Unmanned Store - Future of Retail
Alexis Ben
Market Researcher
Business Analyst
Business Consultant
BearingPoint
Overview and Shoppers’ Perceptions
Introduction
Technology is now redefining commerce and unmanned stores are at the forefront of a new concept called “New retail”. This idea was put forward in 2016 by Jack Ma, co-founder and former executive chairman of Alibaba Group. The concept relies on a seamless engagement between the online presence, offline retail experience and logistics service through data technology.(1)
One emblematic example of “New Retail'' is the Amazon Go store concept launched in December 2016 in Seattle (US) with the promise of revolutionizing shopper experience in store. At Amazon Go, a customer can purchase and leave the store without having to scan a product nor put its credit card out of the pocket. This is made possible through dozens of cameras and weight sensors tracking all movements to recognize the items put in the customer's bag. This unprecedented move of Amazon into Retail through an unmanned store has drawn a lot of attention and media coverage(2). Leveraging its tremendous financial resources, Jeff Bezos’s company invested more than US$11 million to develop the technology able to remove waiting lines and cashiers from the store. With state-of-the-art cameras and complex algorithms, big tech has entered brick and mortar retail. Amazon’s initiative is not left isolated and major Chinese tech companies are following the same path. Alibaba even went further with its “Hema'' store concept where the digitalization of the shopping journey has removed almost all human interactions. A customer shopping in Hema can order fresh food at the store restaurant, pay for one's basket, and request home delivery without having to talk to anyone.
Both Amazon Go and Hema stores are unmanned stores meaning they are leveraging technology to make shoppers more autonomous and remove traditional paint points such as waiting lines. In both cases, robots are taking on human interactions and massive amounts of personal data are being used and stored. Legitimate questions are being raised regarding the social impact of unmanned stores and a breach on personal privacy.
Nonetheless, the concept has been highly criticized concerning the protection of personal data as the fear of being controlled in our choices unconsciously is growing. Moreover, making profit after the technical investment can be hard to manage.
Despite similarities, Amazon and Alibaba actually portrait two different visions of what the new shopping journey should be in the unmanned store. Where Amazon puts forward efficiency, flexibility and data collection, Alibaba embraces a vision centered around shopper’s digital autonomy.
Even if there is no consensus yet on the vision of what the unmanned store should be, huge investments are being made showing that investors trust this new model and its promises. For instance, AiFi, a California-based startup, specialized in systems to build and operate automated stores are working with major retailers all over the world, (Carrefour, Albert Heijn, Valora & Żabka). At the same time, Saha Group, Thailand’s leading consumer goods manufacturing has committed nearly US$1.7 million to digitize its business and bring the first unmanned store to Thailand. ABI Research forecasted that the total number of unmanned stores would grow from 393 in 2019 to 44,138 in 2023 globally(3).
How will the development of unmanned stores change the retail landscape, does this model worth the cost and is it sustainable in the long run?
Part 1:
The concept of unmanned store is built around efficiency, flexibility and data collection
1.a. Increased Efficiency through the concept of unmanned stores
Scholars from several Universities including the VU in Amsterdam have already examined the effects of waiting times on customer satisfaction. In general, they came to the conclusion that high waiting times will have a negative impact on customer satisfaction and can ultimately lead to a loss of purchasing power (4). Inspired by the long waiting lines, the first express check out was introduced in 1984 (5). It brought sufficient results and already had a positive effect, but not satisfactory enough.
As today's society is developing more and more into a fast-paced society, the demand for as little waiting time as possible is increasing.
The waiting time in supermarkets depends on various factors, including the size of the store, the number of checkouts available and of course the number of customers in the store at a given time (6). Population density also plays an important role. In Germany, for example, an average waiting time of 7 minutes can be expected, whereas in Portugal it is only 2 minutes. This can be traced back to the difference in population but also in the store layout (7).
