Algorithms have usurped Amazon’s retail decision-makers.
The e-commerce company once relied on humans to predict demand of certain products, such as anticipating and ordering a glut of the season’s hottest toys ahead of the holiday season. But a report from Bloomberg June 13 shows that the decision-making process has slowly transitioned towards automated ordering and communication with manufacturers, leaving humans in the lurch.
This trend isn’t surprising at an automation-minded company like Amazon, but it’s indicative of a wider trend in analytics-based jobs: The algorithms are coming. Whether it’s insurance adjusting or product buying like Amazon’s workers, there could be software that does an increasingly better job for a lower cost than a human salary.
The transition within Amazon is credited to two forces. The company launched a pilot project called “hands off the wheel” that automated demand forecasting and negotiating prices with vendors, and the rise of the Amazon Marketplace, where vendors can bypass Amazon’s buyers and sell their wares directly on the site on their own.
Amazon’s “hands off the wheel” project started several years ago, according to Bloomberg, and the algorithms originally could be overridden by humans if they thought there was a mistake. But as the algorithms proved their worth, it became necessary for employees to justify why they were overriding the software.
And Amazon Marketplace gave the opportunity to streamline the interaction between vendors and the company, by using similar algorithms to automatically negotiate and buy products based on deals vendors were willing to offer.
“Computers know what to buy and when to buy, when to offer a deal and when not to,” Neil Ackerman, a former Amazon executive who now manages the global supply chain at Johnson & Johnson, told Bloomberg. “These algorithms that take in thousands of inputs and are always running smarter than any human.”