Amazon says it is shutting down a controversial program in 2020 that uses AI to send shoppers free samples based on their buying habits
Annie Palmer / CNBC :
Context & Ripple Effects
The shutdown bookends a short experiment: in January 2019 Amazon quietly began letting brands like Folgers pay to send samples to shoppers whose order history suggested they'd enjoy them. Barely a year later, the company is ending the program entirely.
It is also the second quiet retreat from a shopping-experience experiment that year — in April, sources reported Amazon was [[a:940258|removing or relocating promo spots that gave special treatment to its private-label products]]. Together they sketch a retailer trimming programs that read as Amazon tilting the storefront toward its own commercial interests.
First-order effects
- Brands that had signed on as paying participants — Folgers was the named example in the pilot — lose a paid channel built specifically around targeting shoppers by purchase history.
- Shoppers who received algorithmically matched unsolicited packages under the scheme stop getting them once the program winds down in 2020.
Second-order effects
- Consumer-goods marketers evaluating Amazon retail-media experiments now have a cautionary precedent about building campaigns on purchase-data-driven formats, pushing sampling budgets back toward channels they control.
- The retreat gives rivals' ad and retail arms a talking point against Amazon's broader push into automated merchandising decisions, which reporting has shown already extends to inventory selection and operations.
Third-order effects
- If this joins the private-label promo pullback as part of a durable pattern, Amazon faces pressure to formalize how it governs consumer-facing uses of purchase history — deciding which data-driven programs survive scrutiny rather than winding them down case by case, the way it later did with hundreds of initiatives in its DEI review.
The trend: Amazon is selectively retiring data-hungry shopping features when their optics outweigh their revenue, an early instance of the platform recalibrating how far behavioral targeting reaches into the storefront.