The cashier-less store space is currently dominated by Amazon and China’s Alibaba. But, as Scott Thompson reports, a number of startups have entered the fray, offering traditional retailers a way into this brave new world.
The combination of smartphones and ecommerce has made many customers demand effortless shopping with a click, which leaves bricks-and-mortar brands with a problem: friction and the perceived drag of queuing.
There has been a concerted effort by many traditional retailers to reduce the friction of in-store shopping in recent years. Self-checkouts are commonplace in many large supermarkets and chemists/drug stores, with cost-focused brands encouraging shoppers to use them, but some retailers are going much further. For example, Sainsbury’s in the UK and Walmart in the US have prioritised avoiding the checkout drag completely, by working towards a queue-less system.
The Sainsbury’s SmartShop app lets customers ‘scan and go’ via registered smartphones, or by using in-store handsets. Similar technologies have been deployed by the Co-op, among others, in partnership with Mastercard. Meanwhile, the world’s biggest retailer, Walmart, has trialled a number of different solutions, and has patented a range of new technologies that point towards what some are calling ‘Retail 2.0’.
However, all are being outpaced by Amazon and China’s Alibaba.
The first Amazon Go cashier-less store opened its doors to the public in Seattle earlier this year and plans are afoot to launch six more in 2018. Arguably, Amazon’s Echo and Dot devices go even further by opening a voice- and Alexa-skills-driven shop in consumers’ own living rooms.
Keep smiling!
Alibaba recently took the wraps off its own cashier-less store, at its Hangzhou HQ. Futuremart sells a range of Alibaba merchandise. Customers use facial recognition to enter the store and then scan a QR code with their Taobao, Tmall, or Alipay apps so they can shop.
A ‘Happy Go’ meter measures how happy they are, with big smiles earning discounts. A facial recognition system at the exits recognises each shopper, while the products they want to buy are identified by RFID tags. The store then charges their Alipay accounts.
While the ‘happiness meter’ approach might be too rich a solution for more cynical shoppers in the West, payment with a smile is a concept that is catching on in China via technologies provided by Face++ and others. The unique movement of facial muscles is, these companies believe, a more reliable identifier of a person than their face alone.
Measuring customers’ happiness also has a troubling subtext. In 2020, China will introduce a compulsory social ratings system, which seeks to bring together every aspect of citizens’ lives, from their shopping habits to their credit ratings, employment records, social histories, friend networks, and more. The stated aim of the system is to reward good behaviour and punish bad by, in extreme cases, removing citizens’ rights to buy certain goods or even to travel.
Netflix viewers may recognise the basic idea from ‘Nosedive’, a 2016 episode of Charlie Brooker’s techno-satire, Black Mirror. In future China, smiling for the camera may become a matter of self-preservation. Internet of Business editor Chris Middleton was one of the first Western journalists to write about China’s scheme in two external reports, here and here.
CheckOut Technologies
But not all frictionless shopping programmes have a national or social-engineering dimension, and retail’s big hitters are not alone in wanting to make in-store shopping as swift and friction-free for customers as possible.
A number of startups have also set their sights on the space, including Italy’s CheckOut Technologies, which is working on faster checkouts via AI, computer vision, facial recognition, and behavioural recognition, to create what it believes will be a more involving user experience for customers.
“The inspiration behind setting up the company came from imagining a supermarket without checkout lines, while I was in a checkout line,” says the company’s president, Enrico Pandian. “I had plenty of time to imagine it: there were three full shopping carts in front of me at the time.”
He observes that many traditional retailers are closing down as the ecommerce giants are not only putting the whole world of produce at buyers’ fingertips, but are also moving onto main street themselves, with their vast, integrated logistics systems behind them.
“This is the moment where online giants are starting to open their own physical stores, looking at an incredible opportunity,” says Pandian.
It is also the moment at which to save traditional retailers, to help them to be more appealing, customer centric, and frictionless, to innovate in an industry that has not been updated for almost 20 years,” he says.
The biggest challenge for CheckOut Technologies has been in creating the technology layer. “Deep learning is still a research field, it needs to be explored by highly skilled professionals. Finding the best ones from around the world has been a very difficult task,” says Pandian.
Another major challenge comes with integration: processes, platforms, and data. These have typically been developed in big corporations that have long separated ecommerce, physical stores, apps, and other customer touchpoints, he says.
“Omnichannel retail is a concept that is an internal fight between departments, more than an incredible opportunity at this moment in time. The few that have achieved it, such as Alibaba, are already increasing their revenues.”
Standard Cognition
Another startup, Standard Cognition, launched in the summer of 2017. Several of its co-founders worked together at the US’ Securities and Exchange Commission (SEC), developing AI-based fraud-detection systems. Some had also teamed up at a gaming venture that made use of computer vision technology.
