Amazon will change the retail industry
by implementing AI technology
JULY 7, 2017
Amazon, the US e-commerce giant, just bought Whole Foods, an upmarket chain of supermarket in America, for 465,000 million baht. Combining Amazon’s technological prowess in data analytics and AI with a chain of physical stores could create an unassailable competition for other retailers. Traditional retailers will need to adapt to this new kind of retail if it were to ensure their survival.
Amazon is interested in grocery because food retail is a high-frequency shopping event. Unlike buying a TV, people need to go food shopping on a regular basis. Adding food and grocery will also increase Amazon’s share of wallet for the average consumer. Whole Foods acquisition will help Amazon achieve these aims. It will allow Amazon access to 430 stores that act as both places people can shop as well as a distribution hub for its food deliveries. Another key asset Amazon will gain is the data Whole Foods has on its customers that the business has collected over the years. But historical customer data is not as interesting to Amazon as the future data it will begin to collect.
Amazon is currently trialing a cashierless store, Amazon Go, where people can just go into the store, grab what they want and walk out. It handles all this by using a multitude of sensors that track all the customers behaviour in the stores and feed its AI. By using cameras to track who picks which product, it can add those products automatically to the customer’s basket. When the customer walks out, the AI deducts the correct amount from the customer’s credit card. Unfortunately, it seems that this system cannot handle more than 20 customers in the store at any given time and a wider launch has been delayed. However, the underlying sensors and AI technology that goes into this cashierless store could create a gold mine of customer data if it were to be implemented at Whole Foods.
By equipping stores with similar sensors and cameras, Amazon will be able to track all the behaviours within the store. From the path people take to find a product to the time spent deciding which drink to buy on a hot summer day, all this and more will be collected. It will be able to record which product people pick up then put down again, how long they look at the label before putting it down and which product they then ultimately buy. This unprecedented level of detail is a dream come true for data scientists and will allow Amazon to optimise the stores further than any other retailers in the world, maximising sales and customer satisfaction. From better store layout, assortment and individualised promotions, the sky’s the limit. Not only that, these sensors will also help detect and prevent theft in stores, which can be 1-2% of total sales, a significant amount for a low margin grocery retailer.
Is there a way for existing retailers to fight back? Yes, if they move fast enough. Retailers need to start implementing some of the sensors to collect as much information as they on their customers in the stores and invest in the AIs that will make sense of all this data. At Sertis we are working on an AI that uses cameras to monitor retail stores, to check if things are out of stock and to suggest remedial actions. The same system can be used to analyse customer and employee behaviour within the store, to give retailers a better understanding of what goes on in each branch. Better assortment and reducing shrinkage are additional applications for this system. With this, retailers will be able to start match some of the tools a tech company can bring to play.
Retail is detail, so goes the saying. Amazon is the king of using detailed data on their customers. It understands more about its customers than most other retailers. It tracks every single interaction of users within its sites, and now it will do the same for its physical stores. All this in-depth knowledge of the customers is one of the reasons why Amazon market cap is double that of Walmart. By buying Whole Foods, it will bring all this technology to the physical world. Retailers need to be ready for this new stage of data evolution, or they will risk extinction.
This article first appeared in Thai on Think Data Science Column at Bangkok Biz News.