Amazon disregarded internal warnings on injuries: US Senate probe

For years, worker advocates and some government officials have argued that Amazon’s strict production quotas lead to high rates of injury for its warehouse employees. And for years, Amazon has rejected the criticism, arguing that it doesn’t use strict quotas, and that its injury rates are falling close to or below the industry average.

On Sunday, the majority staff of the Senate Committee on Health, Education, Labor and Pensions, which is chaired by Sen. Bernie Sanders, I-Vt., published an investigation that found that Amazon itself had documented the link between its quotas and elevated injury rates.

Internal company documents collected by Sanders’ investigators show that Amazon health and safety personnel recommended relaxing enforcement of the production quotas to lower injury rates, but that senior executives rejected the recommendations apparently because they worried about the effect on the company’s performance.

The report also affirmed the findings of investigations undertaken by a union-backed group showing that injury rates at Amazon were almost twice the average for the rest of the industry

“The shockingly dangerous working conditions at Amazon’s warehouses revealed in this 160-page report are beyond unacceptable,” Sanders said in a statement. “Amazon’s executives repeatedly chose to put profits ahead of the health and safety of its workers by ignoring recommendations that would substantially reduce injuries.”

Kelly Nantel, an Amazon spokesperson, said the internal studies and recommendations Sanders’ report cited were later found by the company to be invalid. “Sen. Sanders’ report is wrong on the facts and weaves together out-of-date documents and unverifiable anecdotes to create a preconceived narrative,” she said.

She noted a recent ruling by a judge in Washington state that rejected a regulator’s allegations that Amazon required employees to work at an unsafe pace, and she said the injury rates have recently improved. “The facts are, our expectations for our employees are safe and reasonable,” Nantel said.

In an internal study known as Project Elderwand, which Amazon began in 2021, Amazon health and safety personnel identified an upper limit on the number of repeated motions a worker could make while picking items from robotic shelving units before injury rates increased substantially. That limit was equivalent to picking about 216 items per hour over a 10-hour shift. They found that Amazon workers, responding to productivity quotas, typically went well over that limit — picking more than 266 items per hour.

The study recommended that Amazon use software to track workers’ pace and require additional breaks to limit the number of repeated motions and keep workers under the threshold.

In another internal study, known as Project Soteria, which began in 2020 and continued through 2022, Amazon investigators found evidence that a faster pace of work led to a higher rate of injuries. The study recommended that Amazon suspend discipline for employees who failed to meet their productivity targets and that it give employees more time off from work, both of which appeared to reduce injury rates.

But Amazon executives ultimately rejected the recommendations of both reports, according to documents uncovered by Sanders’ office. The Senate labor committee found evidence suggesting that executives were concerned that carrying out the recommendations could lower productivity at Amazon’s warehouses or hurt the “customer experience.”

Sanders’ report also uncovered a third internal Amazon study that took issue with Project Soteria. In it, another team of Amazon researchers disputed the idea that there was a link between the pace of work and the risk of injury. It concluded that some workers are inherently more prone to getting injured.

Amazon said Sunday that Project Soteria was an example of its practice of assigning multiple teams to study safety: One team explored a potential link between speed and injuries and another evaluated the methodology and findings and concluded that they were flawed.

The findings in the Senate report are consistent with investigations conducted by state and federal regulators in recent years. Last year, the federal Occupational Safety and Health Administration cited more than half a dozen Amazon warehouses for exposing workers to high risks of joint and soft-tissue injuries, including back injuries.

The agency has said the elevated injury risks were related to the high frequency with which workers lifted items, the heavy weight of the items, the awkward body movements required to lift the items and the long hours that employees worked. OSHA proposed fines of more than $100,000 across the warehouses it cited; Amazon has appealed the citations.

In California, regulators fined Amazon nearly $6 million this year for violating a law that requires companies to provide written disclosures of quotas, and which forbids quotas that prevent employees from following health and safety laws or taking state-mandated breaks. Amazon said it appealed the citations.

Regulators in Washington state cited Amazon for safety violations earlier this decade, but a state judge threw out several citations this year after a monthslong trial.

The state regulator referenced Project Soteria in its case, but the judge found that the state did not sufficiently establish a relationship between the pace of work and injury rates. The regulator is appealing the decision in state court.

Amazon has said that it has spent hundreds of millions of dollars improving safety in recent years, and that injury rates have declined as a result, including a large drop for the most serious injuries. The company has long maintained that it doesn’t have strict or “fixed” quotas. It says it has performance targets that are evaluated over longer periods and that take into account factors beyond sheer productivity, like an employee’s experience level and how other workers at the site are performing.

But employees have said for years that they are subject to warnings or disciplinary action if they fail to complete a certain number of actions per hour, and interviews conducted by Sanders’ office affirm this. Amazon workers told investigators they could be disciplined for failing to pick items from shelving units at target rates in the hundreds per hour.

The report also identifies what it says are flaws in how Amazon compares its own injury rates with the rest of the industry. While Amazon says its injury rates are roughly average for large warehouses, Sanders’ team argued that this calculation was heavily skewed by including Amazon in the overall data set, which drives up the average. Amazon also tends to restrict the comparison to warehouses with 1,000 or more employees even though it operates many smaller warehouses.

When Amazon is removed from the average and compared with other companies, and when the analysis includes warehouses of any size, its injury rates are more than 1.8 times that of other companies in each of the past seven years, Sanders’ report concludes. The findings are similar to those of a union-backed group.

Amazon defended its methodology, saying that benchmarking against the overall industry average was standard practice.

Sanders’ report also found that Amazon makes it difficult for workers to receive appropriate medical care when they are injured. It found that Amazon often discouraged workers from seeking medical attention outside their warehouse and sent them to an internal health facility that was not equipped to provide more than first aid, even when they had potentially serious long-term injuries. It said the company frequently denied workers the accommodations they needed to deal with injuries sustained on the job.

Amazon has denied discouraging workers from seeking outside medical attention and has said that its accommodations policies meet or exceed state and federal requirements.