Although various apps such as supermarket check-in in the UK or Filalindia in Italy have already been developed to monitor the waiting time in your trusted supermarket, this is only partially efficient. Here comes the principle of unmanned stores in play. With the future driven check out system, waiting lines can nearly be eliminated and therefore, create a never seen before check out speed. This will not only increase customer satisfaction but also generate lots of new customers. The most famous example for the efficiency of these stores is Amazon Go. Customers spend on average around 41(8) minutes grocery shopping per week. Whereas, Amazon Go has an average of 27(9) minutes per shopper. This can be traced down to the reduced waiting time but also on the different shopping behaviour. Amazon Go stores have their highest destiny of customers between 12:00 - 13:00. Therefore, the shop is most visited while lunch. To conclude, you can say that Amazon Go is the perfect alternative during times where short waiting times and efficient shopping behaviour are needed - such as lunch. Overall, consumers still prefer to have their “main” shoppings in big supermarket chains.
1.b. Increased Flexibility through unmanned stores
Unmanned stores have the characteristics of offering shopping flexibility. Supported by the Internet of Things (IoT), unmanned stores opened 24 hours a day and 7 days a week, offering the night shift workers great convenience. Besides, unmanned stores enable automatic purchase, by using electronics, software, sensors etc. to collect customers’ data, there is no need for the cashier (10).
The customer shopping habits change during the decade, they buy more frequently and less quantity each time instead of doing a big shopping on Sunday and late evening. According to change in shopping habits, retail channels also change from shopping malls, hypermarkets, to convenience stores (11). The trend is that stores get smaller, offer fewer categories and be closer to the customer.
The unmanned retail follows the development trend to offer the customers more flexibility and more convenience. Unmanned stores are completely self serve and self-checkout, and there are no waiting lines at the cashier. It not only saves time in rushing hours, but also gives the customer more flexibility to purchase whenever they want.
1.c. Data collection: promise of storing and using personal data to personalized assortment of the store and recommend products to the shopper
The third concept behind an unmanned store is the collection of data. Collection of data for what purpose? It is both a prerequisite to make the store run and a valuable asset to improve operational efficiency and make online recommendations Indeed, unmanned store is powered by hard tech (cameras, sensors) and soft tech (algorithms, AI) to track all movements (items, customers) which is needed to create customer basket and prevent stock shortage
There are different types of unmanned stores, but in most of them like in Amazon Go shops, they require an online account to be able to pay online. Through this account it allows the company to have access to a first line of collection data. Basic information such as age, gender, email address, etc(12).
Understanding who is buying in unmanned stores helps to adjust the offer according to gender, age. preferences.
On top of that, unmanned stores are equipped with computer visions and captors that register and analyse customers habits and preferences. First, it allows the store to run smoothly as no employees are present. Indeed, it helps to run the shop in terms of logistics and to replenish the stock(13). For example, captors on shelves monitor the movement of inventory as it is placed on and taken off the shelves and inform the logistic system when the stock needs to be replenished.
Secondly, it allows the company to learn more about their customers' habits and preferences. Indeed, this technology helps to collect data about customers' behaviors such as where they are looking at, what they prefer, who is buying what, etc. Crucial information that allows retailers to better understand customers and improve offers according to demands(14).
The collection and analysis of those data can become a key competitive advantage for a company. It can help to save costs in terms of stock and logistics. It also helps to improve retailers' targeting strategy in order to increase the attractiveness of their stores, as the old dictum goes, “knowledge is power…”
Moreover, collecting this type of data can help to improve the store's future technology and artificial intelligence. Indeed, such feedback can guide experts to create a more efficient and adapted technology. An important variable as technology is the centre of unmanned stores. Deficient technology can ruin the reputation of those shops (15).
Lastly, collection and storage of data involve a big controversial part in terms of privacy. Indeed, unmanned stores' success is mainly achieved thanks to collection of data. Is society ready to accept such an invasion in their shopping habits in the wake of the Facebook scandal about data privacy? A challenge that we will develop in the following paragraphs.
Part 2.
