切換側邊選單 切換搜尋選單

Retail Revolution

New Tech Unlocks Potential of Unmanned Stores


New Tech Unlocks Potential of Unmanned Stores

Source:Chien-Tong Wang

Late last year, U.S. online retail giant Amazon launched Amazon Go, pioneering a new kind of store without a checkout counter. China and Japan are also battling to deliver a checkout-free shopping experience in unmanned stores. What are their secret weapons in this battle for new retail service models?



New Tech Unlocks Potential of Unmanned Stores

By Elaine Huang, Sydney Peng
From CommonWealth Magazine (vol. 630 )

When Amazon announced Amazon Go with great fanfare at the end of last year, it unleashed a global trend toward cashier-less or unmanned stores. Amazon Go’s “just walk out” shopping experience relies on smart visual recognition technology. Shoppers take products from the shelves, put them in their bags and leave without having to wait in line at a cashier. Meanwhile, China and Japan are also actively plunging into smart retail operations.

Lin Yuan-ching, head of the industry group at the Business Model Innovation Research Division of the Commerce Development Research Institute (CDRI), went to China a few months ago to check out the burgeoning staff-less retail trend there, which is helped by the fact that Chinese consumers already habitually pay via smartphone apps. In less than half a year, at least six foreign companies have joined hands with Chinese counterparts to open unmanned stores. 

“This wave has been created at the technology-end and not spearheaded by the retail service industry. The goal is to improve the retail service industry format,” says Lin, who does research on retail industry innovation.

In Taiwan, where convenience store density is very high, the retail sector is not going to sit on the sidelines of automated shopping. The Family Mart convenience store chain is planning to launch its first staff-less pilot store at the end of this year.

“Be it the United States, China or Japan, we remain at the stage where everyone is competing for technological solutions,” says CDRI Vice President Champion Wang.

Based on the complexity of technology used, the existing staff-less stores can be categorized into three main application models.

The technologically simplest model requires customers to use a smartphone app to scan the QR codes of the items they want to buy themselves instead of a cashier. 

Scanning QR Codes with a Smartphone

China’s Bianlifeng is a typical example of shopping with a QR Code app on a smartphone. The Bianlifeng convenience store brand was jointly founded by Wang Zi, a former executive of 7-Eleven Beijing and the founder of the Lin Jia convenience stores, and Zhuang Chenchao, the CEO of Chinese travel site Qunar. This linkup between a retail giant and an established Chinese internet company drew considerable attention from industry insiders.

Bianlifeng uses its own app within the Chinese social media smartphone service WeChat. After opening the QR Code scanner app, the smartphone connects to the store’s Wi-Fi and scans the store’s QR Code. As soon as the customer scans the QR Code of his purchases, they pay via AliPay or WeChat. Once the mobile payment is complete, the customer can leave the store.

“This does not require very advanced technology; the concept is to make the customer do the job that the staff normally does. Inside the store you only need to add video surveillance,” explains Lin.

Electronic Product Tags

Cashless payment systems with RFID (radio-frequency identification), where consumers’ purchases are detected at a gate or other device, also represent a technology that has been around for some time.

Japan is pulling out all the stops nationwide to promote the RFID model.
Japan’s Ministry of Economy, Trade and Industry (METI) is promoting a “100 billion convenience store electronic tags” program in concert with the nation’s five major convenience store chain operators. Family Mart, 7-Eleven, Lawson, Ministop and Newdays plan to roll out self-checkout computer kiosks in all outlets across Japan by 2025. RFID tags will have to be attached to an estimated 100 billion products.

In the future, customers need only to put their purchases into a shopping basket and put the basket on the self-checkout register, which reads the prices of the items in the basket, eliminating the need to scan each item separately.

“Scanning all items in a basket together for check-out is very convenient,” remarks Chaney Ho, CEO of Taipei-based automation solutions provider Advantech Corporation, with undisguised admiration.

The customer information obtained through the RFID tags can be sent to manufacturers and logistics companies to raise production efficiency.

Self-checkout depends on RFID technology, but a full rollout still presents quite a challenge due to the high cost involved. “The cost for relabeling with electronic tags is very high. Assuming the lowest relabeling cost for a single item is three Japanese yen, 100 million items will then lead to a cost increase of 300 million Japanese yen,” explains Lin Hsiu-yu, vice general manager of the corporate solutions group at NEC Taiwan, Ltd. In Japan, the point-of-sale backend systems in 7-Eleven and Lawson convenience stores come from NEC.

