Amazon’s direct workforce rose by 500,000 people in 2020 — that’s half a million people — to nearly 1.3 million employees. The extraordinary hiring, supporting the rapid expansion of Amazon’s warehouse and delivery operations, raised the natural question: what would be the consequences of all that growth?
An eight-month New York Times investigation, published this week, provides much of the answer, telling the stories of warehouse workers caught up in an unforgiving, error-prone system that struggled to keep pace with Amazon’s growth, the unique challenges of the pandemic, and unprecedented customer demand.
The piece reveals the limitations of Amazon’s automated HR technology, but it also demonstrates the impact of policy decisions by Amazon executives, including founder Jeff Bezos and operations-leader-turned-consumer chief Dave Clark.
Among them, according to the story: a conscious decision to encourage turnover and limit upward mobility among hourly warehouse workers.
Karen Weise, the Seattle-based New York Times tech reporter who reported the story with her colleagues, Jodi Kantor and Grace Ashford, speaks with us about their key findings on this episode of Day 2, GeekWire’s podcast about everything Amazon.
Listen above, and keep reading for highlights, edited for brevity and clarity.
Todd Bishop: How would you describe what you found in terms of the practical realities of these systems, the impact of all this growth, and the policies that enacted these systems in the first place?
Karen Weise: We were really trying take a step back and look, overall, at Amazon’s employment model. There’s been a lot of reporting about safety issues at Amazon and things like that, and the company has begun to really address those. But we wanted to look at how is it as an employer broadly.
We found that a lot of the things that it developed when it was smaller could be more closely managed. But as you scale as fast and as much as it has, they really have been under strain. It’s moved to a very technologically driven work environment. So you have that in everything from the productivity monitoring software to the HR systems. When you’re moving as fast as they are, and growing as fast as they are, it’s really become harder to implement that with precision and care.
One example is Amazon’s productivity tracking. There’s this mythology that you can’t go to the bathroom in the warehouse, and that’s because of this idea of called time off task, which means Amazon tracks every moment you’re not actively producing, essentially, you’re not scanning a product, for example. You can go to the bathroom. But this is a mythology that is developed from a real fear, though, of people feeling constantly monitored.
Time off task was designed to identify operational impediments, to say, “OK, if this machine keeps breaking down, and creates all this time off task, let’s fix the machine or move this product here,” or whatever it may be. But the way it has become translated to employees at this point, at this scale, with all these managers implementing it, is that it’s this surveillance technique.
Very, very few people actually get fired for time off task — less than 1%. But it has this overarching effect on the stress in the work environment, because you’re now translating it through so many people.
And then you also have people who are hired now by machine, functionally. They don’t interview warehouse workers; [prospective hires] fill out a learning assessment. The kind of engagement that you might get with someone when you’re actually interviewing — you not only learn about the prospective employee, but the employee learns about the work environment — you miss that human connection.
One of the overall takeaways from the story is that it’s a numbers game at this point. You have this incredibly high churn; you have 150% turnover a year, roughly. When you’re moving through that many people, it just creates a lot of chaos and unevenness.
TB: It seemed like a lot of this was just unintended consequences. But then you went into the executive ranks. One of the most interesting things about this piece, I think, was that you talked with a couple of highly placed former human resources and technology leaders, and identified them by name, not unnamed sources. And they detailed the fact, for example, that Amazon purposely limited upward mobility among its warehouse workers, which is notable in part because of the racial makeup of the workforce, largely people of color. This, to me was one of the key takeaways.
There was also a story that came out earlier this week from Recode that simultaneously detailed additional problems with diversity, equity, and inclusion in Amazon’s workforce. How would you describe the intentionality of these systems? Did you largely see things in your reporting that were the direct and purposeful result of decisions made by management? Or were they inadvertent things that just happened randomly?
