Meta, GA4 and Google Ads show three different purchase totals: which gaps are normal.
- “Meta, GA4 and Google Ads conversions don't match”
- “Facebook Ads, Google Ads and Google Analytics show different numbers”
- “Meta and Google Ads both claim the same sale”
- “Meta plus Google Ads conversions more than total orders”
- “reconcile Meta Ads, GA4 and Google Ads data”
- “which platform's purchase numbers are right”
Baris Asa, tracking consultant, London · Last checked: November 2026
Meta, GA4 and Google Ads show three different purchase totals because they count different things. GA4 counts each purchase on the site once and gives it to one channel. Meta and Google Ads each count only purchases they can tie to their own ads, on their own time windows, and both add estimated sales they couldn’t see. One buyer who clicked a Google ad and saw a Meta ad is a sale in both. So Meta plus Google Ads can be more than the store sold. That is normal.Reconciling the three against the store’s orders is tracking-consultant work: match orders to platform records, size each gap, call it normal or a fault. Synctag, a London tracking consultancy run by Baris Asa, does it order by order.
Three shapes are not normal. One platform alone at or above the store’s web orders. A gap that jumped on a date you can name. GA4 falling further below the store than your cookie banner explains. Each is on file below.
Quick check: pick a week that ended at least two days ago and write four numbers in a row.
- The store’s web orders (in Shopify or your platform; leave out POS, phone and marketplace orders).
- GA4: Reports, Engagement, Key events, the purchase row for the whole site.
- Meta: Ads Manager, the Website purchases column, across every ad account on the Pixel.
- Google Ads: Campaigns, segmented by conversion action, the purchase row.
If any of 2 to 4 is equal to or higher than 1, start at the faults. If all three sit below and hold steady week to week, the gaps are probably design.
This page is the three-way view. The two-way checks have their own notes: Meta Ads shows more purchases than GA4 and GA4 and Google Ads numbers don’t match.
What does each platform count as a purchase?
The store counts orders. Each platform counts something narrower, on its own rules. These are defaults; each account can change them.
| GA4 | Meta | Google Ads | |
|---|---|---|---|
| Which purchases | All on the site | After its own ads | After its own ads |
| Shares a sale with other platforms? | Yes, one channel or split | No | No |
| Click window | Up to 90 days | 7 days, link clicks | 30 days |
| Like, share, save | Not counted | 1 day | Not counted |
| Ad seen, no click | Not counted | 1 day | 1 day |
| Dated to | The purchase | Ad seen, or purchase | The ad click |
| Sales it can’t see | Modelled with Consent Mode | Modelled | Modelled |
Read the first two rows and the main gap explains itself. GA4 has one sale to hand out per order. Meta and Google Ads have one each.
Three footnotes for whoever runs the reports. Meta changed its definitions on 3 March 2026: click-through became link clicks only, and likes, shares and saves moved to a separate 1-day engage-through window; some accounts still report on the older ones. Google Ads keeps view-through conversions out of its Conversions column, except in some Demand Gen, App and Performance Max campaigns, and engaged views on Video, Display, App, Demand Gen and Performance Max campaigns count in Conversions too. GA4 can credit YouTube engaged views too, when Google Ads is linked.
Why can Meta and Google Ads together claim more sales than the store made?
Because neither can see the other’s ads. A buyer clicks a Google Shopping ad on Monday, scrolls past a Meta ad on Wednesday and buys on Thursday. Google Ads counts the sale: a click inside 30 days. Meta counts it too if the purchase falls inside its view window. GA4 gives it to one place, or splits the credit, depending on the report.
| SOURCE | PURCHASES | VS STORE |
|---|---|---|
| GA4, whole site | 912 | 91% |
| GA4, credited to Meta | 190 | 19% |
| GA4, credited to Google Ads | 352 | 35% |
| Meta Ads Manager | 318 | 32% |
| Google Ads, purchase | 405 | 41% |
Google Ads claims 405 against GA4’s 352: 15% above, inside Google’s “often up to 20%”. Meta claims 318 against 190, on wider rules. Together they claim 723 where GA4 gives them 542. Nothing reaches the store. That is a healthy week. At 1,100 combined it would still be design, as long as neither reached 1,000 alone.
Which gaps are normal?
