Amazon Go is Not the Future of Shopping

Amazon Go has been hailed as the “future” of retail, but questions have been raised about its viability - technologically, financially and ethically. 

 

The article will consider whether Amazon Go is going to be the future of shopping and what issues might prevent it meeting its ambitious expectations as a disruptor of the traditional supermarket.

 

 

 

Amazon Go is the brainchild of Amazon, and their first foray into the market for check-out free shopping. It is a completely cashless system, meaning that customers enter and do their shopping, then leave, all without “checking out” their purchases using the traditional “human shop assistant” method. Instead, a combination of automation, robotics and technology allows for a completely cashless experience that needs little human intervention or supervision.

 

Cameras scan and recognise products on sale and customers buy them using a cashless method of payment (your credit card is scanned and charged as you exit the building). An app in your phone, which you must install before you are allowed to enter the Amazon Go premises provides a receipt for the purchase. After you enter the building, you must scan a personalised barcode at a turnstile and then go about doing your shop.

 

Amazon Go cameras then rely on sophisticated image recognition software to ensure that the correct product is always identified, with several 3 dimensional cameras constantly scanning what is placed in baskets so that it can be correctly identified. The system is supposed to be one where customers never have queue again, and don’t even have to open their wallet. 

 

The idea behind this is that automation reduces the money that the retailer must spend on human shop assistants, and this in turn reduces the overall operational costs of the store. Further, the high level of automation reduces waste, error and theft from the store. The risk of armed robberies is reduced to virtually zero as there is no cash in the store at any given time, as such there are no costs associated with recording, securing or moving cash. 

 

Some reports have suggested that there are plans for another 3000 similar shops before 2021. Hot on their heels are competitors like Walmart who are working with Microsoft to devise their own system of checkout free shopping.

 

 

 

Many have expressed doubt as to the ability of Amazon Go stores to really disrupt the traditional supermarket. With approximately 1000 products in each Amazon Go store as it stands, it is hard to see how the technology can be scaled up to cope with the 1000s of products that large supermarkets have in their stores. Supermarkets typically have 80,000-300000 SKUs in any given location, so the capacity of the technology to manage such a large inventory is in doubt. 

 

The technology Amazon Go stores use to monitor purchasing relies on bulky cameras, currently located in the ceilings of Amazon Go stores. These provide a 3-dimensional perspective and a picture of each item is taken before the AI system identifies it and applies the correct charge. However, it has been suggested that this system will find it difficult to tell nearly identical products apart, which could indicate problems scaling up the system, especially to one that oversees larger inventories. 

 

Furthermore, the sheer volume that the AI system required to manage an inventory of 1000 items takes up has caused some concern. Located in the roof of the Amazon Go stores, hundreds of cameras are operating as people enter and leave the store. Maintenance costs for these cameras are likely to be high, and there is always the risk of fire and overheating as the technology is used over longer periods of time. 

 

Others have raised concerns about the level of adaption a building might need to have to accommodate a bulky system of hundreds of cameras installed. Cynics suggest that whereas the Amazon Go model can be commended for dispensing with the need for renting units with enough space for human-manned checkout stations, the space that is saved is surely used up again in the form of the space needed to accommodate the IT systems? 

 

Furthermore, as it stands, human supervisors are required to manage error in the existing camera systems, so all of these costs and risks have to be factored into the debate as to whether Amazon Go stores are going to be the future of shopping. 

 

 

 

The Amazon Go model is premised on hundreds of pictures being taken of every person entering and exiting the store. Amazon then stores these pictures for long enough to identify the products being purchased, before the images are deleted. Credit card information is captured at the point of registration for the app that is required to use the Amazon Go stores, and then credit and debit card information is captured when purchases are made. Further receipts are issued after purchases are made, and the system also records how long an individual has been shopping for (this is printed on the receipt). 

 

With a need to capture and store such large volumes of information that is both sensitive and personal to people, it is not surprising that many are concerned for the security of the information and the potential for the information to fall into the wrong hands. 

 

Amazon Go stores are in their infancy, and have not had to deal with any negative publicity surrounding data breaches, or cyber security issues. Yet, these very issues are things that could damage the operation of the whole Amazon Go approach. It will only take one major cyber security breach for confidence in the whole system to be shaken, or for the system to attract the attention of regulators or even legislators and policymakers concerned for the future protection of the data that is gathered and held for the purposes of cashless shopping. 

 

These issues will have to be managed very carefully, and proactively if Amazon Go growth forecasts are to be realised. 

 

 

 

There has been considerable controversy surrounding the cashless store, owing to the concern that they create problems for a group of people known as the “unbanked”. These are people who don’t have bank accounts and as such they don’t meet the prerequisites for having credit and debit cards for example, an address or an internet connection. Since you need a phone, an app and a credit or debit card before you can use the Amazon Go shop, one of the most central issues is that the cashless store excludes people who don’t use technology. 

 

Disadvantaged people could find it hard to access these facilities, because often you need a good credit history and an address before you can get a bank account, a credit or debit card, or even a phone that can support apps. These seemingly simple requirements exclude people on low incomes, the homeless and the disadvantaged. These people are being referred to as the “unbanked” a reference to their inability to get credit facilities, open bank accounts or get mobile phones or mobile phone contracts. Some estimates place the number of “unbanked” people in the USA alone at 14 million – a figure set to rise. 

 

Representatives of the “unbanked” have been approaching legislators and policymakers in an effort to gain protection for the “unbanked” population. The result have been various legislative bills, some of which have led to cashless stores being banned in certain cities and localities. In the USA New Jersey and Philadelphia are considering whether to place outright bans on cashless stores, in a move that could force Amazon Go to provide human cash tellers for people who prefer to use cash. Some have suggested that prepaid kiosks may solve the problem, since these are in keeping with the cashless ethos, however they allow unbanked people to exchange cash for items that will allow them to participate in the cashless shopping experience.

 

As such, the unbanked population continues to represent a major problem that may impede the ambitious growth forecasts of the Amazon Go business model, and it is fair to say the mainstay of the problems have yet to be addressed in full at operational level within Amazon Go. 

 

 

 

As with any proposition that aims to disrupt an industry, it takes time to understand potential problems lurking in the operational application of ideas. Amazon Go is no different, and while security, third party data storage and ethical problems related to the “unbanked” population are indeed grave concerns and represent challenges that will necessitate changes to the business model, there is no doubt a place for Amazon Go at the table as disruptors of the supermarket industry. The disruption, may however unfold more slowly than planned as problems are identified and systems adapted accordingly.