Weise: There’s both. We have an example of a woman who was fired for a single bad day because she had too much time off task. She was a top performer. But it wasn’t an erroneous firing. That was the policy at the time, that you could be fired for what happens on a single day. Amazon just announced changes to that specific policy about a week ago. We have been asking about that example and this topic for months. They claim the change has been in the works for months.
We also found that Jeff Bezos had this idea, this belief, that a disgruntled workforce was a threat to Amazon, not necessarily because of unionization. Unionization was an indicator of problems, not the problem. According to David Niekerk, the long-serving vice president who built the HR operations for the warehouses, Jeff Bezos thinks that people are essentially inherently lazy. The phrase that he would say is, essentially, people would expend the least amount of energy necessary to do what they want or need. That is a core idea, actually, of all of Amazon, when you think about it: the whole reason we click is because it’s easier than schlepping to the market or going to four stores to find the dongle for the computer.
Jeff Bezos felt that basically Amazon wanted people who would go above and beyond. This is not just in the warehouse. You see this in the whole bar-raising approach in the corporate environment. But he felt that people would become disengaged over time and wouldn’t go that extra mile, and it would become “a march to mediocrity.” That’s why, instead of giving people pay raises over longer periods of time, they stop after three years. Career Choice is a program to get people to essentially leave Amazon. There’s something called “the offer,” which is paying people thousands of dollars to leave.
It’s an interesting decision. You could say, “I’m going to invest my resources and say, how do we prevent that three-year cliff? How do we make people feel extra engaged when they’re here longer?” as opposed to saying, “OK, they’re disengaged, let’s get them out and get a new batch of people in.”
So in some sense, you have a philosophical push for it. But in another sense, we did just find straight-up errors and mistakes. We had a guy who was on an approved leave, who was fired and wrote these pleading emails into this HR void, saying, please know this, I would really like to keep my job. So that’s why we used the word inadvertent firing. It wasn’t the intent. But it was still there; it still happened.
TB: Amazon, in its response to you, on a variety of fronts, said that many of the issues were outliers, that they were the exceptions to the rule. I recognize that you’re being very factual in this story. Can you give me a sense, without giving a qualitative opinion on Amazon’s statement, for whether what you saw were outliers?
Weise: There are some sentiments that were very common. For example, I’m in these employee Facebook groups, these associate Facebook groups, and I don’t use them to look for individual examples. I look to see, what do people talk about a lot? What comes up regularly? We had issues that we even brought to fact-checking with Amazon as we went through this process, and then we decided, you know what, that is too one-off; we don’t hear that complaint that frequently. And so we took them out of the story. … Yes, the specific examples are singular. But we tried to pick themes that we heard consistent problems with. … Thematically, the issues were common.
TB: We have seen, in recent weeks and in recent months, steps by Amazon to address these things. You mentioned the change in the time-off-task protocol, which is going to be looking at a longer timeframe. Somebody who has a bad day is not going to be punished with termination because of their time-off-task metric. You also have Jeff Bezos saying in his final shareholder letter as CEO that he wants Amazon to become Earth’s best employer. What’s your sense of the ability or the willingness inside Amazon at this point to make real change on these issues?
Weise: People externally don’t understand why Jeff saying he wants to make Amazon Earth’s best employer is a big deal, because it sounds like a PR statement. But you look at these shareholder letters, and they’re prescient about what the company cares about. Since then, we’ve learned more about Amazon’s safety response in the warehouses.
But it’s not clear what he means by Earth’s best employer. They have announced some pay raises, but that’s the tool that they’ve used in the past to improve employment, or the most prominent or public tool they’ve used. And it’s also a labor-market response.
At the same time, you have Andy Jassy stepping up as CEO, who while part of the senior leadership team was in a totally different part of Amazon’s business. So how he approaches this and thinks about it, I have no idea, frankly. He’s been running Amazon Web Services, which as you know, is Amazonian, but it’s really its own operation, in many ways. And Dave Clark, who has been essentially the architect of Amazon’s operations, is now the CEO of the consumer business. So does he have a reckoning in some form? I don’t know.
