Just How Quickly Are Warehouses Being Automated?

An Interview with Beth Gutelius

While conducting research for the brief primer series on this site, I’ve watched countless Youtube videos of different automation technologies at work. They can be quite hypnotizing, to the point where it’s quite easy to lose track of the relative importance of the technologies displayed.

As an example, I took quite a bit of time reviewing the different forms of automation that UPS has said it’s investing in: truck unloader arms, automated forklifts, driverless trucks. From the videos one might guess that humans are pretty ancillary to the functioning of this parcel giant. But then you take a peek inside one of their new “automated” hubs and precisely none of the fancy new technologies are actually being used.

Then again, there’s Amazon, which is investing a wild amount of money in its robotics deployments. When they say that they can imagine a Fulfillment Center running with a skeleton crew of 100 RME workers in ten years time, I’m not sure why I shouldn’t believe them.

To help sort some of these things out, I interviewed Beth Gutelius, Research Director at the Center for Urban Economic Development at the University of Illinois Chicago, and also co-author along with Nik Theodore of “The Future of Warehouse Work: Technological Change in the U.S. Logistics Industry” (2019) and a 2023 update on this report called “Pandemic-Related Trends in Warehouse Technology Adoption.”

Benjamin Y. Fong: Your 2019 report co-authored with Nik Theodore opens in very dramatic terms, describing dark warehouses humming along without humans. So is that what's just around the corner, or is it a myth?

Beth Gutelius: One of the main motivations behind that whole project was to counteract some of the hyperbole that was going on around mass displacement in this industry and a handful of other industries. We wanted to see if, through the course of a set of interviews, that seemed like it was on the horizon. The long and the short of it is no, we didn't think that in the short and medium term, dark warehouses were around the corner.

What we realized is that it's much less a question of job quantity and more of job quality. So you have all these technologies that are set to come into warehouses, which could under some circumstances improve the quality of jobs, but the choices that firms are making around how to deploy those technologies seem to really be pushing in the direction of a net decline in job quality. For the most part, these warehouse operators are looking at labor augmentation and ways of making their existing workforce more productive, more efficient, and not necessarily getting rid of their workforce. I think Amazon is a little bit of a departure, but we can talk about that.

BF: What are the conditions under which warehouse operators invest in automation, and what prevents them from doing so?

BG: In that report, we identified a set of trends that were pushing the industry towards automation and a set of trends that were pushing back against them. Pushing forward, you have first the state of labor markets. In every interview I've done over the last fifteen years with warehouse operators, they say there's a labor shortage—though we can question if it's a labor shortage or a wage shortage. You also have increasing real estate costs and increasing speed requirements, an outgrowth of e-commerce.

The things that are pushing back are the vast amount of variability and volatility in this industry, as well as the outsourcing dynamics and a couple of other things like inertia. Outsourcing is a huge issue. Almost half of warehousing is outsourced. Firms are outsourcing a core function to warehouses, and the margins are thin, there's not a ton of leeway there. And so there's also the issue of short contracts. The contracts for third-party logistics companies (3PLs) are really short, and there's a lot of churn in the industry.

If you're a 3PL, you get a contract from a lead firm, and you don't know if you're going to retain it when the contract is up in 2 to 3 years, you're not going to go out and spend a bunch of money on new automation that is specific to that firm, their needs, their products, etc. So there's a set of disincentives for the 3PL segment, which is a large segment of this industry, against automating.

That said, there are some technologies that especially the big 3PLs are investing in. It's not that they won't invest or can't, they are. It’s just that compared to the vertically integrated part of the industry, it's going to be slow and sort of spotty.

BF: On the issue of labor shortage v. wage shortage, the demand for warehouse workers has not led to a rise in wages for warehouse workers, as one might expect with a tight labor market. Why precisely is that, and has that changed any recently?

BG: No, in the last analysis I did sometime late last year, a similar trend was still in place. Real wages continue to stagnate. There's a range of reasons for that. A large part of it probably has to do with the decline in union density in the industry, which is now about 6%. That’s about the same as the national average, but it used to be three times higher than that.

I also suspect that there's a bit of monopsony happening, where you have warehouses that are concentrated in these industrial parks where there are few other employment opportunities. So warehouse operators dominate the local labor market and take the opportunity to set wages just above service sector jobs, but lower than they might otherwise be, given the demand for the workers.

