Glimpsing the Future of Work in Warehouses

MIT Sloan doctoral candidate Alex Kowalski is researching practical ways to improve jobs in an increasingly important industry.

Alex Kowalski, a doctoral candidate at MIT Sloan’s Institute for Work and Employment Research (IWER), is working on a research project with two members of the MIT Sloan faculty, Erin L. Kelly and Hazhir Rahmandad, to see if jobs in warehouses can be improved in ways that benefit both companies and employees.

As people in a number of countries hunker down at home to slow the spread of the coronavirus, many segments of the economy are suffering deeply. But that’s not necessarily the case for e-commerce, as consumers purchase goods via the Internet rather than venturing out to visit stores. Amazon Inc., for example, in March announced plans to hire 100,000 new warehouse and delivery workers in the U.S. in response to increased demand.

That may be good news for jobseekers, but warehouse jobs can be challenging even in the best of times, when employees aren’t worrying about contracting COVID-19 on the job.

Alex Kowalski, a doctoral candidate at MIT Sloan’s Institute for Work and Employment Research (IWER), has been studying jobs in warehousing for several years. He is now working on a research project with two members of the MIT Sloan faculty, Erin L. Kelly and Hazhir Rahmandad, to see if these jobs can be improved in ways that benefit both companies and employees.  Before the coronavirus outbreak occurred, Kowalski sat down with Martha Mangelsdorf, Director of Strategic Communications for the Good Companies, Good Jobs Initiative at MIT Sloan, to talk about work in today’s warehouses – and what it tells us about how work is evolving in the 21st century. What follows is an edited and condensed version of their conversation.

Good Companies, Good Jobs Initiative: You're studying work in the warehouse industry, where there’s a lot of change going on. Tell us a little bit about that.

Kowalski: The warehouse industry is home to a number of technological, logistical, and economic developments. On the logistics side, consumers are able to do more of their buying at the click of a button, and they’re expecting to get their purchases very quickly. This is reshaping how businesses that traditionally used brick-and-mortar retail stores operate. The warehouses where their goods are stored, which used to be more behind-the-scenes parts of the business, are rising in importance.

The technology used in warehouses is changing very quickly as well. There are robots that are doing impressive functions that weren't available five or 10 years ago.

Finally, warehouses are a growing source of employment. While warehouse work is physically demanding hard work, it pays better than many of the jobs available to those without a lot of educational credentials.

So the warehouse industry is at the nexus of all these trends in innovation and change that are shaping work in the 21st century. This also means we might have the opportunity to nudge things in a direction that improves this work and helps people.

With all these emerging trends, the industry sounds fascinating to research. Tell us a little bit about your project.

Kowalski:  I’m focused on two aspects of warehousing: scheduling and technology. In an earlier project, I had the chance to talk to warehouse workers about their jobs. I learned that what bothered them most was their schedules – how volatile their hours were and how hard that made their life outside of work. MIT Sloan Professors Erin Kelly and Hazhir Rahmandad and I presented this preliminary observation to the company where I had conducted the interviews, and company executives were interested. They said, “We've got to figure out ways to improve retention and employee satisfaction. It sounds like scheduling is a place to do that.” So we spent the spring of 2019 traveling to a subset of this company’s warehouses all across the U.S., interviewing both workers and managers.

We have what we call a dual-agenda perspective, which means a focus on redesigning work in ways that benefit both employees and the company. Part of the dual-agenda approach is that we get to hear both what workers find challenging but also why managers do things the way they do. So we talked to over a hundred people. That part of our research was supported by the Good Companies, Good Jobs Initiative.

What did you learn from the interviews about the scheduling issues?

Kowalski: Most workers in warehousing have what at first blush might look like a stable schedule: they work 40 hours a week, or at least they’re supposed to. So they don't face the problem that many retail or fast-food workers face of not getting enough work to pay the bills.

Instead, warehouse workers’ scheduling issues arise because they don't know when they're going to go home. The reason they don't know is because everybody is on their computers ordering a new pair of shoes or a refrigerator whenever they feel like it, which makes it hard to predict customer demand with really good accuracy. Workers have to respond to customer whims, staying as late as is necessary to fill the orders in a timely manner. As a result, they experience unpredictability and a lack of control over their schedules, which comes in the form of long, late hours caused by mandatory overtime.

