Milestones in Amazon's OPSCM Infrastructure Resource Design

Since its inception in 1994, Amazon has implemented a number game-changing OPSCM infrastructure resource designs, in particular in the area of strategic capacity management, that have been instrumental in allowing it to maintain the supply throughput necessary to keep up with its spectacular rate of growth in sales. The following memo specifies four key implementations: Flexible/As needed construction, warehouse robotics, network layout, and innovative capacity management solutions.

Build Capacity as Needed/Maintain a Flexible and Simple Design

Amazon carefully tracks its facilities' capacity usage and uses a build more capacity as needed strategy. It utilizes a rapid build cycle model that focuses on keeping the warehouse design relatively simple and flexible so that it can build new facilities or add on to existing ones quickly. At a 2012 conference, Mike Roth, then VP of NA Fulfillment Services, explained how Amazon includes capacity forecasts in its S&OP process that are largely based on density maps with current Amazon distribution centers pinpointed, as growth in demand is largely found in areas of high-density population (Banker, 2012) The aim is to be able to expand and contract capacity as nimbly and quickly as possible in response to significant fluctuations in demand. The Introduction of Kiva robots in 2014 and more progressively automated systems in fulfillment centers and DCs has made this flexibility more challenging, however, their introduction has by far made up for it in added capacity utilization and overall efficiency.

Kiva Robotics

Amazon started implementing the use of Kiva robotics system in 2014, and by 2016 it was estimated that they had cut operating costs by 20% and the overall fulfillment cost by 40% in the fulfillment centers that were using them. They are essential to optimal capacity utilization as they increase efficiency multifold: they can find, sort, pick, and pack an order in one-fifth the time of a human (Kim, 2016), however they are most frequently used to carry shelves of inventory to their human-counterparts who can stay then stay in one place and focus solely on the picking and packing, shortening the time to fulfill an order immeasurably. The robots also take up much less space. They are much smaller than people. Thus warehouses can be modified to fit more shelf space and carry more inventory: "Each Kiva-equipped warehouse can hold 50% more inventory per square foot than centers without robots" (Bhattacharya, 2016). The robots run on machine learning and continue to evolve to do more and more advanced tasks.

Facilities Design and Network Layout

As of Q1, 2023, Amazon has 2,373 active facilities worldwide. The majority are fulfillment centers and distribution centers where the items are sorted, packed, and sent on to the smaller delivery centers and hubs that are located much closer to the customer to receive the optimal "last-mile delivery" that Amazon is now famous for. The sheer number of smaller facilities, either Amazon owned or service partners, that then feed the delivery drivers, whether their own or third parties, has been and continues to be optimized, so that the smaller last-mile delivery nodes and modes have become a very significant factor in Amazon's ability to move from the order click to final delivery in record time. Amazon manages a constantly changing, multi-echelon inventory management system, and this network layout allows for great flexibility in how product is broken up and redistributed through its network.

Innovative Capacity Management Ideas: A Capacity Management System for Third Party Sellers

Amazon has led innovation of strategic capacity design. In 2021, Amazon found itself with large excess capacity, having rapidly expanded its facilities to keep up with the pandemic-inspired explosion in e-commerce only to see demand slowdown in 2021. Amazon halted expansion of many facilities and sold some other. But they also found an innovative way of using their excess capacity: they implemented a new capacity management system for third party sellers where the sellers can bid for extra storage space in an Amazon facility. Before then, third party sellers did not have the option to store inventory for their products in an Amazon space. This saves the third-party seller costs associated with having to use storage elsewhere and then move product to Amazon fulfillment centers for processing, but it also means that Amazon can essentially keep track of all excess capacity and rent it out to third party sellers rather than carry the costs of unused storage space.

Conclusion

Though this memo focuses on Amazon's innovative OPSCM infrastructure optimization, it is impossible to not weigh in on the proprietary digital platforms that integrate, support, and run all of their infrastructure. Amazon is an example of an enterprise whose digital analytics and machine learning capabilities are so fundamental that in many ways they precede the actual infrastructure designs and operations. The two are coupled together and continue to be optimized hand in hand. One particularly apt example is the Amazon Warehouse and Distribution App (AWD) which not only underlies Amazon's warehousing, inventory, and distribution capabilities, but which is now directly available to third party sellers, offering them the dual package of physical capacity coupled with digital capacity.

References

Banker, S. (2012, December 3). Amazon plans carefully its distribution capacity growth. Logistics Viewpoints.https://logisticsviewpoints.com/2012/12/03/amazon-plans-carefully-its-distribution-capacity-growth

Bhattacharya, Ananya. "Amazon Is Just Beginning to Use Robots in Its Warehouses and They're Already Making a Huge Difference." Quartz, Quartz, 17 June 2016,https://qz.com/709541/amazon-is-just-beginning-to-use-robots-in-its-warehouses-and-theyre-already-making-a-huge-difference

Garland, M. (2022, September 14). Amazon cancels, delays dozens of warehouses as it looks to rightsize capacity. Supply Chain Dive.https://www.supplychaindive.com/news/amazon-cancels-delays-dozens-of-warehouses-fulfillment-mwpvl/631130/

Garland, M. (2023, January 19). Amazon's New Capacity Management System lets sellers bid for more storage space. Supply Chain Dive.https://www.supplychaindive.com/news/amazon-fulfillment-fba-new-capacity-management-system-seller-storage/640777/

Garland, M. (2023, February 28). Amazon closes, cancels more warehouses as cost-cutting persists. Supply Chain Dive.https://www.supplychaindive.com/news/amazon-warehouse-closures-cancellations-delays-2023-mwpvl/643623

Kim, E. (n.d.). Amazon's $775 million deal for robotics company Kiva is starting to look really smart. Business Insider.https://www.businessinsider.com/kiva-robots-save-money-for-amazon-2016-6

Roser, C., Kevin, Raam, & John. (2022, March 22). The inner workings of the Amazon Fulfillment Centers Parts 1 - 6AllAboutLean.com.https://www.allaboutlean.com/amazon-fulfillment-6/

Wulfraat, M. (n.d.). Amazon Supply Chain and Fulfillment Center Network. Amazon Distribution Network Strategy | MWPVL International.https://www.mwpvl.com/html/amazon_com.html

"How Amazon Leverages Last-Mile Delivery for Its Logistics?" JungleWorks, 23 June 2022,https://jungleworks.com/how-amazon-leverages-last-mile-delivery-for-its-logistics/

Questions

1. What are some OPSCM infrastructure design (nodes, modes, lodes) changes or adjustments that companies adopted in their in response to lessons learned during and/or post pandemic?

2. What could be one possible next evolution in Amazon's OPSCM infrastructure design?