Global Retail Lessons From Amazon Go

Innovative retailers are checking out the benefits of cashierless stores

In 2016, Amazon opened its first Amazon Go store in Seattle, Washington. Since then, the ecommerce giant has launched in five other locations. The concept, as described by the company, is ‘a new kind of store with no checkout required’. The thing is, it isn’t exactly new. In 2015, long before Amazon Go’s initial store opened, Alibaba started its own cashierless grocery chain called Hema.

Amazon Go might be spearheading the cashier free movement in the US, but as Internet of Things technology becomes more available, they’re not the only ones. The question is, will any business be able to catch up?

Amazon at the front of the queue?

Across the world, various different companies are already experimenting with cashierless shopping including Alibaba, Dutch supermarket Albert Heijn, South Korean conglomerate Lotte and fashion label Jack & Jones. Many of these stores use a combination of biometrics and body recognition technology to equip physical shops with digital efficiency. The advantages are clear – purchasing is quicker, easier, and there is less friction. Inventory can be tracked in real time, and visibility across stores presents a huge data opportunity. Not only this, but it combats fraud by tying products to specific people.

What sets Amazon apart? Well according to Anshuman Singh, Head of Digital Business at Mindtree – a company that delivers digital transformation and technology services – Amazon have a number of distinct advantages.

“Most people think of Amazon as a retailer, but it’s not the only thing they do. 16 per cent of their business in non-retail. They have their cloud business and media business which have significant margins, so it’s possible for Amazon to subsidise innovation in another line,” he explains. “This opportunity is not available to other retailers. They operate on margins of one to two per cent, which doesn’t leave room for innovation.”

Innovation, then, becomes a money problem – and with such extensive product diversity, Amazon is well ahead of the pack. They have different product categories which grow at different rates.

“Most retailers will tell you they are spending close to 70 per cent of their IT budget on just keeping things running, 20 to 25 per cent on fuelling the growth side of the business, and maybe zero to five per cent on innovation,” says Singh. “The question is how you unlock innovative value from the 70 per cent by deconstructing legacy systems and reducing operational costs. The third way to look at it is diversifying the core business to find a larger customer base or alternate business-models.”

Taking advantage of the model

Retailers don’t necessarily need to totally overhaul existing stores to benefit from automated shopping. One easy to install service is Mindtree’s Flooresense, which uses video analytics from regular cameras to figure out what is happening on the shop floor, whether a customer requires assistance, and which areas are most visited. But while early adopters have a chance to get ahead of risk averse competitors, there might be sense in holding back.

“If you go back 10 years, Waitrose was the first store to introduce scan and go shopping. It worked for the demographic that Waitrose serves, but you don’t see that technology in any other supermarket chain,” says Singh. “If you go back seven years ago, it was much the same for the first self checkout services. It took all that time for them to become commonplace.”

As with all new ventures (case in point: ‘unexpected item in the bagging area’), it’s expected that cashierless shopping will experience some teething problems. Amazon Go assures customers that when they replace an item, it will be removed from their online basket. However, what happens when an item is put back in the wrong place, or a product that looks very similar to another is accidentally selected?

“In the first version that Amazon rolled out, there were examples of people putting things in their bag accidentally that didn’t get recognised. Once you have a crowded store environment and products that look similar, how do you distinguish between items?” asks Singh. “Look at the store formats – they are designed so that cameras can detect when items are picked from the shelf without any obstruction. Now imagine walking into an M&S store at Paddington Station at 6pm. You’ll find at least 20 people in an aisle. Can you accurately track those items? I don’t think it’s practically possible, today.”

Scaling up

Nonetheless, Amazon Go is now looking to scale up by opening 3,000 more stores. And when Amazon does something, other businesses start to pay close attention. Will competitors want, and be able, to offer an experience of the same quality?

“They might want to, but can they? It usually requires years of experimentation to make this work, and then you have your traditional challenges depending on what store format you have. In Amazon Go’s videos you see fairly empty stores with long spaces between the aisles, so there is an element of practicality. You may have to look again at packaging, layouts, format holistically to deploy a solution like this. Not every store would be predisposed to that and the cameras would simply not work.”

Singh believes that, in the long run, other more established ecommerce retailers will catch up and release their own answers to Amazon Go. In terms of a time frame, he sees five years as a realistic estimate. At the moment, though, Amazon has the margins to drive innovation at a pace and scale that simply can’t be matched by less influential competitors.

In other words, new entrants will need to wait in line.


Read more about retail disruption in our free weekly newsletter.