A technological cost hard to support, cultural & legal barriers to overcome jeopardized the development of this model
2.a. Technologies behind the concept of unmanned stores
Technology behind unmanned stores lies in a combination of cutting-edge technology (16) to improve retail business efficiency as well as the customer experience. Unmanned store is running by artificial intelligence to operate and run retail operation, the use of AI enables today’s retailer to obtain large troves of customer data and insights which allow in the long run to identify the ideal combination of in-store items and preferred user experiences, which in turn further optimize and improve their own AI models for such store. Within these 4.0 stores, we also found radio frequency (17) identification, Internet of Things, security camera, facial recognition and image recognition track buyer’s activity that ensure the effective functioning of the store.
This new business model which is booming since 2016 in some tentacular cities starting from Shanghai to Beijing or Singapore and Korea (18) is here to help retailers to shrink their overhead cost and leverage profit margin by reducing staff costs as labor cost average 20 to 35 percent (19) of gross sales but it may vary according to the business scope. The reason behind relies on retailers’ budget expenditure in marketing which is more and more focused on Media and particularly on Social Media (20) at the stake of labor cost, which could be balanced through the right staffing level and sales forecast since workforce remains the main asset for every business and avoiding any reduction.
Nonetheless, running a complex IT system with AI technology and in our case an automated store would require an investment for a 7-Eleven size store from $100 000 to $300 000 (21) presenting some challenges and periodic costs to operate 24H/7. Including the rental space which will be in major cities $50-75 per square foot a year knowing that the average retail shop is around 1 000 square feet(22). Although, retailers can still enjoy some benefits like staff reduction, data collection to have a full picture of the consumption trend to propose a more accurate product selection which will drain more traffic and enhance the store turnover.
Consequently, the overhead cost to run a store with the IT system to always update requires the retailer to have a huge capital expenditure budget as the investment cost is expensive and as most retailers have not been widely using these technologies which are still not mainstream, leading to high technological maintenance costs of unmanned stores where technologies need to be updated over and over (23).
Some external facts which are relevant in this case is to include the high rent space of those stores in some cities like New York, Hong Kong or Paris where the price per square meter can exceed $ 20 000 USD (24). Does it mean that only Big Retailers can afford such a location? When the hardware cost for Amazon average for each store $1Milion according “RBC Mark Mahaney”(25) Consequently, including the diverse operating and maintenance costs as well as renting space cost a lot and retailers still need to break even on their investment.
Another point about customers would also require having the right target since unmanned stores are hard to develop as citizens are not yet accustomed to this shopping experience and more particularly in Europe where Europeans still are in favour of Traditional selling.
2.b. Endangerment of the cashier job
The development of unmanned stores could have a tremendous impact on the job market and there has been legitimate concerns expressed by union workers According to ACS, in the US alone over 3,65 million people currently work in grocery stores. Out of these 2,99 million people approximately 24,9% work as cashiers (27). Germany for instance has 2,95 million people working in retail (28). 25.1 % of these work as cashiers in supermarkets. These two examples show that the cashier job is substantially for the US and German workforce.
According to the U.S bureau of labor, the job outlook for cashiers looks grim. They prognose that between 2018 and 2028 over 138700 jobs in this sector of the industry will be erased due to technological advancement – including self-service checkout stands and increased online sales (29).
Nevertheless, there is hope. Companies like Walmart stated that they will shift 7000 cashier jobs to the sales floor as their stores get more and more self checkout counters. Another positive outlook is the shift towards higher-value tasks and more flexibility according to Roman Zitzelsberger, head of the German IG Metall (30). These higher-value jobs could include being the supervisor for such self checkouts, advisor for customer decision making and in general delivering outstanding customer service.