They wanted to come up with a solution that merged AI and computer vision, in a market where it would have a big impact. After carrying out initial research, they decided to focus on retail and the challenges associated with the checkout. This was in 2016, before the world had even heard of Amazon Go.
“The threat that Amazon poses to bricks and mortar retailers is a huge issue in retail,” says Jeff Hsu, intelligence and logistics manager at Standard Cognition. “Another is the rising cost of labour and real estate, especially coupled with the Amazon threat. Retailers can’t just raise prices to make up the difference. Instead, their margins suffer. We help with all of this, levelling the playing field for them.”
The startup claims it is in talks with “nearly every major retailer in the US and many abroad”, and pilots will begin later this year. Its solution differs from Amazon’s in several key ways: it is available to any retailer; it has a much lighter footprint with far fewer sensors and cameras, which leads to faster installs; and it doesn’t require shoppers to use an app – there is a guest checkout feature.
“Retailers love the concept of autonomous checkout. No one likes to wait in line – retailers recognise that,” says Hsu. “Scan-and-go type approaches haven’t really taken off, as they are prone to error and enable theft, and most customers hate using them. Our solution is far more accurate, easier for the customer – they just shop, and walk out – and safer for the retailer.”
Teach and go
Like CheckOut Technologies, Standard Cognition’s journey has not been without its challenges, however. It takes time to teach a computer vision system how to identify items that look very similar. Even a small convenience store might have thousands of stock-keeping units (SKUs), and the system has to learn each one.
“We’ve also had some more pedestrian challenges, like how to see what items a person is carrying once they enter a dressing room, for instance, where we obviously can’t place any cameras. For those reasons, our initial pilots will be in smaller store formats, mostly in convenience stores,” says Hsu.
Investors have also been asking tough questions such as, doesn’t this invite theft? “Actually no, it doesn’t,” says Hsu.
“Computer vision-based systems are far more accurate than human security guards in terms of knowing when a person has picked up an item, even if they’ve put it in a pocket, inside a jacket, or in a bag. And we don’t rely on an honour system where customers are asked to scan their own items, so autonomous checkout reduces theft by a wide margin.”
But will customers be alarmed by all the cameras involved? “Virtually every retailer already has security cameras all over the ceilings in their stores,” says Hsu. “Our system doesn’t require any more cameras than they already have, just slightly different ones that are integrated with our in-store servers.
“Our cameras can double as security cameras and also enable retailers to take accurate, real-time snapshots of store inventory, so retailers get additional uses out of our system.”
In summary, therefore, reports of the death of physical retail may have been greatly exaggerated. There are numerous issues and challenges to be addressed, but the future will increasingly be powered by new ways of thinking and by tapping the likes of AI, computer vision, and mobile payments.
Internet of Business says
The influx of new technologies into big-name supermarkets, chemists, and other stores is clear evidence of the high-stakes, multibillion-dollar fight for customers’ hard-earned cash that is taking place on the streets of every town and city. And Internet of Business has reported on a range of other technologies, such as smart changing rooms and augmented reality applications in clothes shops, which offer new opportunities to sell products, increase customer loyalty, and deepen the relationship between buyer and seller.
But these approaches won’t work in every type of shop, or for every type of shopper; many people simply crave better, more human service and a more informed, personal, or even analog approach. So what’s needed is a large dose of common sense and customer sensitivity – particularly to the many people who hate being on camera, for example, or to people whose mental health challenges may also make them feel uncomfortable with the use of invasive systems.
While Hsu is right that there are cameras on the ceiling in most stores, many shoppers may feel that is different to putting them at the checkout and making shopping feel like being at the passport or security gate of an airport. How consumers feel in any shop is, after all, a multibillion-dollar industry in itself, and stores should consider the risk of alienating as many people as they attract. Sensitivity will be the key.
Also, many companies may find it impossible to resist linking facial recognition systems with AI and other technologies to start targeting advertising at people, or to enter into what some may see as invasive relationships with their customers – particularly if a supermarket also sells life insurance services, for example, or acts as their local pharmacy or even doctor’s surgery.
There is a fine line between what many in the West might see as China’s draconian social ratings scheme, and a large supermarket thinking, “Wouldn’t it be great if we could link someone’s shopping basket with their health records and life insurance premium – and maybe find out who their friends are?”
In the arms race for new technologies, retailers should also remember that the multiple scandals that have afflicted Facebook and other social platforms in recent months revealed that many consumers are increasingly distrustful of organisations’ data gathering and security. GDPR exists for a reason – to reset that balance – and Internet of Business predicts that more people will exert their right to be forgotten than some businesses realise.
Our advice: put yourself in your customer’s shoes and think about what is right for them. Invest in the technologies that – genuinely – make their lives better, and not in what would simply make their lives more useful to you.
Additional reporting: Chris Middleton.
- Read more: Why Alibaba’s ‘New Retail’ revolution is all about AI
- Read more: Retail: Walmart aims for majority stake in India’s Flipkart