The Discovery-to-Demand Gen Migration Is Complete
As of February 2026, Google has fully sunset Discovery campaigns. Every advertiser has been migrated to Demand Gen whether they asked for it or not.
For most ecommerce brands, this feels like a rebrand. Same placements, same audiences, right? Not quite.
What Actually Changed
Demand Gen introduces several structural shifts that matter commercially:
- Lookalike segments replace custom intent audiences — Google now controls similarity thresholds
- Video and image assets are now blended across YouTube, Gmail, and Discover in a single campaign
- Performance reporting has been consolidated, making it harder to isolate which surface drives conversions
The Commercial Implications
If you were running Discovery as a top-of-funnel awareness play, the migration is mostly cosmetic. But if you were using Discovery to drive direct-response conversions — particularly for high-AOV products — the blending of surfaces introduces noise.
The risk is not that Demand Gen performs worse. It is that you lose the ability to attribute performance to specific placements.
What Ecommerce Brands Should Do Now
- Audit your legacy Discovery campaigns — check that conversion actions migrated correctly
- Separate brand and non-brand creative — Demand Gen blends everything unless you force separation
- Monitor cost-per-acquisition by surface — use placement reports to identify waste
- Set clear contribution margin targets — do not let Google optimise for volume at the expense of profit
The Bigger Picture
Demand Gen is part of Google's broader push toward AI-controlled campaign types. The pattern is clear: less manual control, more algorithmic optimisation. For brands that understand their unit economics, this is manageable. For brands running on ROAS alone, the margin erosion will be invisible until it is too late.
The question is not whether Demand Gen works. It is whether you can measure whether it works for your specific commercial model.