There are two big problems with this situation. For workers, the problem is that it's very hard to plan life when you don’t know when you’re going home. Many people we talked to are parents, and some are single parents. Figuring out how to get your kid from a sports practice or get them to school becomes really difficult. Most parents who are able to make a warehouse job work have to outsource these kinds of family responsibilities to someone else. Add to this that because they work long, late hours in a physically demanding environment, warehouse workers are exhausted. They're fatigued, and they're often not doing their best on the job because they're so drained.

This feeds into the second problem, which is for the employer. Managers experience a negative cycle because their workers are so tired they start calling out sick, which means the remaining workers have to stay even longer. This leads to a detached, disengaged workforce and high levels of turnover because the schedules are so hard.

Why doesn't the next shift just pick up where one shift leaves off? Are these warehouses mostly just operating one shift?

Kowalski: No, they are operating 24 hours a day. But there's another layer of difficulty here, which is that a lot of these warehouses don't carry out the actual shipping of the orders to customers. They often outsource this task to a large transportation company, whether it's FedEx, UPS or some other third-party provider.

So the warehouses have to meet the expectations of the people who do the shipping for them. There is a set time of day where they must fill every order if it is to be sent and delivered by the next day, for example, or the next two days or the next three days. They can't just do all the work later in the day or they will miss these deadlines.

Got it. Does this mean that when, as a consumer, you're clicking that button on a website that says, “I want delivery in two days,” you're creating misery in a warehouse somewhere?

Kowalski: Very potentially. There's a good chance that’s the case, especially depending on what you're ordering and when you're ordering it. If it's the weekend before the 4th of July and you're buying beach chairs or towels, then you're probably going to be making people work late right on the eve of a holiday when they would like the time off themselves.

One thing that was clear in my interviews was that unstable, unpredictable schedules have real costs to people's personal lives. They weigh on people's relationships and make it hard to form new ones. And they also lead to physical exhaustion that creates secondary problems. I've heard about car crashes. I've heard about leaving children at daycare past their pick-up time. I've heard about missing important life events. Often, people are trying to find a different employment situation largely because their schedules are so difficult.

Wouldn't it be possible to have a pool of people who are interested in taking on additional work? It seems like a company should be able to manage the amount of mandatory overtime.

Kowalski: Our hunch is that there are a couple ways we might devise a solution. One is what you're talking about, which is the use of new staffing strategies -- hiring people who either express an interest in a certain schedule or creating different employment categories with different schedules. That's one path forward.

Another path forward is job rotation, giving people the skills to perform more than one function. Within warehouses, there are multiple kinds of jobs. And often, depending on the size of the warehouse, workers do just one. On a high-volume day, one way to prevent people from staying overtime too long is to ask those in another part of the operation where they don't have as much work that day to come help the people who do. This requires an emphasis on training because workers need the skills to be versatile. It also relies on really engaged managers who work as a team.

The third path we’re considering is giving workers some control or input over when they work. This could be accomplished by some kind of tool -- an app, for example -- where workers are able to either bid on the overtime available or to preplan days when they won’t work overtime. This could extend all the way to helping set schedules in advance or to shift swapping. We think there are variations on these kind of flexible scheduling applications that might make a lot of sense.

You’ve talked about the impact of scheduling practices on workers’ lives. What else have you learned in your interviews that surprised you?

Kowalski: One thing that was surprising to me was hearing from managers about the variety of different ways that they manage their specific operations, all geared to serve the customer as quickly as possible. While there is some similarity across all the warehouses we studied, managers have unique styles and ways of addressing challenges. You might expect that if you have one company, with a clear mission, a shared culture, and a set of business imperatives that are felt across its multiple operations -- get things out the door as quickly as possible, satisfy the customer, and don't make errors – that managers might settle on one optimal solution.

That's not what we see. Managers handle scheduling in very different ways, whether it’s because they once worked on the warehouse floor themselves, want to manage in accordance with their personal values, or care only about the bottom line.  And managers’ different approaches to scheduling have very different consequences for workers.

I see this as a really positive finding because it indicates that while you might be in the business of warehousing, you're not beholden to a just-in-time, precarious, unpredictable way of scheduling your employees. You have some leeway. One of the things I hope our research project can do is identify those approaches that work well for both managers and workers and share them widely.

That makes sense. In a high-pressure environment like these warehouses, managers probably don't have the time to go out and find innovative solutions being used elsewhere.