2.c. Legal barrier: GDPR in Europe, the notion of customer’s approval
Data protection is a stormy subject since numerous fraudulent cases with Facebook and some of the biggest online players have encountered issues with European consumers. Ever since, the European Commission through the Parliament has incorporated since 2016 the “General Data Protection Regulation” (31) as a protection against usurpation or misuse of data, giving citizens back control over their personal data. GDPR is applied within European countries where any firms operate and process their activities using personal data. Data through GDPR can be used whether the user has given its consent, fulfils a contractual obligation, this consent must be minded by an affirmative act, such as checking a box online or signing a form. It is important to note that the users will be able to withdraw their consent whenever they want (32) The only friction point here will be translated that beforehand customers would have to download the retailer application to get access to its unmanned store, raising the question of discrimination (33) for those who are not willing to give their consent by using the retail software to shop. Therefore, what could retailers implement in this case or under which condition to avoid any discriminatory issue.
In our case with a store without staff, prospective customers might be reluctant to be screened through their banking data, facial recognition or fingerprint as those technologies have not been implemented yet in Europe and not as mainstream as in Asia, a region where technology has been on the rise for several years now and where personal data protection has not been legislated as same as the European one however, it is relevant to point that today many countries from middle-East to Asia reinforced their protection in terms of data to strike against the misuse of customers data.(34)(35) By this means, unmanned stores in Europe would require to extend somehow the GDPR including the use throughout the new information and communication technologies and artificial intelligence to be ethics vis a vis of European consumers with strict rules and regulations where technologies failure can occur at any time and lead to high prejudice that could incur cost for both parties.
Part 3
Unmanned store may be relevant in some cases but shall not replace traditional retail
3.a. Removing queue may be achieved by other means which make more sense in terms of customer journey
Queuing is a major pain point when going to a store resulting by losing customers, whether it is waiting for a human cashier or self-checkout where customers have often a hard time with the scan-and-go technology, (36) consisting of a trio of scan, pay, check that is increasingly adopted by retailers via its functionality, which makes the checkout process more fluid, a sensitive point for supermarkets ideal for everyday shopping(37).
Amazon Go has popularized since 2018, cashier less stores all over the US and has influenced huge retailers worldwide such as “AIFI or GRABANGO” (38) who took the veil on this technology and brought some disruption unmet by Amazon like: “sensor fusion and camera web tracking from AIFI” customers would just need to get Apple pay or Google Play app to be scanned at the turnstile enabling to follow their ID whatever the surface within the store tracking their purchase. A study from Utlinous through Retail week has shown that 89% of shoppers left the store resulting from a long queue and among those 65% conceded to have visited a rival store right afterwards to get what they wanted (39).
Consequently, some retailers have integrated in their business model other mechanisms to cool down the overwhelming queue frustration. Like Nespresso, who has implemented in every store virtual queue to consolidate their traffic by using a digital platform to track customers while assigning each of them a customized avatar for ease of identification where these avatars are entered into a virtual queue management system (40) freeing therefore, the customers to wand around the store and increased in the meanwhile its cart. Another way to shrink a long queue as it creates frustration will be the application of a queue management system, (41) which through AI could gather real-time data about the wait time and customers number informing shoppers of their status in the queue (42).Through this process, it will increasingly reduce the customer’s tenseness (43).
3.b. A small store format and the use of a franchise business model will create more advantages in the early stage development of unmanned stores.
Unmanned stores make more sense to keep the store small and use the franchise business model in the initial phase of development as they can somewhat reduce the renting costs, and operating cost as well as technology and hardware maintenance.
Namely, over Amazon Go, Bingo Box a Chinese retailer took advantage of this business model to implement in urban areas, surfaces fitting with city’s landscape that optimize strategic location which is in terms of reproduction less restrictive to copy versus Amazon model that in surfaces required huge space since its smallest and biggest store average 450 to 2700 square feet respectively (44). In contrast, Bingo Box has experienced an exponential growth of its store’s expansion with over 500 stores all over China (45) under its belt aiming to an easy installation and relocation as its biggest surface cover is 140 square feet (46) operating 24/7. Along these lines, we can easily acknowledge that surface allocation of unmanned stores will be the prerequisite to franchise this new business model.
