At Amazon, our employees’ safety is and always will be our top priority—nothing is more important to us. While we respect Sen. Sanders and his work chairing the Senate Committee on Health, Education, Labor, and Pensions (HELP), the senator has issued another report that’s wrong on the facts and features selective, outdated information that lacks context and isn’t grounded in reality. Our voluntary, good-faith cooperation with this investigation was premised on the reasonable expectation that any report would be even-handed and truthful, even if that truth was inconvenient for people who want to claim that our workplace is anything other than safe. We’re proud of the progress we’ve made and our commitment to continuously improving, and we were eager to share that progress with the Committee. Unfortunately, it’s now clear that this investigation wasn’t a fact-finding mission, but rather an attempt to collect information and twist it to support a false narrative. Here are some actual details about our record and the false claims in this report.
The main premise of the report is fundamentally flawed. The title of the report is “The ‘Injury-Productivity Trade-off’: How Amazon’s Obsession with Speed Creates Uniquely Dangerous Warehouses.” If that were accurate, what you’d see is that as our productivity and speed goes up, injuries go up. But what’s actually happened over the past five years is exactly the opposite – we’ve increased our delivery speeds, while decreasing the injury rates across our network. How is that possible? Because speedy delivery doesn’t come from pushing people harder – it comes from getting products closer to customers and reducing the number of steps involved in going from a supplier to a customer. We’ve spent years re-designing our network to do just that.
Our safety record continues to improve. As we have publicly disclosed and discussed with committee members during this investigation, we’ve made, and continue to make, meaningful progress on safety across our network. And unlike most companies, we provide a public report each year with our safety data. Our most recent report shows that, from 2019-2023, we reduced our recordable incident rate (which includes anything that requires more than basic first aid) in the U.S. by 28%, and we reduced our lost time incident rate (which only includes more significant injuries that require an employee to miss at least one day of work) by 75%.1 Contrary to the narrative in this report, these large safety gains occurred alongside surging demand, unprecedented organizational growth, and during a global pandemic.
The report also accuses us of having safety policies in place but not following them, which is hard to square with our significant progress: strong policies—and adherence to them—are helping us create a safer work environment every day, and that’s clear in the data. It’s also clear to objective observers like the judge in Washington state who heard multiple weeks of evidence and dismissed similar claims against us earlier this year, also acknowledging our robust safety programs and employees’ engagement on safety.
While we’re encouraged by the progress we’ve made, we aren’t satisfied and will continue to invest and invent to be the best and safest workplace in the industries in which we operate.
We benchmark ourselves against similar employers because it’s the most effective way to know where we stand. We often use the industry averages from the Bureau of Labor Statistics (BLS) to see how we compare to other employers on injury rates. To do that, we compare ourselves to two categories – one for warehousing (where we compare rates from across our large warehouses) and one for transportation (where we compare rates for the transportation portion of our network, which typically includes smaller buildings).
In warehousing, we compare ourselves against the average for employers with over 1,000 employees. The senator alleges that it’s improper for us to benchmark against employers with more than 1,000 employees because many of our buildings are smaller than that. But what he misses or ignores is that we have on average 1,400 employees at the buildings that are included in this particular comparison. As a result, it makes sense to compare ourselves to employers of that size, and putting ourselves in a different category would be misleading.
The senator then goes on to claim that our safety record for 2023 is “far more dangerous” than the industry, which is again wrong. In fact, Amazon’s RIR in the transportation-focused Courier and Express Delivery category was significantly less than the 2023 BLS industry averages. Amazon’s Courier and Express Delivery RIR rates for 2023 were 6.3, while the industry average (regardless of number of employees) was 9.7. Amazon’s facilities are not “uniquely dangerous” as Sen. Sanders continues to claim.