Kowalski: Exactly. One might expect that if somebody had arrived at an optimal solution, it would quickly travel within the company, but that’s not the case. Even those who supervise multiple warehouses are not always aware of what's going on in one versus the other.

What’s the next phase of your research?

Kowalski: The interviews were the initial step in a larger project. We conducted them to get a clear diagnosis of the problem. The next step is to quantify the costs of the problem, looking at warehouse-wide data. And then we’re going to develop an intervention, a solution where we test out experimentally some of the ideas in a subset of warehouses and see: Can we improve the situation? Can we improve employee well-being and satisfaction, reduce turnover and absenteeism, and then tie that to productivity and potentially profit?

If workers have more control in their workplace, that's likely to make them more committed and engaged. It's also likely to make them have less work-life conflict, which will make them enjoy coming to work more. And it's also likely to reduce fatigue and stress in ways that make employees able to do a better job at work, which we think can then lead to better business outcomes. In 2019, we received grant funding from the Washington Center for Equitable Growth to help us develop this intervention and bring on some great new team members, including Kirsten Siebach, a pre-doctoral researcher. We hope to pilot a potential solution sometime this year.

Sounds exciting. You spoke earlier of the technological changes going on in warehouses. How does that weave into your research?

Kowalski: There’s a lot of momentum in warehouse robotics right now. Lots of startups have emerged, many of them in the Boston area, and they’re developing some cutting-edge technologies. There's so much interest and enthusiasm for what they’re developing that when you talk to CTO types in warehousing, they'll say, “If our company is going to survive and be part of the next generation of warehousing, I have to consider this technology.”

But things are still nascent. There's a great deal of experimentation going on. There hasn't yet been a market winner who has determined the way that all warehouse robots are going to look. Instead, there are many different approaches and visions out there. And so it's an intriguing time to compare these different visions and ask: What are their implications for the workers who are going to work with the various technologies?

What I'd like to do is to ask, “Here's robot X. Here's robot Y. If a company invests in them, how is robot X going to affect workers differently than robot Y will?” My hypothesis is that the robot that is better for workers will also be better for business. Given my exposure to some robots in use on the ground in visits I've made to warehouses, I've seen that the robots that are designed to be used in collaboration and interaction with workers give warehouses a lot more flexibility to deal with contingencies and unexpected events.

Think of the conveyor belt, which is in a way a robot that doesn't on the surface need humans at all. When that thing jams, it shuts down the entire warehouse. It‘s pure paralysis. I think robots that are designed with the intent of totally removing humans the way that a conveyor belt does are actually going to create a lot of problems -- at least initially, as organizations try to adjust to them. Companies should be asking themselves this same question, and they should be taking input from their workforce to arrive at an answer.

What you're describing sounds like an industry in tremendous flux. People are experimenting with lots of solutions, both in management and with technology -- in the sense that, even within this one company, managers at different warehouses are implementing different solutions to the scheduling problems, at the same time that all these startups are trying different approaches to the technology.

Kowalski: Exactly. And companies are not sure which technology is going to be the best for them, or which robotics startup is going to still be in existence in a couple of years. So many larger corporations are testing or even investing in more than one startup until a winner is decided.

I recently read “On the Clock,” a book about low-wage work by journalist Emily Guendelsberger. As part of the research for the book, Guendelsberger worked in an Amazon fulfillment center before Christmas, picking orders. At the warehouse where she was employed, workers would get a handheld scanner that told them where to go next in the warehouse and that gave them a certain amount of time to do each task. And so part of the experience she wrote about was about being micromanaged by the algorithm. Did you see that kind of set-up where, literally, the software is telling workers, “Go here”?

Kowalski: In today’s warehouses, there's all kind of variations in how work is performed. At one extreme, a worker is basically following instructions on a piece of paper. At the other, there are devices you wear that tell you where to go and track your moves at all times. Different methods are more or less controlling. My understanding is that Amazon is particularly controlling in that sense.

There is so much unfolding in warehouses right now. You have an algorithm telling the person where to move. You have robotic technology that might be replacing people or might be helping them. You have businesses that are trying be ever more closely connected to the market, adjusting their operations in real-time, and passing some of this onto their workers. All of this is creating a new kind of workplace.

Is this the kind of workplace in which more people are going to find themselves? What do these developments mean for how we experience work, in warehouses and outside of them? Can we come up with ways to make warehouse work better, and can we come up with the kinds of policies and institutions that move things in a positive direction? That last question is the one I hope my work can help answer.