GA4 vs Google Ads gap diagnoser

What's inside

• 1 · View-through conversions — Ads counts impressions that led to a conversion; GA4 never sees them. Expected +30–60%. Definition: report Ads view-through separately.
• 2 · Consent Mode basic drops EEA users — Ads models the consented-out clicks; GA4 in basic mode cannot. Expected +5–25% of EEA. Fix: advanced consent mode (artifact A-002).
• 3 · Data-driven attribution shares credit in GA4; Ads keeps 100% of what it touched. Expected ±10–20%. Definition: compare on Advertising › Attribution paths, not the key-events table.
• 4 · The Ads tag and the GA4 import both counting — the same purchase twice in Ads. Up to 2×. Fix: one conversion action per outcome (artifact A-007's audit).
• 5 · Conversion windows differ — Ads' click window vs GA4's event date. ±5–15%. Definition: compare 'by conv. time' in Ads.
• 6 · No server-side tagging — browser-side GA4 loses hits that Ads' click ID still catches. +10–30%. Fix: server-side Tag Manager; since 1 May 2026 sGTM setups with Google Signals narrow this on their own.
• 7 · Currency and value rounding — affects value, not counts.

Install

1. Write down your setup: import or tag, attribution model, consent mode basic/advanced, click window, server-side yes/no.
2. Walk the seven causes; mark each definition or fix.
3. Sum the expected sizes of the definitions: if your gap sits inside that band, fix only the flagged rows.

Breaks on

The logged changes that can invalidate this artifact. A new one triggers a re-check.

• 16 Jan 2026 · Per-conversion attribution settings (beta) for Google Ads customers
• 1 May 2026 · Improved conversion tracking and attribution for Google Ads-linked properties on server-side Tag Manager
• 5 Nov 2025 · User-provided data attribution improvement
• 18 Jan 2024 · Consent mode required for EEA audience and measurement features (the v2 signals)