Technology Analysis of “Amazon Go” – The Smart Shopping Experience

Amazon has a history of defining the customer experience when it comes to shopping. Apart from its famous one-click patent that changed the online shopping experience, amazon is now ready to change the offline shopping experience for its customers.

Recently, Amazon launched a video showing its concept retail store under the brand name “Amazon Go” which will provide the “just walk-out” experience to the customers which means customers will not have to pass through long checkout lines and payments shall be automatically calculated and deducted from the account of the customer.

It will not be wrong to say that the concept of “Amazon Go” is somewhat similar to the one-click patent of the Amazon that is going to expire this year.

The concept video seemed very fascinating and took the entire world with a storm. It also left us very curious and we tried to find out how Amazon will make this store practically possible.

Techniques to Identify Customers Entering the Store

The Amazon Go store will be equipped with various data-gathering devices that will track the activities of the customers inside the store. The items picked by the customers inside the store premises will be identified based on the data received from a variety of data gathering devices.

Amazon store will also be able to identify that whether the item which was earlier picked by the user is placed into inventory or shelf again or not, thus it will make the user experience more intuitive.

The data gathering devices used by Amazon in its smart store will be the camera, RFID tag, a volume displacement sensor, pressure sensors, infrared sensors, load cells, Safety Light curtains and/or a Bluetooth tag.

A RFID reader, will be located at or near the transition area and collect data (e.g., images, RFID tag identifiers) that will be processed to determine that a customer is entering and/or passing through the gate of the Amazon Go store. Alternatively, load cells transducers (e.g. strain gauge load cells) may be positioned on the surface of the transition area to detect when a customer is passing through the gate of the store.

Techniques To Identify Customers & Their Locations Inside The Store

Once the presence of the customer inside the store is detected, customer data will be collected by one or more of the data gathering devices located in the Amazon Go store and will be used to determine the identity of the customers. A customer may also be identified using facial recognition techniques and/or a customer ID which is used to enter in the store. Additionally, other customer identification methods such as QR code scanning or providing a unique pin representing the customer can also be utilized for identifying the customer. Upon identifying the customer, customer information such as customer profile date, purchase history, and potential pick list associated with the customer will be used to support further analysis.

Further, the camera located within the store will collect images of the store that will be used to detect the presence of the customer inside the store near a specific item on the shelf. Amazon can also collect customer’s GPS information using his smartphone to determine the exact location of the customer inside the store.

Techniques To Detect Whether A Product Is Picked And/Or Returned Back

For this store to become practically possible, it is very important to detect whether a product is picked from the shelf. Also, the product which was earlier picked by the customer can be returned back on the shelf and the detection of this step is also very essential for the working of the Amazon Go model.

To achieve the above-mentioned objectives, Amazon may also make use of image detection for determining whether the product is picked or not by the customer. For example, images of the customer’s hand can be captured and image analysis can be performed to detect whether the customer has picked any product from the shelves or not. Amazon Go store will also be able to detect the skin tone color of the user’s hand to enable accurate detection of the products.

In addition to image analysis, Amazon may use a pressure sensor and/or weight sensor to detect when an item is picked and/ or returned to the shelves.

Also, an infrared sensor placed near the shelf may also be used to distinguish between a user’s hand and items.

Amazon may also use light curtains or other presence detection devices to detect the customer or when an item is picked and/ or returned to the shelves.

Alternatively, Amazon may also attach a RFID tag / NFC tag / Bluetooth tag to the exterior of a product to determine whether the customer has picked the product or not.

Techniques to Calculate the Amount of Purchase

After determining that the product/item has been picked by the user, a unique identifier representing the picked item will be linked with the customer profile to represent that the customer has picked this item. Similarly, if it the customer has placed the picked product/item on the shelf again then the unique identifier representing the product/item will be removed from the customer’s profile data.

The customer profile data store may be updated regularly to update the list of the items/products that are picked by the customer. Customer will also be able to see the details of the items picked by the customer on the Amazon App installed on his smartphone in real-time.

Unanswered Questions Revealed

1. Still, there has been many questions which people are not able to find answers, such as if a customer enters the store with his/her friend how Amazon will determine the total shipping amount?

We have also tried to find out the answer to this question, and we feel that Amazon will be able to associate different people using various detection devices who are related to each other or who have entered the store using one App/identity. So, if your friend enters in the Amazon Go store with you using your App/ID then Amazon will track products picked by you as well as your friend and will create the final checkout list accordingly.

2. What if some products are picked by store employees and handed over to the customers?

In such cases, Amazon Go store will be equipped to differentiate the customers from the store employees, and the items that are picked by the store employees and are handed over to the customers will be identified and added to the final billing of the customer.

3. What if amazon is not able to identify which item is picked by the customer?

In such a scenario, Amazon may make use of the customer’s past purchase information to assist in identifying the currently picked item by the customer. Also, even if it is not identified that which product or how many products were picked by the customer, in that case, the data may be provided to the store employees who may review the images/other data and confirm the identity of the item picked by the customer.

Wrapping Up Things

Brands have become part of a rat race and everyone wants a better hold over the market. As a part of the strategy, numerous new technologies are introduced for enhancing the user experience and for attracting more and more customers towards their brand. “Amazon goes” seems to an innovative and fresh approach towards revolutionizing the customer experience. It remains to be seen that the practical implementation of the concept turns out to be equally exhilarating.