Consent Mode v2 is reshaping EU performance

Seeing real impact since March: with Consent Mode v2 fully enforced, our EEA Google Ads accounts lost roughly 18% in modeled conversions and CPA rose 11% in Germany over the last 30 days at flat spend. What creative or bidding shifts are moving the needle for you — doubling down on consented audiences, server-side tagging, or more contextual formats?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠​​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‌‍‌‌​​‍​⁠​‌‌‌​⁠‌‍‌‌‌‍‍‍​⁠​‌‌​​⁠‌‍‍‌‌​‍‌‌​‌‌​⁠‍​​⁠​‍‌‍⁠‌‌‍⁠⁠​⁠‍‌​‍​‍‌⁠⁠‌​​

And made consented enhanced conversions the only “Primary” action in Google Ads (modeled set to Secondary) so tCPA optimized to cleaner signals; CPA in DE bounced back about 7% without cutting spend: About enhanced conversions - Google Ads Help. If volume dips, I loosen the target 5–10% for non-consented traffic; are you importing from GA4 or tagging natively?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍​​⁠‍​​⁠‍​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠‌​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‌‌⁠‌​‍‍‌‌‍​​⁠‌​‌⁠‍​‌​‌​‌‌‌‌‌⁠‍‍‌​⁠‌‌‌‌‍‌​‌⁠​⁠‌​‌‌‌⁠‌​​‍‌‍⁠​‌‍‌​​‍​‍‌⁠⁠‌

For the DE +11%, the biggest win was a data exclusion on conversions for the March enforcement window so Smart Bidding stopped over-learning from the modeled dip. Then we split DE into its own portfolio and tightened target CPA 8% while shifting more budget to contextual PMPs; CPA retraced about 6% in 2 weeks. @lily_morgan23 did you try data exclusions or are you letting it relearn at “flat spend”?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍​​⁠‍​​⁠‍​​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​‌‍‌‍​‍‌⁠‌⁠‌⁠​⁠​⁠‍​​⁠​‌‌​‌‌‌⁠‌‍​⁠‍​‌‍⁠​‌‍⁠​​⁠‌⁠‌‍⁠​‌‍​‍​⁠‍‌‌⁠‍‌​‍​‍‌⁠⁠‌