The doors never close on digital shopping

Initially, the thrill of online shopping was that the shops never closed; customers could browse online and make purchases far beyond conventional opening hours. Receiving those purchases, however, was often another matter. Placing orders online at 11pm on Christmas Eve was easy, but did not guarantee delivery in time for Christmas. The initial leaders in internet shopping were high street names which had efficient warehousing and delivery systems.

The definition of a disruptive technology is that it destroys the incumbent, but this has not happened in the case of retail. In a pre-pandemic survey by TimeTrade, three out of four consumers said they preferred shopping in a physical store. Reasons given were that customers can see and feel products, see availability in terms of size and colour and – for grocery shopping – compare freshness. A 2020 survey of retailers by Future Stores and WBR Insights (‘Engaging Digital Natives in Physical Retail’) found that 38 per cent of retailers believe digital natives (i.e Millennials and Gen Z-ers, who have never known life without the internet) prefer in-store shopping because it provides instant gratification and enjoy a social element to bricks and mortar shopping. At the same time, they prefer the personalised discounts and convenience of ‘open all hours’ online shopping.

Internet shopping quickly digitised the practices of experienced sales staff, who would up-sell. If buying a suit, for example, a skilled sales assistant might suggest a shirt and tie to complete the look. The internet does this with suggestions of what other products people also looked at or bought when they chose your selection.

This artificial intelligence (AI) is creeping into physical stores. Customers like discounts and AI can help retailers identify loyal customers and target special offers for particular products. Some retailers use smartphone apps to send an alert when a loyal customer enters the store, or a customer’s online searches could be used to let the retailer know they were looking for a pair of summer shoes and send an alert that there is 10 per cent off sandals today.  

Online and offline retailers use customer data to tailor promotions according to their shopping history. Not all use the internet. Beacons, small wireless devices, dotted around the shopping area, can send content to a user-enabled smartphone app. It can detect when a shopper is near an area where there is a promotion in-store, nudging them to look at a particular product there.

The internet is used in vision systems within the store. AI is used for facial recognition to categorise a shopper according to age and gender for generally targeted promotions, such as men’s or women’s fragrances, without any data being exchanged. Retailers and even specific brands can use the internet to make these offers more targeted.

Supermarkets charge brands for particular shelf slots so it is important that appropriate products are in place. Brands such as Pepsi, Hershey’s and Hasbro use AWM’s vision computing and sensors to ensure appropriate products are on the shelves or in the aisles.

Vision computing systems are internet-linked cameras that can recognise a customer and guide them to particular areas where ranges have special offer pricing or promotions. Cameras and weight sensors linked to servers track what people pick up from the shelf, gondola, freezer section or hangers. “Our weight sensing is down to 0.75g,” says Kevin Howard, CEO at AWM. The smart scales help manage inventory effectively not just with the appropriate versions and quantities but also by tracking when a product was placed on the shelf to detect when it is close to its expiration date. The shelf price is reduced in real-time and customers who have signed up to an app get a notification directing them to the special offer. The technology can also be used if there is a spike in sales of a particular product, for example headache tablets, and an alternative can be promoted.

Within a store or selected area of a shop, AWM’s Retail Data Engine can recognise a loyal customer and, with an understanding of their purchasing habits, guide them around the store. Skeletal structure tracking, or posed detection, is the technology that tracks where shoppers go and what they pick up. AI allows the retailer to learn more about the shopper’s behaviour and create predictive modelling to optimise the layout of the store and suggest a route for the next visit. The aim is to reduce the time spent in the store by 30 minutes, says Howard.

Access to internet search history means it can make suggestions of associated products that they typically put in their basket as they go around the store.

AWM also provides a Demographics Engine which uses computer vision and machine learning to capture shoppers’ data like age, gender and ethnicity. “One of the things that is on our roadmap to implement is an immersive environment,” says Howard. For example, if a parent enters with a crying child in a pushchair, soothing music or visuals can distract the child. This immersive environment can also extend to enticing environments, for example peat moss smells or highland imagery around whisky promotions.

Internet-based intelligence can also be used to tailor support for customers. “If you are selling a recipe, sometimes you might target women differently to men or one age range differently to another,” explains Howard.

For a more interactive encounter, there is Pepper, the retail robot. One of the reseller partners of the SoftBank-created robot is Robots of London. Pepper uses AI to understand questions asked in a variety of ways: from ‘I want to know how much is your most expensive pen’ to ‘What’s the highest-priced pen?’.

Robots of London’s founder, Adam Kushner, explains that the software is written specifically for each retailer. “It could be ‘what does [this product] do? Is it in stock?’ It can also provide more information, for example, by displaying a video to show it working,” he says.

If the customer is persuaded and wants to buy the product, Pepper can take care of that too. After accessing inventory, a payment screen can be displayed and the payment made via an internet connection.

Pepper can recognise returning customers and recommend items that might be of interest based on shopping habits. If a loyal customer spends a certain amount, Pepper can also tailor a discount when it takes payment, or point the customer to related special offers.

As a store’s electronic greeter, Pepper can recommend products. The camera scans the customer to tell if they are happy, sad or indifferent and selects an appropriate welcome message. It might then ask if they would like to look at a particular product the store is promoting, based on its assessment of their age and gender.

Storage capacity is 32GB but an internet connection expands that to allow the robot to access data on servers too. “I think people actually prefer interacting with a robot,” says Kushner. “They can get all the information they need, location, price, any further information about the product, without disturbing someone filing shelves,” adds Kushner.

There are Peppers in stores in the UK today but Kushner believes countries like France, Germany and Austria are more accepting of robots in a retail environment. “And outside of Europe, in Japan, China, it is a common occurrence when you go inside a retail store to see robots.”

However, it may be that shoppers don’t share Kushner’s professional enthusiasm. In June this year it was reported that SoftBank is ‘pausing’ production of Pepper robots, blaming the pandemic and economic slowdown.