if several Google ads were involved in one conversion, Google splits it between them. That is why you sometimes see odd values like 40.7 conversions there. How much Google models also depends on whether you use Consent Mode in Basic or Advanced mode, which our article on Consent Mode Basic vs. Advanced explains.
6. Your shop counts things no platform sees
Your shop or CRM also contains orders from organic search, newsletters, returning customers, phone orders and referrals. Conversely, the platform usually keeps counting a purchase that is later cancelled or returned. When you compare revenue, also check gross vs. net, shipping and vouchers. Your shop may report the net value while the pixel sends the gross amount.
Not normal: reasons that point to tracking errors
You can and should fix these. They distort not only the report but also how your campaigns optimise.
7. The platform sees too little: signal loss
If Meta or Google shows clearly fewer conversions than actually come in through their ads, signals are getting lost on the way. Typical causes:
- Ad blockers block the pixel in the browser.
- Intelligent Tracking Prevention in Safari limits the lifetime of cookies and click IDs.
- Visitors decline in the cookie banner, and the platform may not measure them.
- The conversion page loads so slowly or redirects so quickly that the pixel never fires.
Server-side forwarding like the Meta Conversions API helps against the first two, because it does not depend on the browser. No technology helps against declined consent, only a better, honest banner.
8. The platform sees too much: duplicate events
If Meta reports more purchases than your shop has in total, something is counted twice. The most common reasons:
- Two pixels fire the same event, for example once from the theme and once from a plugin or Google Tag Manager.
- Pixel and Conversions API run without event deduplication. Both report the same purchase, but without a shared event ID Meta cannot merge them.
- The thank-you page is loaded several times, because customers reload it, bookmark it or come back from the confirmation email.
- Google Ads counts every instead of one conversion. For leads, a conversion action should count "One", for purchases "Every". If a lead form is set to "Every", a user who submits twice counts twice.
9. Wrong event, wrong page
A URL rule like "contains /thanks" also matches /thanks-newsletter and /thanks-voucher. A click event on "Buy now" also counts visitors who abandon the purchase afterwards. If conversions look too high, first check exactly what is defined as a conversion.
10. Test purchases and internal traffic
Test orders from the setup, visits from your own team and staging sites with the same script end up as real conversions in the platforms. In small accounts this can make up a large part of the gap.
How much difference is normal?
There is no fixed number, because it depends on your channel mix, your windows and your consent rate. These rules of thumb help:
| Observation | Likely cause |
|---|---|
| Sum of all platforms is above the shop | Normal, multiple attribution |
| A single platform is above the shop | Almost always an error: duplicate events, wrong rule, test purchases |
| Platform is slightly below the orders that came through its ads | Normal, windows, consent, attribution date |
| Platform is far below and the gap keeps growing | Signal loss or broken pixel |
| The gap jumps suddenly from one day to the next | Technical change: theme update, new plugin, banner change |
More important than the absolute gap is its stability. A constant difference can be priced in. A difference that suddenly jumps is almost always a technical problem.
How to reconcile the numbers in five steps
- Pick the same period and time zone. Compare at least 14 days so date shifts even out.
- Document the windows. Note which click and view windows are set in Meta and Google Ads.
- Compare each platform with the matching slice of your shop. Don't compare all orders with Meta, only those whose visitors came via Meta. You can tell by the UTM parameters or the fbclid on the first visit.
- Check for double counting. Search the page source for multiple pixels, check deduplication in Meta Events Manager, check the counting setting of your Google conversion actions.
- Exclude test purchases. Always test with clearly recognisable test data and mark it so you can remove it later.
How LeadSignal helps
LeadSignal does not replace attribution, but it removes many of the avoidable errors:
- One script instead of several pixels. Meta, Google Ads, TikTok and LinkedIn are fed from the same source. Duplicate pixels from theme, plugin and tag manager disappear.
- Deduplication built in. For Meta, pixel and Conversions API send the same event ID, and the same goes for LinkedIn's Insight Tag and Conversions API.
- Your own platform-independent view. The event log shows every conversion with the session timeline, UTM parameters and click IDs. The analytics view can be filtered by UTM source and shows your top campaigns. That lets you check how many conversions came in via Meta clicks before you judge the number in Ads Manager.
- Less signal loss. Meta, TikTok and LinkedIn also receive conversions server-side, independently of ad blockers in the browser.
If you also want to know which channel contributes which share of revenue across several touchpoints and weeks, and include deals from your CRM, that calls for a real attribution solution like our sister product LeadMetrics.
Conclusion
Different numbers in Meta, Google Ads and your shop are not a sign of broken tracking as long as they can be explained: multiple attribution, windows, views, date logic and modelling. Get suspicious when a single platform reports more than your shop, when it is far below, or when the gap suddenly changes.
The best protection is a lean setup without duplicate pixels and with server-side backup. Our guide to simple conversion tracking shows how that works without a developer.
