Could a New Amazon App Weaponize Receipts?

Amazon, which has crafted an online retail juggernaut behind its own customer shopping data, is now engaging customers to directly assist in research of purchases they aren’t necessarily making online.

A newly released mobile application called Amazon Shopper Panel will pay select shoppers to upload receipts from their non-Amazon purchases and answer questions about them, essentially providing a platform for “tattlers” and a potential trove of intelligence Amazon might not otherwise discover, including information on the items and prices for goods shoppers are not necessarily acquiring from them—letting them know from whom, and what’s in the basket with them. This is turn could inform the Seattle-based e-commerce giant on product development, pricing strategy and advertising and marketing to both shoppers and suppliers, sources say.

The app, which was released earlier this month, offers invited shoppers the opportunity to earn monthly rewards of $10 in Amazon store credit—or a charitable donation of $10—in exchange for them uploading 10 eligible receipts per month. Additional rewards can be earned by completing surveys the app can feed to consumers. An example posted in an app screenshot by Amazon indicates these could offer rewards related to the estimated time to complete the surveys: a 1-minute survey offering a 25-cent credit, 2 minutes for 50 cents, and so on.

Eligible receipts include what Amazon terms as “most receipts for purchases made within the past 30 days,” including those from retailers such as grocery stores, drug stores, department stores and entertainment outlets such as restaurants, movie theaters and theme parks. Receipts from Amazon.com and other Amazon stores, including Whole Foods Market and AmazonGo, are not eligible.

Multiple Data Points

While retailers conducting so-called “opposition research” is nothing new—nor necessarily are the ideas of “crowdsourcing” or receipt-based research, which the Shopping Panel also incorporates—industry observers contacted by WGB were largely admiring of the cleverness with which Amazon stitched such efforts together.

“This is a strong move, especially for Amazon’s nascent advertising business,” Pradeep Elankumaran, co-founder of San Francisco-based online food grocery merchant Farmstead, said in an email. “It gives them multiple interesting data points—missing products to carry, along with which products are being purchased in which retail and online stores outside of Amazon. Also, it gives Amazon account managers a list of brands to target communications to drive Amazon purchase adoption at potentially higher margins. It also drives consumer loyalty with customers getting Amazon gift cards back. A very well thought out program.”

James McCann, the former Ahold USA CEO who today is chairman and CEO of Food Retail Ventures, an early-stage investment firm, told WGB there would be little to stop other retailers from developing a similar technology—if only they would be able to afford the hundreds of millions in monthly discounts such an effort could potentially represent.

“This can be very good for Amazon,” McCann predicted. “If you assume that a customer does 20% of their shopping with them, then the other 80% is dark to them.”

Amazon for its part is governing the availability of its Shopping Panel. Customers who download the app can be placed on a waiting list, which could potentially allow for the company to target particular shopper demographics or geographic areas. Such data could be useful for Amazon as it develops and launches the chain of Amazon Fresh grocery stores, noted Gary Hawkins, a former food retailer who today leads the Center for Advancing for Advancing Retail & Technology.

“I think dedicated Amazon shoppers will probably participate and provide that data that will give them that much more information about what shoppers are doing off Amazon properties.”

Other forms of connecting with shoppers when they are not in stores are also emerging from the tech world such as so-called hyperspersonalization apps like Anagog that use edge computing to inform marketers of shopper activities outside of their stores. Walmart also used competitor receipts as a kind of promotion with the since-retired Savings Catcher app that allowed it to provide shoppers with store credit in the difference in cases where competing prices for like items showed differences.

AI Processed

A description of the Shopper Panel on an Amazon web page suggests the company would use hybrid artificial intelligence methods to process the data. “The program relies in part on supervised machine learning, an industry-standard practice where humans review a small sample of submissions to help train our systems to correctly process receipt contents and eligibility,” Amazon said. “For example, a human reviewing a receipt can surface that the Amazon Shopper Panel misinterpreted a receipt for orange juice as a receipt for oranges. During our supervised learning process, Amazon reviewers only have access to the contents of receipts and do not have access to any other customer information.”

Amazon noted it would not share personal information collected via the Amazon Shopper Panel with third parties, saying however that the collected data would inform decisions in advertising measurement and models, and research and development of products. “We may use your purchase information and survey responses to improve the product selection on Amazon.com and affiliate stores such as Whole Foods Market and to improve the content offered through Amazon services such as Prime Video," it said. 

Hawkins said he believes the notion of engaging shoppers to participate in the process of providing shopping information is something of a future trend in itself, and likely to be influenced at some levels by privacy legislation.

“There’s something to this idea that a retailer can vacuum up all sorts of data, not only on their own customers, but on people in their market areas that are not customers, and intelligently go after them. I think Amazon is sort of early going down that path, but it’s no big surprise,” he said. “Amazon can ratchet that up or down as they see the need to, but I think it begins to open the door to this notion about paying the shopper for use of their data. And I think we’re going to see more of that, especially as we see more states roll out privacy laws like California has done. And I know of some startups that are working in this space, creating a capability for a consumer to control their data and only release it to companies that they want to, or that they want to do business with.”