ROI Talk Is Back Because Nobody Trusted the Old Metrics
CMOs are being told to rethink ROI just as AI reshapes ecommerce ad measurement. The real issue isn't methodology—it's that the last decade's attribution models were never honest to begin with.
The conversation about how CMOs should think about ROI lands differently when you read it alongside Amazon scrapping planned API fees after backlash, and Alexa quietly becoming a shopping and advertising layer. These aren't separate stories.
The API fee reversal tells you that Amazon needed third-party tech firms on-side—measurement and attribution partners chief among them. Alienate them and the ad ecosystem's plumbing breaks. That's leverage, and the tech firms used it.
Meanwhile, Alexa-as-shopping-layer is about to make attribution even more contested. When a voice assistant closes a sale that started with a display ad three days ago, who gets credit? The answer depends entirely on whose measurement stack you're running on.
The ROI rethink Adweek is gesturing at isn't really about frameworks. It's about the fact that for years, brands accepted last-click attribution because it was convenient, not because it was accurate. AI-native commerce surfaces—Alexa, Netflix's ad agents, YouTube's creator ecosystem—are forcing the question back open because the old models plainly don't fit.
Brands that invest now in first-party data and platform-agnostic measurement will have something real to bring to these conversations. Everyone else will keep paying for attribution stories that flatter the platform selling the ads.