1) You’re comparing different attribution models
GA4 reports are usually data-driven attribution (or last-click depending on the view), Google Ads has its own conversion logic and settings, and Meta reports are often view-through + click-through within a chosen window. If you compare “Purchases” across platforms without aligning windows and attribution rules, mismatch is guaranteed.
Why this matters
Different attribution rules mean each platform assigns credit differently. Without alignment you can misallocate budget and misinterpret channel performance.
How to check: compare conversions using the same attribution window and define a canonical source-of-truth for reporting (e.g., server-side conversions reconciled with GA4).
- Quick test: pull the last 30 days and compare conversions by aligned 7‑day click windows.
- Output: a short note describing the expected delta and an adjusted reporting dashboard.
2) Consent mode and ad blockers remove events upstream
If consent is denied (or an ad blocker blocks requests), GA4 may receive modeled or reduced signals, and ad platforms may receive nothing. The result is not a “reporting issue” — it’s missing input data.
Why this matters
Missing signals mean you underreport conversions in analytics and may over- or under-invest in channels. Server-side capture and modeled conversions can reduce this gap.
How to check: compare client-side event counts to server-side logs and consent-mode reports. Look for a consistent percentage of missing events by browser and device.
- Quick test: instrument a server-side purchase event and compare counts over a week.
- Output: ratio of client-side to server-side events and recommended mitigation (server-side tagging, modeled conversions).
3) You have duplicate or missing purchase events
The most common technical issues we see: duplicate events fired on page refresh, missing ecommerce parameters (value, currency), or server callbacks failing under load.
Why this matters
Duplicates inflate conversions and break attribution, while missing events hide true performance. Both lead to poor decision-making.
How to check: run a test checkout at low traffic and inspect the network calls, GA4 debug view, and server logs for one transaction ID.
- Quick test: complete 5 instrumented test orders and verify one-to-one mapping between order IDs and events.
- Output: list of missing/duplicate events and fixes (dedup keys, backend callbacks, GTM filters).
4) Your source/medium is getting overwritten
Even if purchases are correct, you can still get channel chaos if UTMs get stripped, redirects are misconfigured, or you’re losing gclid/fbclid through checkout. This shows up as “Direct” growing mysteriously or paid being underreported.
Why this matters
When source attribution is wrong, budget decisions based on channel ROI will be wrong. Fixing UTM persistence and redirect handling restores trustworthy channel reporting.
How to check: test flows from ad click through checkout and confirm UTM/gclid persistence across redirects and third-party payment flows.
- Quick test: click an ad with test UTMs and complete checkout; inspect final URL and stored utm/gclid values.
- Output: list of pages or redirects that drop UTMs and proposed fixes (cookie persistence, URL parameter forwarding).
5) The right goal is “decision-grade,” not “perfect match”
The objective isn’t making GA4 equal Meta to the cent. It’s building an instrumentation layer where:
Why this matters
A decision-grade system gives you consistent directional signals and a reliable basis for budget allocation. Perfection is expensive; consistency and reconciliation are valuable.
How to check: establish tolerance bands for expected deltas between platforms and focus on trends rather than exact counts.
- Quick test: define a reporting tolerance (e.g., ±15%) and compare weekly trends rather than day-to-day spikes.
- Output: documented reporting tolerances and a reconciliation playbook.
Want us to run the audit?
We’ll audit your tracking, reconcile platforms, and deliver a prioritized list of fixes and tests. Our approach includes server-side validation, UTM persistence fixes, and a 4‑week reconciliation plan. Book a call.
Implementation checklist
- Compare attribution windows and choose canonical reporting rules
- Instrument server-side events and compare to client-side
- Fix duplicate/missing events and persist UTMs across redirects
- Document reconciliation steps and tolerance bands