Challenge
Expanding reach beyond Amazon Audiences
Brands expect their programmatic campaigns to reach the right shoppers, but audience reach has natural boundaries. Amazon Audiences provide a powerful foundation for engaging in-market shoppers based on Amazon shopping signals. Beyond that, brands wish to have visibility into shopper behavior across third-party retailers, brands, and publisher sites, where they can discover potential customers during the research and browsing phase — before those shoppers ever arrive on Amazon.
A home essentials brand, managed by Flywheel, wanted to understand whether Criteo Commerce Audiences could help them with the extended discovery. Criteo’s shopper graph draws on browsing and purchase behavior across brands, retailers, and publishers — identifying in-market shoppers in early purchasing phases. The goal is to determine whether layering Criteo Commerce Audiences with Amazon Audiences through Amazon Ads campaigns could reach new buyers on the open internet and support higher conversion rates when those shoppers later encountered the brand’s Amazon onsite ads.
The hypothesis: If Criteo Commerce Audiences could reach shoppers during their initial research on the open internet, and Amazon Audiences could deepen consideration once those shoppers arrived on Amazon, then the brand’s Sponsored Brands and Sponsored Products campaigns might convert those shoppers at higher rates, creating a full-funnel path from discovery to purchase.
Solution
Testing complementary audiences to enhance brand consideration
In August 2025, Flywheel worked with the Criteo Commerce Grid team to determine how combining Criteo audiences and Amazon audiences within Amazon Ads could drive better campaign outcomes for the brand. The pilot was designed to test a full-funnel path: Criteo Commerce Audiences reaching new shoppers on the open internet, Amazon Audiences deepening consideration on Amazon, and Sponsored Brands and Sponsored Products driving conversion onsite. The question was whether these three touchpoints, working together, could outperform any one in isolation.
In the controlled group campaign, Criteo built audiences using its cross-retailer shopper graph and delivered them to Amazon Ads through a Deal ID. In one campaign, those Criteo Commerce Audiences ran alone — reaching in-market shoppers on the open internet who may not yet be browsing on Amazon. In the experiment campaign, Criteo Commerce Audiences were combined with 16 Amazon In-market segments (spanning categories like Lighting, Smart Home, and Home Improvement).
Results
Criteo Commerce Audiences and Amazon Audiences: Complementary signals with distinct strengths at every stage
By combining both audiences from Criteo and Amazon Ads, Criteo Commerce Audiences extended the brand’s Amazon Ads campaigns to a broader set of in-market shoppers on the open internet, while Amazon Audiences deepened engagement with shoppers showing strong purchase intent. For the brand, this meant reaching shoppers earlier in their journey and guiding them toward Amazon, where the brand’s onsite advertising could drive consideration and conversion.
- Combined Criteo and Amazon Ads audiences drove 2.18× higher DPVs per unique user: When Criteo Commerce Audiences and Amazon Audiences worked together, shoppers viewed product detail pages at 2.18× the rate of Criteo-alone shoppers.¹ This indicates that combining Criteo’s cross-retailer signals with Amazon’s shopping-intent signals helps identify shoppers with a stronger likelihood to purchase, ensuring impressions are delivered to engaged, in-market consumers. Both signal sets contributed — Criteo identified the shopper on the open internet, and Amazon audiences confirmed their purchase intent.
- Highest observed conversion lift was 62% when offsite and onside touchpoints combined: An Amazon Marketing Cloud (AMC) path to conversion analysis examined how Criteo offsite exposure interacted with Amazon Sponsored Brands and Sponsored Products across the brand’s full advertising program.² Shoppers exposed to the full combination — Criteo offsite, Sponsored Brands, and Sponsored Products — purchased at 2.73 vs 0.04% over Sponsored Brands + Sponsored Products alone. This demonstrates the full-funnel value of combining both strategies: Criteo Commerce Audiences introduced shoppers to the brand on the open internet, and Amazon’s audiences (Sponsored Brands and Sponsored Products) deepened engagement and converted them at a higher rate. Each touchpoint amplified the next.
- New-to-brand acquisition signal: When Criteo offsite ads served as the first touchpoint in a shopper’s path, 92% of resulting purchases came from new-to-brand customers.² Criteo Commerce Audiences introduced the brand to shoppers who had never purchased from it before — and Amazon’s onsite advertising then guided those new shoppers to conversion. This suggests that combining Criteo’s top-of-funnel reach with Amazon’s onsite advertising can help brands acquire new customers, not just re-engage existing ones.
The takeaway is clear: Criteo Commerce Audiences and Amazon Audiences deliver the strongest results when they work together within Amazon Ads. Criteo extends reach to new in-market shoppers on the open internet; Amazon Audiences and Sponsored Brands and Sponsored Products deepen engagement and drive conversion onsite. Each strategy amplifies the other — and both are more effective together.
Together, Criteo Commerce Audiences and Amazon Audiences helped support a full-funnel view of how shoppers discover, evaluate, and purchase the brand’s products. The results confirmed the hypothesis: each touchpoint amplified the next, from offsite discovery, to onsite consideration, to conversion, delivering the strongest outcomes when all three worked together.
For brands running Amazon Ads campaigns today, Criteo Commerce Audiences offer a way to extend those campaigns to additional in-market shoppers across the open internet— expanding the top of funnel while strengthening the performance of every downstream Amazon touchpoint.
¹ Amazon Ads reporting, US, August–October 2025. Results may vary. This case study reflects one advertiser’s experience during a defined campaign period and is not indicative of typical results.
² Amazon Marketing Cloud (AMC), US, August–October 2025. Results may vary. Note: This NTB figure reflects early-stage pilot results based on a small sample and should be considered directional.