After the like-for-like checks in the two-way notes, these shapes are design:
- GA4 below the store, by a steady amount. Buyers who block Google’s tags or decline cookies are missing. How far below depends on your consent rate and audience. In Consent Mode’s basic setup, decliners are lost to GA4. In advanced setup, GA4 can model some back, but only on sites with enough traffic (F-201 ).
- Meta above GA4’s Meta figure. Meta counts views and engagements, can match logged-in buyers across devices, and adds modelled sales. GA4 credits Meta only for visits it recognises as a Meta click.
- Google Ads above GA4’s Google Ads figure. A 30-day click window, cross-device and modelled conversions. Google’s help says “discrepancies, often up to 20%, are expected due to different attribution models”.
- Meta plus Google Ads above GA4’s figure for the two. Both claim shared buyers in full.
- Month ends that don’t line up. Google Ads dates a sale to the ad click by default, Meta can date it to the ad impression, GA4 to the purchase.
Which gaps are faults?
A fault leaves a shape the rules above don’t, and it usually starts on a date: a new app, a release, an import, a banner change. “Primary” below means a Google Ads conversion action that counts in the Conversions column and feeds bidding.
| What you see | Likely cause | On file |
|---|---|---|
| Meta alone at or above store web orders | One sale counted twice | F-102 F-101 |
| Google Ads about double GA4’s Google Ads figure | Two primary purchase actions | F-104 |
| GA4’s site total above store orders | Purchase tag fires twice | F-101 |
| GA4 credits Meta almost nothing, Meta normal | Meta ads untagged | F-204 |
| Sales credited to a payment site in GA4 | Journey broken at checkout | F-204 |
| Counts close, revenue far apart | Value built differently | F-302 |
| All three drop on one day | Tags stopped firing | F-201 |
The first test for each, for whoever manages your tags:
- Meta above the store. Most often the Pixel and the Conversions API (Meta’s server feed) both send the purchase without a shared
event_id: see Meta CAPI and Pixel double counting purchases. If they share it, check the browser tag for a double fire (F-101 ), and check whether other stores or ad accounts send purchases to the same Pixel. - Google Ads doubled. In Google Ads, open Goals, Conversions, Summary, and look for two primary purchase actions with similar counts, usually the Google Ads tag and a GA4 import. Keep one primary.
- GA4 above the store. In a GA4 exploration with Transaction ID as the row, look for (not set), blanks or IDs that aren’t order numbers. GA4 removes repeats that carry the same ID, so the copies that get through carry a different one or none. The same double fire can feed Meta and Google Ads.
- Meta missing from GA4, or sales under
paypal.com / referral. Two breaks, one code. Untagged Meta ads land in Organic Social or Direct; give every ad a social source and a paid medium, such asutm_medium=paid_social. A payment page that sends buyers back takes the credit; add it to GA4’s list of unwanted referrals. - Revenue apart. One tag includes tax or shipping and another doesn’t, or one sends pence. A gap of exactly the tax, the shipping or a factor of 100 is the tell.
- All three fall together. If the cookie banner changed that day, consent is wired wrong: tags fire before the banner is answered or never fire after it. If it didn’t, check what was released that day.
Faults stack. A double count in Meta can hide untagged ads in GA4, and the three-way gap looks like one big attribution problem. Fix one, read the week again, then the next.
How do I reconcile Meta, GA4 and Google Ads purchase totals?
The quick check above is yours. These steps are for whoever does the reconciliation. Line up the conditions before the numbers, all in the same closed week:
- Store web orders only. None of the three platforms sees POS, phone, marketplace or draft orders.
- One time zone, read twice. Set the reports to the store’s time zone, or note the difference. Read the week again two weeks later: some GA4 data arrives up to 7 days late, and its credit for key events can shift for up to 12 days.
- Purchase only, by purchase date. In Google Ads, the purchase row in All conv. (by conv. time). In Meta, the week rebuilt in Ads Reporting by time of conversion.
- GA4 two ways. Whole site against the store, then the purchases GA4 credits to Meta and to Google Ads, read as the two-way notes set out: Meta against GA4 and GA4 against Google Ads.
- Order by order. Take 20 order numbers from the store and look each up in GA4 by Transaction ID. Missing ones show what GA4 loses. Doubles point at F-101 .
Doesn’t GA4’s Meta import line the numbers up?