BF: When we talk about warehouse automation, it’s natural to think about Amazon. But is Amazon really driving trends in warehouse automation, or is the company simply more representative of certain trends?

BG: I think it's a little of both. E-commerce has been driving the desire to automate, and that's in part because of Amazon. Amazon has definitely changed the landscape of competition, conditioning consumers to have really high expectations around delivery times. In that way, it's definitely impacting the industry. But e-commerce as a phenomenon is just very different from the typical pallet in, pallet out, or even case picking, warehouse operation. It looks a lot different, and it takes a lot more labor, like three times the amount of labor.

Couple that with the need for increased speed, and I think you find people scrambling to find workers, to pay for workers, to get orders out the door, at whatever their promised service level is. So e-commerce presents a distinct set of challenges in itself. And then you add Amazon to the mix, where all these people are competing in the same labor pool and Amazon can snatch a thousand workers, out of that labor pool when they open. So there's some labor market competition happening.

But Amazon as a company is also unique, in that it's subsidized by its other arms and can invest a lot more in R&D, in robotics. Most other warehouse companies just don't operate like that, and they don't have an AWS that's able to subsidize other operations.

BF: In your latest report from 2023, updating the work of the 2019 report, you focus on 3PLs in particular. Just how much are 3PLs shaping the market? Might they be somewhat like Walmart suppliers, subservient to parent company interests and squeezed in various ways?

BG: Yes, they are squeezed, and yes, they shape labor because half the industry is outsourced. They're obviously playing a big role in providing services and providing labor to lead firms. Lead firms are always the locus of control and power, but differently positioned 3PLs can negotiate different contracts and very different material conditions in the contract. Of course, that doesn't necessarily mean the working conditions are any better just because the company can negotiate for say, better margins or whatever. But I do think there are noteworthy power differentials across 3PLs. Employers in general, whether it's a temp agency or a 3PL or a lead firm, are not passive beneficiaries of existing labor market dynamics. They are active participants in shaping that labor market from which they're recruiting their workers.

BF: You note that 2021 was a banner year for automation, with 28% growth in warehouse automation. And I was curious as to what that means exactly. Are those technologies implemented for the long term, or really just companies experimenting, or maybe even putting on something of a show for investors?

BG: It's so hard to get real numbers on investments, so that statistic should probably be taken with a grain of salt. Sometimes a company will invest in some really localized technology for a very specific issue or problem that they're having, but that doesn't mean they're not still using a whiteboard to schedule workers. That is a literal example from a warehouse that I toured: really high-tech, scanning conveyor technology, and they were also moving around this whiteboard.

In 2019, a third of warehouses didn't even have a warehouse management system. You can't have another technology built on top of anything until you have that. If you're doing inventory management on paper, it would be really hard to integrate, say, an AMR into your system. It's impossible. So in order to connect your inventory and where it's sitting in the warehouse to a robotics system, that data has to exist somewhere in a virtual form. So it is the first building block for other technology.

BF: The Amazon Kiva is probably the most ready referent for warehousing technology, and there’s been a lot of investment in AMRs (Autonomous Mobile Robots). What makes that technology so appealing? And what are the conditions under which that investment is made?

BG: The AMRs are really modular sorts of technologies. You can buy one or two, or even lease them. You can start small and see how it works. I think especially for firms that are a little cautious about investing in technology, it's a way to wade in slowly. They're scalable, so you can scale up and down depending on your volumes. They're physically movable. So let's say you have three different operations, and you put them in one of your facilities. If it's not really working out, you can try them in another one.

But in order for them to really reach the high ROI that's often promised, your setup has to be such that you have a low pick density, so that they're not all jammed in the same aisle at the same place. And you have to have enough pick volume in order to really realize their productivity enhancements.

BF: Given your knowledge of the industry, what is the best point of entry for labor organizing?

BG: First and foremost, it's wherever workers are organizing, where are workers in motion. That said, I think we really missed the boat on Amazon. Unions should have been in motion much earlier to try to organize Amazon. 30% of workers in the industry work for Amazon, so we have to try.

There's been a lot of debate over where the chokepoints are. Maybe the air hubs, as there's just less redundancy in their air network than there is in their land network. But more broadly, going after lead firms is the way to go. That's who’s holding the purse strings. But if the workers at XPO or whatever want to organize, then that's where the organizing is going to happen, with the understanding that the goal is always to get concessions from the lead firm.