Attribution Discrepancy After Consent Mode v2
A European e-commerce retailer implemented Consent Mode v2 in March. Before implementation, Google Ads reported 12,000 conversions in March; after, it reports 8,500. GA4 reports 9,200 conversions for the same period. The CRM shows 10,000 orders with a 30-day attribution window. The marketing team is debating whether to increase Google Ads budget to compensate for the 'lost' conversions. The analytics lead notes that 40% of users reject cookies, and Consent Mode modeling is enabled but has a 7-day learning period. The average order value is €80, and the Google Ads cost per acquisition (CPA) was €25 before the change. The team needs to decide on budget allocation for Q2.
1. What is the most likely explanation for the discrepancy between Google Ads, GA4, and CRM conversion counts, and what immediate action should the team take before changing budgets?
Read the worked answer
The discrepancy is primarily due to consent mode modeling and attribution differences. Google Ads uses modeled conversions for users who reject cookies, but the modeling may undercount compared to CRM. GA4 uses a different attribution model (data-driven) and may have longer lag. CRM is the source of truth for actual orders. The team should not increase budget based solely on platform-reported conversions. Instead, they should: (1) Compare CRM orders to platform conversions over a longer period to assess modeling accuracy; (2) Use CRM data to calculate true CPA: €25 * (12,000/10,000) = €30 if pre-change CPA was based on platform conversions, but actually pre-change platform overcounted? Need to reconcile. Better: Calculate true CPA from CRM: total ad spend / CRM orders. If spend was €300,000 (12,000*€25), true CPA = €300,000/10,000 = €30. Post-change, if spend unchanged, platform CPA appears €35.3 (300k/8.5k) but true CPA still €30. So no need to increase budget; instead, adjust bidding to use CRM-based conversion imports or enhance conversions. Also, ensure consent mode modeling has enough data and consider using blended ROAS. Immediate action: set up offline conversion imports from CRM to Google Ads and use value-based bidding with true order values.
2. How should the team adjust their bidding strategy to account for consent mode and data loss?
Read the worked answer
They should implement enhanced conversions for leads and offline conversion imports to feed CRM data back into Google Ads. Use value-based bidding with true order values from CRM, not platform-reported values. Additionally, they can use consent mode modeling but monitor its accuracy against CRM. For users who reject cookies, they can use modeled conversions but should set a buffer in target CPA to account for undercounting. They should also consider using GA4's behavior modeling and export to BigQuery for more granular analysis. A/B test bidding strategies with a holdout group to measure incremental lift. Finally, they should ensure that conversion windows align across platforms and CRM to avoid discrepancies.