It's wrong to rely on analytically unsound documents like the Project Soteria paper in any objective report – yet, it’s relied on here while disregarding and minimizing the evidence that shows Soteria is unsound. If this report were seeking the facts, it wouldn’t include rely on an outdated, inaccurate document called Project Soteria that Amazon never used (a fact that we made abundantly clear to Committee staff), and that a judge in Washington state evaluated before ruling that the claims that Amazon’s pace of work is hazardous weren’t supported by the evidence.
Among other things, Project Soteria examined whether there’s a relationship between employee performance feedback systems and injury rates. However, Soteria’s methodology and hypotheses were evaluated by expert economists at Amazon, and they determined that the project was flawed and inaccurate. Nonetheless, Sen. Sanders and his staff chose to rely on the debunked Soteria analysis because it fits the false narrative he wanted to build.
This report also disregards the results of our aforementioned case in Washington state – even though the senator said the citations in that case were one reason he was opening his investigation in the first place. So, we’ll share the result: after six weeks of testimony from employees, ergonomists, and other experts, a Washington judge ruled in Amazon’s favor and vacated the citations issued to Amazon by the Washington Department of Labor and Industries (L&I), including one that alleged we “willfully” put employees at risk from ergonomic hazards. In addition to hearing testimony on why Soteria was flawed and invalid, the judge ruled that L&I did not establish that pace of work at Amazon was hazardous. The judge’s decision was also upheld on appeal.
Disregarding this ruling merely illustrates that the outcome of this investigation was decided when Sen. Sanders opened it, and no amount of facts that prove otherwise would change his mind.
Project Elderwand shows how we regularly examine our safety processes to ensure they’re as strong as they can be – nothing more, and nothing less. Our ergonomists regularly review our processes to make sure they’re effective and align with ergonomic best practices. That pursuit of continuous improvement led us to pilot Project Elderwand to understand whether forced breaks reduced pace of work and affected the rate of musculoskeletal disorders (MSDs). In sum, the pilot showed that the intervention was ineffective. In fact, we’ve made a number of meaningful ergonomic enhancements over the last few years, including:
- Sophisticated computer algorithms in our fulfillment centers that direct employees to pick and stow products in ergonomic power zones—the area between the shoulder and mid-thigh. As a result, there was an average 5% improvement in the cumulative forces acting on the lower back, and an average 7% on the shoulder.
- Adjustable height workstations at select fulfillment centers.
- Ergonomically redesigned packing stations at specific fulfillment centers.
- Self-adjusting carts that enable employees to handle totes (plastic bins used to move products) in an ergonomically friendly way.
Our performance expectations are reasonable and achievable. Like any business, we have performance expectations for our teams. But unlike many businesses, the expectations at Amazon are based on what the team at a site is actually accomplishing in any given week (adjusted for things like seasonality). In other words, we only ask our team to accomplish what the vast majority of them are showing us is doable. And employees in our fulfillment center network can see their own performance anytime and can talk with their manager if they’re having trouble finding the information. While of course some people might not meet those expectations, it’s the exception and not the norm.
This report also alleges that the optional games available on employees’ screens are intended to make them work faster, but that isn’t accurate. All games on employees’ workstations are voluntary things that can make work more engaging for some people – some employees like them, and others don’t. It’s our employees’ choice whether to engage with them or not.
“Total injury rate” claims are a red herring. Sen. Sanders’ interim report made false assertions about Amazon’s 2019 “total injury rate” and his reiteration of those claims in the second report are just as misinformed as the first. We report our recordable injuries to OSHA as required, which includes anything that requires more than basic first aid.
Our onsite wellness centers—formerly known as AMCARE—are for first aid, not medical treatment, and there’s no evidence of systemic issues. As he did in the interim report, Sen. Sanders again paints a false picture of what onsite wellness centers do, and do not, provide. These centers are for first aid and first aid only, and there’s no truth to the allegations that we use them to diagnose medical conditions that require more than first aid, to delay outside treatment, or to hide recordable injuries. The senator makes these sweeping claims based on anonymous anecdotes and a citation at a single warehouse, and in that citation, the allegation was actually that front line first aid responders did not follow Amazon’s procedure. Senator Sanders’ mischaracterization of this citation is both inaccurate and misleading.
