As Covid-19 spread in 2020, homebound customers turned to Amazon at an unprecedented clip. Orders neared that of the holiday season and the company was short-staffed and often out of stock on key items, pushing delivery windows from 2 days to weeks on some items. Founder Jeff Bezos greenlighted a strategy, guided by a revered internal forecasting tool, that overshot the long-term projections for demand. Instead of a permanent shift in consumer behavior, the pandemic-fueled growth in online shopping has slowed as in-person shopping has bounced back.
Early in the pandemic, Amazon opened hundreds of new warehouses, sorting centers and other logistics facilities, and doubled its workforce from 2020 to 2022, to more than 1.6 million people. But demand hasn’t kept pace with that planned capacity. Now new CEO Andy Jassy is cutting back the excesses. He is subleasing at least 10 million square feet of warehouse space, deferring construction of new facilities, and finding ways to end leases with outside warehouse owners. Jassy has also abruptly closed down the company’s bricks-and-mortar retail operation—68 stores—and is paring back its bloated head count.
Part of Amazon’s e-commerce challenges today, writes The Wall Street Journal (June 17, 2022), stem from a piece of technology long prized as a secret weapon, an internal forecasting system called Supply Chain Optimization Technologies, or SCOT. It was designed to incorporate a multitude of factors and spit out projections for product demand and the growth in logistics needed to fulfill it.
SCOT forecasts produced low, medium and high estimates. Because of unprecedented volume in the early days of the pandemic, Amazon repeatedly chose the higher end of SCOT’s estimates. Those estimates meant that the company needed many more fulfillment centers and other infrastructure to keep up. So Amazon aggressively built out new warehouses and transportation hubs, and went on a hiring spree to get customers their packages. But the forecasting technology wasn’t equipped to process an unforeseeable event like the pandemic and caused the company to commit to building infrastructure early in the pandemic that take 18 months to 2 years to come online. When the virus receded, Amazon was left with more planned capacity than orders. After being understaffed for 2 years, the company was suddenly overstaffed.
Classroom discussion questions:
- Using Chapter 4 terminology, what type of forecasting system did Amazon employ?
- What could the company have done differently, in hindsight?