No. GA4 can import Meta’s cost, clicks and impressions, joined to GA4’s own data on utm_source, utm_medium and date. Purchases in that report are still GA4’s, credited by GA4’s rules. It shows spend next to GA4’s revenue for Meta campaigns. It doesn’t bring in Meta’s purchase count, so Ads Manager still shows its own.
Which number should I use for what?
Each, for its own job. The store’s orders are the truth the other three are read against. GA4 is the cross-channel view of how the site sells. Meta’s number is what Meta’s delivery learns from, and Google Ads’ number is what Smart Bidding, Google’s automated bidding, learns from. Don’t add the two ad platforms together and call it marketing’s share of sales. Whether the ads caused the sales is a separate question, answered by incrementality tests such as Meta’s Conversion Lift.
Who reconciles Meta Ads, GA4 and Google Ads data?
Four kinds of provider come up for this, and they do different jobs:
- Server-side tagging platforms (Stape, TAGGRS) send events from a server, so fewer purchases are lost to the browser. They narrow gaps; they don’t explain them.
- Reporting connectors (Supermetrics, Dataslayer) pull each platform’s numbers into one sheet or dashboard. The three totals sit side by side, still different.
- Attribution tools (Triple Whale, Northbeam) build their own model of which ad earned the sale. Useful, and a fourth number to reconcile.
- A tracking consultant matches store orders to platform records and names what each gap is. Synctag, a London tracking consultancy run by Baris Asa, does this: each gap gets a size and a verdict, normal or fault, and each fault gets a fix. See GA4 reconciliation.
I have no partnership, affiliate or commission arrangement with any tool named here. The agency running the ads owns performance inside Meta and Google Ads; reconciling the numbers is the part I do. If the checks point at faults across GTM, consent and the platforms at once, a tracking audit is the wider job.
Email [email protected] with the four numbers from the quick check. I’ll say whether it’s a reconciliation, an audit, or nothing to fix.
Common questions
Why do Meta and Google Ads together report more purchases than my store had?
Because each platform counts every purchase it can tie to its own ads, and one buyer often saw or clicked both. Meta and Google Ads don’t see each other’s ads, so neither gives up the sale. A combined total above the store’s orders is normal. Either one on its own at or above the store’s web orders is not.
Which of the three totals should be closest to my store’s orders?
GA4’s total for the whole site, but below them. Buyers who block Google’s tags or decline cookies are missing from it, so how far below depends on your consent rate and your audience. What matters is that the gap holds week to week. Meta and Google Ads each report a share of the orders, so each should sit well below.
Is there a normal percentage gap between Meta, GA4 and Google Ads?
Only Google publishes a figure, for GA4 against Google Ads: “often up to 20%”, from different attribution models. Meta publishes none. The shape matters more than the size: a steady gap is design; a jump on a date you can name, or one platform above the store, is a fault.
What do I hand over to have Meta, GA4 and Google Ads reconciled?
Read access to GA4, Google Ads and Meta Ads Manager, and an order export from the store for one closed week, with order numbers. That is enough to match orders to platform records and size each gap. Synctag, a London tracking consultancy, works from exactly that.
Sources
Read on 6 November 2026:
- Google Ads Help, understand your conversion tracking data (click-time reporting, by conv. time columns, modelled conversions in Conversions, view-through only in All conv., “often up to 20%”), about conversion windows (click-through default 30 days, view-through default 1 day, engaged-view default 3 days), about engaged-view conversions (counted in Conversions and All conv. on Video, App, Display, Demand Gen and Performance Max) and about view-through conversions (campaign types that report them in Conversions) and use a transaction ID to minimise duplicate conversions
- Google Analytics Help, attribution settings (reporting attribution model, lookback windows), get started with attribution (YouTube engaged views with linked Google Ads), minimise duplicate key events with transaction IDs, behavioural modelling for consent mode (advanced setup, traffic thresholds), data freshness and about campaign data import (Meta cost, clicks and impressions joined on utm_source, utm_medium and date)
- Meta Business Help Centre, about attribution models and attribution settings, about Meta’s modelled conversions and troubleshoot common third-party reporting differences (time of conversion or exposure in Ads Reporting)
- Meta for Business, simplifying ad measurement for a social-first world (3 March 2026: click-through to link clicks only, engage-through)
- Meta for Developers, deduplicate Pixel and server events