3.c. Brick & Mortar retail is about building a long-lasting relationship, not removing human interaction in-store
Brick-and-mortar retailer refers to a traditional street-side store that offers products and services face to face to its customers such as a local grocery store or the corner bank. In brick-and-mortar stores, consumers can speak to the staff, ask questions about products, and engage with sellers.(47) Face-to-face interaction is proven to comfort us and provide us with some important sense of well-being, whether it’s with friends or friendly cashiers in the checkout line. There is something intangibly real and valuable about face to face interactions that can be highly valued by customers.(48)
In fact, the Timetrade Report found that 90% of shoppers are more likely to buy when helped by a knowledgeable salesperson and 50% of consumers said they value the recommendations by a real salesperson.(49) Indeed, trust and legitimacy are shared commodities for buyers and not applicable in unmanned stores. By investing time and effort to know clients and serving them to the best of your abilities, you can build a successful relationship. This can help framing your reputation and increase awareness but most importantly increase customer base loyalty and thus creating life-long connections.(50)
In conclusion
Throughout this paper, we saw the major advantages behind unmanned stores, such as customer benefit of shopping at any time at “online speed”, or retailer benefit to better understand shopper’s behaviour by collecting data. On top of that, we saw the chance for some companies to expose their technologies and take advantage of this “free” publicity. Indeed, we saw that an efficient and easy to use technology in unmanned stores could change customers' image of the concept. Therefore, a good technology will positively impact the unmanned stores but also the company's technological sales. This is what Amazon started to do by selling the technology used in its Amazon Go stores to the general public (51). Amazon Go may just be a showcase to sell Amazon’s technology.
Along this paper, we also saw key challenges, such as the viability of unmanned stores. Are unmanned stores a revolutionary way of experiencing shopping or just a form of a showroom to expose brands’ new technology and to keep or increase their brand visibility and therefore to draw attention from the general public? Also, from a technology point of view, this concept required a lot of initial and ongoing investment to continue to offer the best solution. A costly investment that not all companies are ready to do. Even if data collection helps to reduce costs, a strong argument since the cost of technology remains high, the profitability of unmanned stores has not been reached yet.
Furthermore, behind unmanned stores, there is a dilemma around data collection and privacy, especially in Europe, where the legislation and even the mentality are stricter. However, as technology and data collection are more and more present in our daily life the debate is already open and legislation is starting to structure this new sector. A point that will help to reduce uncertainty and risk for retailers.
Lastly, there is the issue of customer adoption of this new type of store. We saw that some customers of unmanned stores are attracted by the futuristic aspect of the concept, but will not this craze fade when a new technology is commercialized? How many customers will stay loyal? An essential question in a context where the desire to return to the source with a more simple and sustainable economy is growing, and where social interaction and advice are valued. However, offering sustainable products from a short
circuit is entirely doable, and the human touch can be added with artificial intelligence such as advice like in Casino Four in Paris where a big interactive board helps you to choose a wine according to your taste and meal. However, building loyalty on technology is risky since the competitive advantage can be turned away by a new technology coming up on the market. Building a long-lasting relationship is a more secure approach. To do that, offering complementary channels that give autonomy to the customer is key. Unmanned stores are one model relevant in some contexts (convenience stores, airport, etc.) but not the forthcoming of retail. The future is more on automation of some low value task to free staff for high value task and digital interactions
All those inquiries about the future of unmanned stores and the viability of the project are necessary because a lot of risks and uncertainty are still associated with this new type of stores. The legislation, the society and the way of consuming are not 100% ready with all the aspects that come with unmanned stores, but the majority will agree with the fact that those stores filled an existing need. A need to reduce time spent in groceries, a need to simplify paying systems, a need to have products better adapted to customer demands etc. However, based on the retailers race and investments to become the first one to win market shares in this domain, combine to the craze those past years about new technologies, IoT, AI who completely cross our doorstep like with “Alexa” (Amazon virtual assistant), we can conclude that unmanned stores will be an integral part of the upcoming retailing landscape.
