It’s a great window into the extent to which things have changed. That’s maybe an obvious point, but it’s actually underestimated among union organizers, because we have a tendency to talk about the 1930s as the great labor breakthrough. People study it a lot and ask, how can we copy the methods? People are always talking about that as a reference point.

But it’s a real question about whether we can copy those methods, because the structure of life has changed so much. Working-class people are much more dispersed. Essentially, what that means for organizers is you need to find new mechanisms through which you would bring together workers who aren’t organically linked via the neighborhood and via their workplace in these very dense, associationally rich environments.

So the short version is it makes organizing harder. And it makes it so that even when you get one organizing win, that win is less likely to set off an organic chain reaction. The metaphor I use is: if you set up dominoes too far from each other, knocking one over doesn’t set off a chain reaction. When workers are more spread out, it takes more work to organize large numbers of them.

The extent to which and the way in which this sprawl has disaggregated working people varies by industry and region a lot. There’s a general trend since the 1930s, but there are still some workplaces that are more concentrated; there are still certain neighborhoods that are more adjacent.

You have to take it on a case-by-case basis. For instance, often big universities almost function like big factories. You have a lot of different types of workers and a concentrated space. Similarly with hospitals — the workplace itself becomes a big enough environment, where a lot of workers will mix.

But if you’re talking about where the spaces are outside of work, this again becomes harder with suburbanization, essentially because people are commuting much longer distances. Part of the reason then that social media becomes more important is precisely because you have far fewer of these common, shared spaces outside of work in which you could have this organic worker-to-worker connection.

The term used at the time was “concentration.” You’re going to concentrate on the biggest industries, the biggest workplaces. The understanding was, if you’re able to organize at the heart of industry in these massive workplaces, that will set off a chain reaction elsewhere.

That was basically true — it had to do with the geography of it. It had to do with the nature of mass-production industries, where if you could paralyze one or two key factories that could set off a disruption to the whole process. That is not possible to the same extent today.

That doesn’t mean you can’t target certain places; that doesn’t mean there aren’t some industries or some workplaces that are more strategic than others. We don’t want to throw out the baby with the bathwater. But because there aren’t large workplaces to the same degree, you don’t have the same degree of concentration; labor unions have to find ways to spread more widely and find ways to encourage organizing that go beyond this hyper-targeted approach.

Right. This gets overlooked. Everybody knows that industry has declined. But what’s missed in this is the spatial dynamic, because services, for obvious reasons, have to be very decentralized.

You’re not going to have all of the Walmarts in the country in one or two cities or states — it wouldn’t make sense. Or hospitals, or schools, or Starbucks. You have to have them spread out, in relative proximity to where the population lives. Because of that, the spatial dynamics of deindustrialization mean that you have more and more workers spread out ever more widely.

For instance, with Starbucks, you have fifteen thousand small workplaces spread out all over the country. If you compare that to General Motors, in which you had a dozen or so strategic massive factories, you get a sense of the scale dilemma [and why organizing] has become so much harder. To get a critical mass when you have so many small workplaces is just an order of magnitude more difficult. It helps explain why even though workers might have as much potential power as they had in the past, tapping that power is much harder than it used to be.

I was talking to one of the staff organizers for the Utility Workers of America. They were organizing wind techs in Texas, and this guy is not from Texas. He says, these are workers who, in the past, we would have had all working in one power plant. And now they’re spread all over Texas. He says, “They’re all over creation [in] Texas. Good luck finding them.”

Oftentimes, these workers don’t even know their coworkers, because they’re operating over hundreds of miles of space in which they’re really on their own. He said that the challenge for organizing large numbers of workers is something that they had never seen before. It’s not something you’ve ever experienced before when you would have had one hundred workers in one plant; that’s different than one hundred workers spread out over a massive state.

Right. And it’s not the case that the United States has ever had a majority of workers working in massive factories. Even back a hundred years ago, there were a lot of small shops. There has been a decline. But in particular, you see the decline in that the biggest companies today, unlike a century ago, tend to have smaller workplaces.

This makes a big difference. The bigger companies are the most important for unionizing because they have the biggest pockets to give concessions, and they also are most central to the economy. So if you can’t organize a Starbucks or an Amazon, you’re not going to be able to organize the working class as a whole.

