Why Your Google, Meta and GA4 Numbers Never Match (and What to Tell the Client)
Three platforms, three different conversion totals for the same campaign. Here is exactly why they disagree, roughly how far apart they run, which number to trust for which decision, and how to present it so the client stops asking.
Every agency has had this conversation. The client pulls up Meta, then Google Ads, then GA4, and asks why the same campaign shows three different conversion numbers. The instinct is to apologize, as if something is broken. Nothing is broken. The platforms are answering different questions, and once you can explain that clearly, the discrepancy stops being an embarrassment and becomes a sign you actually understand the data.
Why don't the conversion numbers match across platforms?
They will never match, because each platform uses different tracking rules, attribution windows, and modeling. This is by design, not by error.
Meta Ads Manager reports how many conversions its ads influenced — credited to a person it recognizes by login, inside the attribution window you chose, and including view-throughs. Google Ads reports how many conversions its ads drove, dated to the click rather than the purchase. GA4 reports sessions that arrived and converted on your site, stitched together by a first-party cookie, with credit shared across every channel that touched the journey. Three questions, three answers. Expecting them to agree is like expecting revenue, bookings, and cash collected to be the same number.
How far apart do the numbers usually run?
Far enough to matter, and in predictable directions. According to a cross-platform analysis compiled by EasyInsights:
- Meta over-reports by around 26% on average versus analytics tools, driven by modeled conversions and last-event attribution.
- Google Ads over-attributes by 15–20% when Enhanced Conversions or Consent Mode modeling kicks in.
- GA4 under-reports paid conversions by 18–35% when cookies are rejected or blocked.
So the same campaign can genuinely look like it drove 100 conversions in Meta, 85 in Google Ads, and 65 in GA4 — all "correct" by their own definitions. Knowing the rough size and direction of the gap is what lets you reassure a client instead of scrambling.
What actually causes the gap?
Four mechanics do most of the damage.
Attribution windows. Meta defaults to a 7-day click / 1-day view window; Google Ads lets you set anywhere from 1 to 90 days. A conversion inside one window may fall outside another.
Counting rules. Google Ads can count "one per click" or "every conversion." GA4 counts each event but splits its credit across channels via a data-driven model — so a sale Google logs as "1" might appear as "0.3" in GA4 because organic and email also touched the user.
Dating methods. Google Ads dates a conversion to the day of the click; GA4 dates it to the day it happened. A sale from a click 40 days ago lands on different days in each report.
Signal loss and modeling. When consent is denied or cookies are blocked, each platform fills the gap with modeled estimates — and each models differently. That is where the over- and under-reporting spread comes from.
Which number should you actually trust?
It depends on the decision. For channel optimization, trust each platform's own number within that platform — Meta's number to optimize Meta, Google's to optimize Google. For a picture of what the business truly earned, trust GA4 (or better, your own server-side / warehouse data) as the more conservative, de-duplicated source of truth.
The mistake is picking one platform as "the truth" for everything. Meta's number is right for the question Meta is answering and wrong for the question the CFO is asking. Both statements are true at once.
How do you present this to a client without looking wrong at scale?
Reconcile, don't apologize. The winning move is a single reporting layer that shows all three numbers side by side, labels what each one measures, and adds a blended, de-duplicated view on top. When the client sees Meta, Google, GA4, and a reconciled total in one place — with a one-line note explaining why they differ — the question disappears. You look like the adult in the room instead of the agency whose numbers "don't add up."
That is a data-layer problem, not a spreadsheet problem. Pulling three APIs into one place every month by hand is exactly where agencies end up confidently wrong at scale. A per-client data layer that normalizes the definitions once, and reports them live, is what makes the discrepancy explainable instead of awkward.
FAQ
Should I just use one platform's numbers to avoid confusion? No. Use each platform's number for optimizing that platform, and a single de-duplicated source (GA4 or your warehouse) for business-level truth.
Will server-side tracking make the numbers match? It narrows the gap by recovering lost signal, but it won't make them identical — the attribution-model and windowing differences remain.
Why does Meta always look the best? Because it counts view-throughs and uses last-event attribution inside its own window, which structurally credits Meta with conversions other tools assign elsewhere.
Is GA4 "more correct" than the ad platforms? It is usually more conservative and de-duplicated, which makes it better for total business reporting — but it under-reports paid when consent is denied, so it is not perfect either.
Reporting that reconciles instead of confuses is the whole job. Sifra builds white-label, per-client reporting dashboards for agencies — Google, Meta and GA4 normalized into one live view with a blended, de-duplicated total your clients can actually trust. Get a free mock dashboard built on one of your accounts and see the reconciliation done for you.
Sources: EasyInsights, Trackingplan, Ruler Analytics.