Aside from providing self-checkout functionality, the tags will also be linked to customer data in the future.

In Japan, the social messaging app Line has joined hands with general trade enterprise Itochu Corporation. Once consumers have registered their credit cards on mobile payment service LinePay, consumer behavior can be predicted based on previous purchases. Such information can be provided to retailers and upstream manufacturers so that they can adjust output accordingly at any time.

Like Japan, China is betting on RFID application models for its staff-less stores. BingoBox, a new chain of unmanned snack stores established jointly by Zhongshan Bingo Network Technology Co. and hypermart chains RT-Mart and Auchan, also uses RFID technology. 

As Lin observes, BingoBox outlets are mainly located in posh, upscale new residential areas and white-collar worker business districts. “They feel more like high-end vending machines,” he notes in describing the shopping experience at the glass box-like stores.

Scanning Faces, Judging Buying Behavior

Amazon Go’s “Just Walk Out” model relies on visual recognition, the most technologically challenging and capital-intensive model. 

Amazon Go, whose first concept store is located in Seattle, cuts the shopping process short, offering the most convenience.

Before shopping, customers need to first register an Amazon account and install the Amazon Go app on their smartphones. They then use the app to scan their account number for identification when entering the store. Inside, computer vision technology automatically detects when shoppers take products from the shelves or return them there. The customers' smartphone GPS and Wi-Fi assist with positioning.

When customers are finished shopping, they ‘just walk out’; when passing through the door, their purchases are automatically charged to their Amazon account.

China was quick to replicate the new walkout approach.

After poaching key Amazon Go developers for its A.I. Lab, the Chinese e-commerce giant Alibaba Group in July released the first application of its “unmanned retailing plan,” the Tao Café.

Tao Café is an offline, brick-and-mortar model store with a surface area of some 600 square meters. When entering the store through gates resembling subway ticket gates, shoppers only need to open the Taobao e-commerce app on their smartphones and scan the QR Code to obtain an electronic admission number. After the smartphone has been placed on a reader for authentication, shopping can begin; the smartphone will not be needed again. Customers can just walk out. When passing through the exit, the system automatically senses the amount spent and informs the customer of the purchase with a voice message.

“Basically, these are technological applications such as visual recognition, biometrics and machine learning. This is also the most convenient approach for the consumer and the provider,” says Lin.

However, so far both Amazon Go and Tao Café remain at the pilot stage due to cost and technology-related delays.

The Shopping Experience is Key

Ultimately, the essence of retailing means providing valuable services to the customer.

“When the customer comes to a store to consume, there should be an emotional connection, or else the unmanned store can only be a fad,” posits MC Chiang, AVP in the Service-IoT business group at Advantech. Chiang adds rather doubtfully, “Can the unmanned stores give the customer an emotional connection at the present stage? I feel we have not reached that stage yet.”

Last August, Advantech teamed up with French supermarket chain Carrefour to set up Taiwan’s first smart hypermarket in the Bade store in Taoyuan City. The store will use Advantech’s customer flow analysis software, which allows customer flow data to be sent to the central control console in real time as a basis for business reports. Customer smartphones will receive Wi-Fi advertising that will allow for positioning inside the store.

Thanks to a satisfaction survey taken during checkout, store staff can immediately address possible complaints. The store manager can take remedial measures immediately, even before the customer has left the outlet.

Retailing is, after all, a service industry. Does it make sense to make human staff obsolete and take away the warmth of human interaction? “The point is how to enable people to make use of things most efficiently, to achieve optimization,” says Lin.

The cost of running stores without any staff remains high. Most unmanned stores are still in the experimental phase. A key question is how the service process can be broken down into individual segments to experiment where the injection of smart technology would lead to a better process and more valuable services.

Lin points to the ordering process in convenience stores to make his point. In the past, it took about six hours to complete orders, which were placed category by category from fresh produce to miscellaneous goods to frozen foods based on the previous order history. Now, however, the system itself can suggest order quantities based on various parameters such as weather and events thanks to big data analysis and artificial intelligence. Consequently, the process of placing orders has been shortened to one hour, which translates into a reduction of unnecessary staff and inventory.

“It is very difficult to deliver service technology in a single step. After the service process has been broken down, we need to understand which segments can become smart or unmanned and which are not suitable, or else technology will not be able to add value to services,” Lin concludes.

Translated from the Chinese by Susanne Ganz