What you’re talking about, then, is the heart of the economy, the most important workplaces today, are spread out into a lot of small workplaces, averaging from fifteen or so workers at a Starbucks to even something slightly bigger at a Walmart — you’ll have maybe a couple hundred. Compare that to the average size of a factory for steel or auto, in which you’re often talking about three to four thousand workers at an average plant. It makes organizing more difficult because you can’t concentrate your relatively limited forces on a relative handful of targets.

It’s a good question. My research found that turnover isn’t as different as we might expect.

One of the excuses unions sometimes give today is you can’t organize high-turnover industries because there’s just too much churn. What this confuses is that even back in the day, in factories before unions — these were also high churn industries. What made turnover lower was unionization.

The data I found suggested that on the whole, we haven’t seen as dramatic a shift in turnover rates if we look to the pre-unionization period. Of course, if you look to the 1950s after unionization and more standardized jobs, then you do see a big shift. But if we’re asking, what does it take to organize workplaces when there’s no union? Then the turnover obstacle is relatively similar to what you saw a hundred years ago.

That doesn’t mean it’s easy. It’s definitely a challenge, and it’s definitely harder to organize when workers are working for half a year or not much longer. But unions did it in the past, and that would suggest that we can do it again today.

They do. The question is, to what extent?

There is a line of argument from various people who suggest that logistics is essentially like the new manufacturing or the new auto. In the same way as organizing General Motors sort of turned everything around for labor in 1937 and the 1930s generally, the idea is if you could organize Amazon, that would provide a similar leverage point.

The difficulty is that logistics isn’t actually as central as manufacturing was. And its mechanism of functioning, its production process, doesn’t lend itself to the same level of centralized targeting and disruption.

To give a couple of examples: If you could organize one of the mother plants at General Motors, and this is what they did in the famous sit-down strike in 1937, you could paralyze almost the entirety of production of, say, Chevrolet, or different types of lines in the company. Today, at Amazon, the distributed nature of the supply chain actually means that if you were to have a strike at just one warehouse — for instance, JFK8 in Long Island, where the Amazon Labor Union famously won — the process of distribution allows Amazon to essentially reroute around that one warehouse.

It’s not just strikes or unionization: there are weather issues, there are all sorts of issues that can happen in the supply chain, and companies aren’t stupid. They have consciously built in rerouting mechanisms to mitigate against a complete breakdown of the supply chain.

What’s missed in some of this talk about choke points is that, for instance, during the pandemic, a lot of the reasons you had a breakdown in the supply chain were on the production side of things. One of the factories, I believe in China, that made some of the parts for cars stopped producing for COVID-related reasons. That set off a chain reaction, but it wasn’t so much that the logistics-distribution side of things blocked everything.

There are certain places where there are major choke points, ports in particular. A lot of things pass through ports. Those tend to already be unionized, and the state can intervene to prevent workers from striking in some of these places.

So it’s not to say that logistics isn’t important. It is — it seems to me clearly strategically central for labor, but it just doesn’t have the same level of centralization and disruptability as factory work did in the 1930s. This raises the need to essentially have a more economically diverse and geographically spread out approach than was possible a hundred years ago.

The two main things I would stress are, one, you need to move toward a more worker-led, worker-to-worker organizing model, and rely less on staff because there are so many more workplaces. You need to be able to set off the type of explosions we’ve seen at Starbucks, where you have unionization go viral. Workers are training other workers over Zoom on how to unionize, rather than using the traditional model of labor, in which you have a staffer choose to target a workplace and then meticulously help [workplaces organize one by one]. It costs too much and takes too long to be able to organize a large enough number of workers.

Two, concurrently with that, you need to be able to rely on digital tools so that workers at, say, a Buffalo Starbucks can train workers in Arizona in organizing methods. You don’t have to rely to the same extent on staff as you used to, because you have these cheap digital tools that allow workers to talk to workers across these spatial divides. One of the reasons to be optimistic is there’s more of an ability than there was ten years ago or even five years ago for workers to be able to help train workers and connect with workers across these divides, in part because of the rise of digital tools.