Look but Don’t Touch — Amazon’s Supply Chain Strategy

“But why can’t I touch it?” the child wailed plaintively while tugging on their parent’s hand.

“Because if everyone touched it, there wouldn’t be anything left”, the parent patiently explains.

This simple exchange may be familiar to each of us as either the child, the parent, or both. It is certainly true that the oils on human skin can become detrimental to works of art, organic materials, and even stone over time. We have therefore come to the consensus that added handling reduces the value of an object. Nowhere is this more understood than in the supply chain. Within the classic understanding of the supply chain, movement of goods, without converting or changing them, is considered a non-value-added activity.

Allow me to propose that the right movement is the key to adding value. I would also submit that this is the essence of Amazon’s overall business approach leveraging advanced systems engineering, data analytics, machine learning, and Artificial Intelligence (AI). To better understand this, let us look at the example of a relatively mundane product, the athletic shoe.

Athletic shoes are almost universally produced in the countries on the western edge of the Pacific. These shoes are manufactured by the millions and yet are barely purchased in the region. It is the ability to transport these products across the globe, to markets where they are valued, which allows for the economic fabrication of a very light, mundane item on the other side of the world.

If the shoes are not sent to the right market in the right quantity, their value can fall precipitously. Additionally, the seasonal aspect of these shoes means that an older shoe is immediately less valuable to the consumer and thus to the market.

Amazon is striving to create a supply chain with near anticipatory capabilities. It seeks to use enormous computing power and consumer data to predict consumer needs and position themselves with the shortest supply chain in terms of time and in terms of handling. If we continue with our running shoe example, Amazon may subscribe to metadata that is related to Point of Sale (POS) purchase of running shoes in a region as well as their own datasets on sales. They may also then examine metadata on fitness and running applications used by runners to better understand the usage of the shoes.

We have all experienced that moment when searching online that several strangely coincidental suggestions pop up in our social media or search feed. This is the probing of someone’s AI attempting to determine if you have an interest in that item. Each electronic footprint or fingerprint we leave about running shoes, whether on a social page, online retailer, fitness app, or through a point of sale can be aggregated into trends and then subdivided by region to help Amazon drive the right inventory at the right time for their consumers.

If we have chosen to “opt-in” for more personalized data usage, we may find that very specific data is analyzed to deliver a more customized experience for us as a running shoe consumer. We may get notices as our shoes begin to degrade, information on our recent running trends and how to run more comfortably or other levels of engagement meant to tie us to Amazon in a beneficial, symbiotic manner.

All of this requires an enormously connected and efficient operation. In an ideal environment, the system would understand our running behavior, anticipate our needs, and place an order for new shoes automatically. Additionally, it would combine that order with other items coming from the manufacturing region to minimize costs and maximize the cube in a container. These may be wildly dissimilar products, sometimes referred to as a calico shipment. Their common aspect is that they are needed in the same region at the same time.

Once this order is dispatched, the accuracy of the predictive system permits the entire container to be shipped to the destination where only then is it opened and items are removed for final transport to the consumer, optimally via autonomous systems or crowdsourced transports. In this manner, Amazon leverages its expert systems to reduce the number of touches as low as possible. In a normal supply chain, the system’s inefficiency would require buffer stock, operates from outdated sales forecasts, and disconnected demand signals. No matter the sophistication of the physical handling of the product, the wrong items in the wrong place have essentially zero value and in fact, represent a loss to a business.

Amazon’s incredible advantage starts with its obsession over the customer. They have invested heavily in research, including predictive technologies, like those being developed by Ping Xu, who holds an operations research Ph.D. from MIT, forecasting science director within Amazon’s Supply Chain Optimization Technologies (SCOT) organization. Her role is forecasting for both the business and the overall supply chain. At Amazon, systems engineers are aggressively recruited, with average salaries north of $140K. Their role is to accelerate the capabilities, design and validate the supply chain of the future. One in which the needs of the customer are paramount and the product may arrive without ever being touched by human hands.

References

Beck, C. (2014). Predictive Policing: What Can We Learn from Wal-Mart and Amazon about Fighting Crime in a Recession? The Police Chief, 7.

Calvo, D. J. (2019, May). An analysis of Amazon’s digital supply chain. Retrieved from The Future Factory: https://www.thefuturefactory.com/blog/30

Clark, G. (2019, April 19). Inside Amazon’s AI-powered supply chain. Retrieved from Retailbiz: https://www.retailbiz.com.au/latest-news/inside-amazons-ai-powered-supply-chain/

Daisuke Tsukamoto, E. P. (2020, June 09). Optimizing Supply Chains Through Intelligent Revenue and Supply Chain (IRAS) Management. Retrieved from AWS Partner Network (APN) Blog: https://aws.amazon.com/blogs/apn/optimizing-supply-chains-through-intelligent-revenue-and-supply-chain-iras-management/

Fuse Inventory. (2020, February 01). UNLOCKING AMAZON’S SECRET TO SUCCESS: SUPPLY CHAIN. Retrieved from Fuse Inventory: https://www.fuseinventory.com/blog/unlocking-amazons-secret-to-success-supply-chain

Gantenbein, D. (2020, June 05). Predicting what Amazon’s customers will need tomorrow, next week, and beyond. Retrieved from Amazon Science: https://www.amazon.science/working-at-amazon/predicting-what-amazons-customers-will-need-tomorrow-next-week-and-beyond

Grand View Research. (2018). Trail Running Shoes Market Size, Share & Trends Analysis Report By Type (Light, Rugged, Off Trail), By Distribution Channel (Online, Offline), By Region, And Segment Forecasts, 2019–2025. San Francisco, CA: Grand View Research.

PayScale Inc. (2021, March 15). Average Systems Engineering Manager Salary at Amazon.com Inc. Retrieved from PayScale: https://www.payscale.com/research/US/Job=Systems_Engineering_Manager/Salary/3658a311/Amazon.com-Inc