It’s obvious to anyone who uses Amazon (AMZN) how the company tailors its website to customers’ preferences. Your purchase history and browsing history are factored into what you see when you show up to the home page.
As the retail giant expands its brick-and-mortar footprint with bookstores, Amazon Go stores and 4-star stores, it must take a different approach to target customers. Unlike online, each customer cannot have their own store.
The death of physical retail has been overstated; Amazon and other previously online-only or online-centric retailers – including Glossier, Bauble Bar, and Warby Parker – have seen value in and adding brick-and-mortar stores.
Unlike traditional retailers, these online players have been able to leverage what they’ve learned online to make a store do more with less.
Amazon’s forays into physical retail — absent its acquisition of some 500 Whole Foods stores — has been especially novel given its cashier-less model, and it makes good use of its online edge.
Knowing thy customer
The 12 Amazon Go stores currently in existence (soon to be 13) are laid out and frequently updated based on feedback from customers through the app. But since the Amazon Go app identifies a customer’s account when they scan into the store, the company has another level of information beyond simple inventory levels that show what people are buying. Knowing which account is buying helps.
“We plan and adjust the selection in our Amazon Go stores to cater to what we think customers shopping at that location will most likely be looking for,” an Amazon spokesperson told Yahoo Finance.
These stores have vast arrays of cameras to determine who picks up what and buys what. But the cameras give the company more than that. In most well-capitalized brick-and-mortar retailers, cameras from companies like RetailNext monitor the floor and windows to see how people interact with the store, giving companies useful insights through heatmaps that show where people gather, indicating what may be working or not working.
Amazon Go was first developed and tested in the company’s headquarters in Seattle, but each location hosts a different selection of products based on the likely clientele’s preferences — not dissimilar from how most stores operate, but gleaned from Amazon’s extensive data and testing. After all, Amazon has the addresses and order history of its customers.
At the 19 Amazon bookstores and three Amazon 4-star locations, this is the most evident. The concept behind the 4-star store is, of course, to feature products that have garnered four-star ratings from Amazon customers. The curation and data application is in the name. For those 4-star stores as well as the few bookstores the company has, pre-orders, sales, trends, reviews, and ratings all work to curate the selection. After the fact, Amazon surveys customers to get more data on their preferences. With so few stores, these projects — which are at a small scale — have a pilot project feel.
While the Amazon web and in-person experiences in its cashier-less Go stores are linked, shopping history data is not collected for the company’s non-Amazon-branded Whole Foods stores, unless someone decides to identify themselves to get Prime deals.
Surveys are used throughout the network of physical stores at both Amazon-branded and non-branded stores alike.
“We get lots of great feedback directly from customers through the Amazon Go app and use that feedback to evolve selection to fit our customers’ needs,” a spokesperson said.
The same is true for Whole Foods stores, which see extra value in these surveys as customer data collection is limited to the shoppers who choose to identify themselves as Prime members, often to get discounts.
As of now, the company’s integration with Whole Foods has been done somewhat slowly and deliberately, as it has done with expanding its brick and mortar presence with Go, 4-star, and bookstores. With fewer examples in the wild, that means most of Amazon’s learnings from all of this data are still hidden.