OSHA requires employers to ensure ready access to basic first aid, and we do. Our wellness centers are staffed with over 1,200 onsite medical representatives who are expected to follow a detailed procedure manual – and employees can seek outside care whenever, and wherever, they wish, even if it’s for first aid. In fact, our policies state that employees are urged to seek care outside if they want or need it. While the report goes on to accuse us of systemically failing to record or “hiding” injuries, an OSHA investigation into our recordkeeping practices found the opposite – the minor issues OSHA says it found were for only small clerical errors.
We’re proud of our comprehensive programs and procedures to protect workers from heat-related incidents, and our buildings are climate controlled, contrary to what this report insinuates. We have industry-leading preventative measures in place to help reduce the risk of heat-related illnesses and injuries at our sites, both indoors and outdoors. We ensure our heat mitigation practices meet or exceed state requirements and federal guidance, and in many ways, go above and beyond to set more stringent standards for our sites. Nearly all of the employees in our North American fulfillment centers work in climate-controlled workplaces, and Amazon is one of only a few companies in the industry to install climate-control systems in our fulfillment centers. We’ve also invested in prevention procedures and training to ensure employees stay safe while working in our facilities. We train all of our employees and partners on the signs and symptoms of heat-related illness, environmental and personal risk factors, hydration guidance, and other tips from public health and medical professionals. Most importantly, everyone is encouraged to take a preventive cool-down rest break anytime they need, and employees and safety teams are empowered to speak up to address any temperature-related issues.
Unverified anecdotes are not facts. When allegations of injuries or safety violations are brought to our attention, we work to quickly and thoroughly investigate them so that we can address them. We understand Sen. Sanders and his staff spoke to 135 employees. In a workforce of more than 800,000 in the U.S. alone, that represents about 0.018% of our current frontline employee population.
Anonymous and unverified anecdotes like the ones in the report make it difficult for us to determine the details, and whether the claims are true or accurate. It’s even worse to use these unverified anecdotes as “proof” of anything, yet the senator does just that in this report. The reality is, we offer multiple mechanisms for employees to share their feedback, including all-team meetings and one-on-one discussions to engage with leaders, raise issues, and suggest improvements. For example, we have an ethics line for anonymously reporting issues that’s open 24/7. We have boards in every building where employees can write questions or feedback (anonymously or not) and get responses from their managers. As part of our commitment to safety, accountability, and transparency, these boards are located in common areas throughout each site, so all employees can easily see the questions and answers. We also have a program called Dragonfly, which allows employees to report potential hazards, near-misses, or incidents, and raise safety suggestions from an app on their phone or at terminals in sites where employees work. In 2023, we actioned on more than 200,000 pieces of feedback collected via Dragonfly. And those are just a few examples. We know that—with a workforce the size of ours – not everyone will have the same experiences. But we work hard to support our team, listen to their feedback, and keep getting better every day. If anyone wants to learn more, we encourage them to visit one of our sites, meet some members of our team, and decide for themselves.
We’re disappointed that Sen. Sanders’s has published a pre-conceived and one-sided narrative instead of a factual report. We’d hoped this report would have taken into account the thousands of pages of information, data, and details we provided throughout this investigation, and the conclusions drawn would be based on facts. But, the false information in this report doesn’t change reality: Our safety progress is well documented, and we’re proud of it. We’re grateful to our 1.1 million frontline employees and 9,000 health and safety professionals around the world who work hard every day to ensure a safe, comfortable, and inclusive working environment. Our work together is far from done. We’ll continue to invest in safety and continuous improvement for years to come as we work toward being the very best in the industries in which we operate.
1. All numbers and comparisons in this letter are based on the rates we have reported to applicable regulators or are otherwise derived from the same tracking systems used for that reporting.
