A fresh approach to the retail experience

Londoners can now enjoy checkout-free grocery shopping in one of five Amazon Fresh stores.

Amazon’s ‘just walk out’ concept had just been launched when I was completing my book, Contemporary Retail Design – A Store Planner’s Handbook. I was excited to see this for myself, so visited the Canary Wharf store. It was a seamless experience – as perhaps we’ve come to expect from Amazon.

The store offers a range of grocery goods familiar to anyone who buys lunch or a daily shop from an express or metro-type supermarket, although Amazon’s 250 sqm retail footprint is perhaps slightly more compact.

Passing through the Amazon Fresh green branded storefront, the customer is faced with an entry gate equipped with scanners. Access is gained through the Amazon app which generates a QR code which the customer scans to open the gate. From then on you can put your phone away and simply pick what you want to buy, put it in your bag and walk out when you’re finished. No till, no scanning, no cash or card transaction required. Afterwards and returning to the app, the items you just took away appear in the Your Orders section among your usual purchases.

The aesthetic of the store is clean and simple with a disciplined and ordered layout. Canary Wharf is blessed with a great open ceiling void and looking up above the display you’ll see what makes Amazon Fresh function – technology that Amazon describes as ‘computer vision, sensor fusion and deep learning’. This allows the store’s systems to track the merchandise from shelf to bag without the need for asset-tagging.

Essentially the shopping experience could be human-free but during my visit, access to wines and spirits was controlled by a sales associate. This is to prevent the obvious risk that alcohol could be purchased by minors. In theory the same deep learning already employed in the store could be deployed to verify age. One of the case studies in Contemporary Retail Design explains the work of Aura Vision whose software analyses sequences of camera images using artificial intelligence to determine demographic information from even partial data, for example when the object is obscured, a person’s face is not visible, or the image blurred. The system’s accuracy is ninety percent for age where the results on adult age demographics fall into seven groups, each one spanning ten years. Perhaps the need for human intervention is nevertheless a requirement of the licensing authorities?

Relieved of their sales duties, the staff are on hand to happily guide and explain and it’s clear that this is a great opportunity to take convenience shopping customer service to previously unknown levels. This is where Fresh really could be revolutionary. Rather than walk out shopping being a cold, robotic experience, the time won back for human interaction with the customer could set this type of retail apart.