Amazon’s satanic mills

Amazon’s executives need to think hard about the human costs of their business model. Photo: Reuters

With Britain suffering through its worst cost-of-living crisis in decades – owing to high inflation and soaring energy prices – hundreds of workers at an Amazon warehouse in Coventry this month demanded a wage hike. If the demand is not met, they say they will go on strike in November, just ahead of Black Friday and the holiday shopping season. As with other recent labor actions by US rail workers and British Royal Mail employees, the Amazon workers' move has kicked off a debate about who is to blame for the threatened disruption: the elves in the workshop or Father Christmas?

Amazon owes its success to a variety of factors, including a sophisticated data-driven approach. But its real genius lies in its logistics breakthroughs – including route optimization, fleet planning, and metadata management – that allow it to minimize "click-to-ship" time and provide customers with unprecedentedly fast and reliable on-time deliveries. Amazon Prime-branded planes and trucks shuttle packages around the world, operating like clockwork even through a pandemic that grounded much of the rest of the economy.

The mastermind behind the operation is a man named Jeff Wilke, who combined Taylorism (dividing production into narrow, closely monitored and measured repetitive tasks) and Fordism (assembly-line techniques) to create a warehouse model capable of processing more than a million units per day. With the help of robots and close surveillance, human "pickers" and "stowers" now process several times as much merchandise per hour as they once did.

But the system has become notorious for testing human employees' limits. Recent investigations have shown that much of the convenience that Amazon customers enjoy comes at the expense of Amazon's lowest-paid workers.

For example, last year the New York Times found working conditions at Amazon's New York "fulfillment center" to be utterly Dickensian. After passing through airport-style security gates, workers say they are subjected to hard physical labor, long shifts (10.5-12 hours), and a high incidence of injuries and accidents (double the rate of non-Amazon warehouses). Compounding the indignity, all are closely monitored by a dystopian surveillance system that punishes infractions like talking to co-workers or missing productivity targets (which are often as high as processing 30 packages per minute or requiring a minute in total to unshelve, box, and ship an item).

The threat of being fired – or what the company calls being "released" – looms large, and workers who seek recourse through human resources run into a Kafkaesque systems that specializes in stonewalling, especially when it comes to requesting disability leave or pay. Horror stories include Amazon drivers having to urinate into plastic bottles or defecate into plastic bags to stay on schedule. There have been reports of workers selling their wedding rings or relying on food stamps to make ends meet. In response to these accounts, the company has offered ham-handed corporate responses like "meditation rooms" that resemble large coffins.

No wonder unionization efforts at Amazon facilities have been growing. Despite the company's systematic efforts to suppress organizing, a union drive at an Amazon warehouse in Staten Island succeeded earlier this year, following a narrow loss for a similar effort in Alabama. In 2018, US Senator Bernie Sanders introduced the Stop Bad Employers by Zeroing Out Subsidies ("Stop BEZOS") Act, which would tax companies for 100% of the public government benefits they receive. And now, the US Occupational Safety and Health Administration has opened investigations into working conditions at Amazon.

These skirmishes have fatally undermined the narrative that Big Tech tells about itself. Amazon may be a logistics pioneer, but it is no less reliant on worker exploitation as the "satanic mills" of the First Industrial Revolution. According to Amazon's origin story, it all started with Jeff Bezos selling books from his garage and ringing a bell every time an order came in. Yet even in the early days, there was an incipient culture of overwork (employees were expected to put in at least 60 hours per week), rule-bending, hazardous workplace conditions (unpackaged knives falling off conveyer belts), and Orwellian performance monitoring.

Amazon is now one of the world's biggest companies. But, as I've argued elsewhere, bigger is not always better. While some of its practices could be framed as innovative and adaptive when it was much smaller, today it systematically reduces employees to data points. Before stepping down as CEO last year, Bezos viewed employee churn as more of a feature than a bug of Amazon's model. To have an entrenched workforce, he reportedly said, was to "march to mediocrity." Hence, the company reports an employee turnover of roughly 150% per year – twice the industry average – meaning that its entire workforce is replaced every eight months.

This model is not only unethical and inhumane; it is also probably unsustainable. Studies show that happy workers are more productive. And, as an internal company memo warned earlier this year, "If we continue business as usual, Amazon will deplete the available labor supply in the US network by 2024."

With an estimated net worth of around $140 billion, Bezos was the world's richest man from 2017 to 2021. He has clearly become as disconnected from the ordinary Amazon employee as the scale of his wealth would suggest. As one worker put it in 2020, "I'm sure Mr. Bezos couldn't do a full shift at [the New York City warehouse] as an undercover boss."

The Coventry workers who are demanding a cost-of-living adjustment would doubtless concur. Amazon's executives need to think hard about the human costs of their business model. If they need a quiet place to consider the issue, they can always try one of the meditation coffins.

Antara Haldar. Sketch: TBS

Antara Haldar is Associate Professor of Empirical Legal Studies at the University of Cambridge.

Disclaimer: This article first appeared on Project Syndicate, and is published by special syndication arrangement.