I think one of the major factors that could create a change is the labor market and Amazon’s growth needs. One of the fascinating things I found in reporting this is that people kept describing to me, essentially, a palpable fear of running out of workers, that this is an existential problem. That not only is the growth demand so big, but the turnover means you need this endless stream. I definitely hear from workers who leave and come back and leave and come back, there’s no doubt about that. But you need so many fresh bodies still to feed this turnover machine.
We see it now: hiring bonuses, pay increase, not screening for marijuana. They’re making changes to their policies to bring more people in, and you can only tweak on the edges so much without addressing the core of the job and getting people, frankly, to stay longer.
There’s actually an interesting debate right now. Since we’ve reported this, there have been these various takes on the piece. Clearly this model has worked for Amazon so far as a company. The question is, will it work going forward? And some people say yes, and some people are saying no. The fact that I was hearing so much from people internally expressing concern about it tells me that there is pressure to change, but what form that takes, I don’t know, because it is such a metrics-heavy company.
The productivity approach — which is, again, not the singular reason people leave, but is a dominant part of the work experience — that has incredible allure at Amazon. And the demand for that is even greater the faster you promise shipping, the more precise you get for delivery estimates. … They need to know exactly how a building is producing at every moment. It needs to be consistent, and predictable. And that creates even more productivity pressure. So you have all these countervailing forces of the labor market pushing one way, and the business model pushing another way. And the idea of being Earth’s best employer being co-equal to Earth’s most customer-centric company, that’s going to be tested.
TB: There was the landmark piece about Amazon’s bruising workplace several years ago by the New York Times that drove the conversation about Amazon for weeks, and one of your colleagues on this piece, Jodi Kantor, wrote that past piece with David Streitfeld, another New York Times colleague of yours. I was struck, reading this new piece, having paid a lot of attention to that prior piece: it felt like the three of you on this piece went out of your way, took great pains, to present all sides. There are sections of the piece where it’s clear: this is a great place for some of these folks to work. Did you approach this piece and the reporting of it with lessons from that past piece in mind, as an institution?
Weise: I didn’t work on that one. I can say, I have been talking with Amazon about this story since, I think, early fall, somewhere in there. There was a multi-month effort to try to get more executives on the record, more people on the record, and to push for interviews. In the end, we landed somewhere in the middle. We got this tour of the Staten Island warehouse that we focused on from the general manager, which was very helpful, to get the tour from her and to see the warehouse through her eyes, as well as to have an interview with her.
Then we spoke with the pretty new vice president of HR for the warehouses on the record, as well. I was so grateful for those. I wish we had more, frankly. An hour with someone, I’ll take it, I will always take it. But it’s not the same as having an open-ended, ongoing conversation with someone and with the most senior leadership of the company. We put in a request for everyone and anyone. I don’t mean that loosely, but we put in specific requests for Jeff Bezos, Dave Clark, [Amazon HR chief] Beth Galetti, all these people.
I was asking people who would be good, who else should I talk to? And there were some really interesting suggestions of people who had risen through the warehouses into leadership roles, not the tippy top, but a pretty unique perspective. Same thing on the HR side.
We had then an extensive, I would say almost borderline epic, fact-checking process with Amazon, over many, many weeks, to try to make sure we were accurate and precise, and if there was any context from them that we should include.
So we worked hard to do that, because I think most things in life are not cut-and-dried. And we knew that, and so we wanted to include as much as we could get. I am grateful for that fact-checking process because it definitely elicited more information, more context for the story that I think really serves everyone and most importantly our readers, to really understand, and they can make their own decisions about what they think about things. People are human, they live in the world, and they can they have their own judgment.
Subscribe to GeekWire’s Day 2 podcast in Apple Podcasts, Google Podcasts, Spotify or wherever you listen.
