In the popular imaginary, Ford River Rouge complex has come to represent an industrial age long gone. Completed in 1928, the Rouge complex comprised ninety-three separate buildings in Dearborn, Michigan, occupying about 1.5 square miles in total. About 80,000 workers were employed there at its peak.

In addition to its vast assembly plants, the Rouge complex had its own docks, railroad tracks, electricity plant, and steel mill. All of this was surveilled and guarded by some 8,000 “service men,” the private army of thugs hired by Ford and run by the infamous Harry Bennett. A sprawling manufacturing facility with a mammoth workforce, the Rouge complex embodied everything we supposedly lost with deindustrialization: centralization, vertical integration, manufacturing, workforce, and working-class community concentration.

The Rouge was a particularly gargantuan example of mid-century industrial geography, but similar organizing targets existed around the Northeast and Midwest: Goodyear in Akron, General Motors (GM) in Flint, stockyards in Chicago, steel mills in Pittsburgh, General Electric plants in Schenectady, New York, and Lynn, Massachusetts. These were the key targets in the 1930s, the strategic chokepoints of the industrial society of the period, and in a remarkably short amount of time, between 1937 and 1941, the Congress of Industrial Organizations (CIO) conquered them all. Where that particular spatial dynamic was absent, the CIO either struggled (as in the West) or largely failed (as in the South).

The CIO’s success derived from two key sources: First, a relatively friendly political environment in which Franklin D. Roosevelt and various New Deal politicians were not willing to send in the guns, or at least were willing to be somewhat neutral in conflicts between capital and labor. And second, the ability of unions and workers to overcome previous divides and unite in effective strategic disruption. This did not just mean coming together to strike. It meant bringing the gears of production to a grinding halt. The Flint “sit-down” strike was won at the moment when workers seized Chevrolet Plant No. 4, the only motor plant in the complex. If they were not able to do so, it is not clear that GM would have come to the table, at least as swiftly as they did.

The organization of the Rouge is typically discussed as something of a footnote to the more heroic year of automobile organizing in 1937, but here, too, workers had to demonstrate their disruptive power. In the months prior to the decisive strike in April 1941, Ford employees lost their fears of the service men and began meeting company violence with their own. Then when a facility-wide strike started at the beginning of April, workers not only quickly formed pickets around the twenty-seven gates to the Rouge complex but also established automobile barricades at those sites and at key access roads for about 10 square miles around the factory.

Nothing moved in that area without workers letting it move. It was a dialectical irony probably not lost on the Marxists at the time: automobiles playing a crucial role in stopping the production of automobiles. Of course, it helped that Ford had been recently awarded a $123 million government contract for the production of aircraft engines, and the last thing they wanted were production holdups.

But to get their union, the Ford workers still had to demonstrate the ability to shut things down, completely and until they agreed to let production resume.

The CIO moment is the one reliable referent in American history for a true labor upsurge. In the span of a few years, the country’s core industries went from unorganized to fully organized. But given the relative dearth of manufacturing complexes like River Rouge and concentrated working-class communities after deindustrialization, what lessons can we really learn from that moment?

Manufacturing is certainly not dead in the United States, but it is in decline and certainly much less central than it used to be. Working-class associational ties and values are radically different than they were in the 1930s. In awe as we might be of the organizational power the CIO was able to build, are not our fights so different from theirs that it is just wistful nostalgia to look back to that era for lessons?

There are innumerable differences between then and now, but there is one key lesson from that period that should never be minimized by any amount of historicizing of the particularity of our situation: go after the big targets and disrupt their operations until they recognize the union.

This is an obvious claim to make in one sense, but what does it concretely mean today? There is nothing quite like a Chevrolet Plant No. 4 with tens of thousands of employees packed around it today. So how do we go after the big targets and disrupt their operations when the targets are more nimble, workplaces are fissured, and working-class communities are fractured and diffuse?

This article is an attempt to apply that one key lesson from the CIO period to the present. The big targets today are no longer manufacturing companies as they were at mid-century, but rather a mix of retail and parcel companies (see chart below), all of whose strength lies in the sophistication of their logistical operations — as Peter Olney has written, the other “historic basis of labor’s power.”

Just as it was apparent to CIO leaders that the labor movement was not playing to win if it did not take on GM and US Steel, so too is it increasingly clear that labor must go all in on organizing Amazon and Walmart today.

Amazon and Walmart are the biggest fish to fry, but most of the other large employers in the country, including FedEx, Home Depot, Target, Kroger, and Lowe’s, have built up their own highly efficient logistical capacities. If the biggest targets in the 1930s were all manufacturers, all of the key targets today are in logistics. But what kind of opportunity do contemporary logistics operations really pose for the labor movement?

The New Industrial Geography and the Problem of Organizational Purview

Recently, a debate has emerged between the left-labor scholars Kim Moody and Eric Blanc, (see “How to Unionize One Million Workers Every Year,” by Blanc) on this precise question. In his book, On New Terrain: How Capital Is Reshaping the Battleground of Class War, Moody argues that the logistics revolution in the United States has resulted in massive concentrations of logistics workers in certain clusters, roughly on the scale of workplace concentration in the mid-century industrial geography.

Blanc, by contrast, believes that increasing decentralization and downsizing trends in workplaces and in residential life mean that “we can’t copy 1930s union targeting strategies.” He is particularly skeptical of Moody’s proposal that logistics organizing today presents the same possibilities as manufacturing organizing in the 1930s. As he contended in an interview with Doug Henwood:

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. . .. [This is] 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.

To make this claim, Blanc engages in a somewhat surface-level analysis of the United Parcel Service’s (UPS) organizational structure that makes it appear that there is far less concentration in its system than there is. On the other hand, Moody does something similar in exaggerating the concentration in logistics clusters: at one point, he points to Memphis as an air, rail, and trucking hub employing 220,000 workers, but that number appears to be derived from the number of both full- and part-time jobs generated by the Memphis cargo airport for the entire state of Tennessee.

The truth is somewhere in the middle: there are some organizing targets out there (like the Memphis cargo airport itself, which employs about 13,000 FedEx workers) that constitute huge opportunities for the labor movement, according to the traditional perspective of industrial unionism. But in many key logistical hubs, there is far less spatial concentration than there was in the CIO period. And in any event, where there is workplace concentration, it is often spread over multiple companies, which themselves utilize a maze of subcontracting relationships.

There is a difference, however, between the raw numbers of potentially organizable workers in a particular place and the economic importance of that place (and thus the leverage that might be gained through its control). There are far fewer longshore workers at the ports than there were in the 1930s, but they are still able today to command the respect of the carrier industry because they can shut down facilities of massive economic importance. This is to say that a particular node in a supply chain might be worth targeting not because it will net a certain number of workers, but because it affords a great deal of leverage over companies that care much more about the flow of their goods than the actual places that they flow through.

This is perhaps the biggest conceptual hurdle in thinking through the organization of the large employers today. Big manufacturing companies rely on key sites of value extraction; the big logistical companies rely instead on key flows. Organizers at mid-century had their work cut out for them, but strategically, things were quite simple: go after the big factories and try to shut them down. There was some thinking through of supply chains even in that era — for instance, when communists were dynamiting segments of railroad track to prevent the movement of steel — but the organizing purview was predominantly place-based.

Organizers today often still think in these same terms, asking questions like, “What are the individual facilities where there’s interest in action?” or “How do we build support for a recognition campaign on that individual facility?” If they’re thinking outside the box, maybe they will also consider how community support can be marshaled.

All of this is fine and well in certain cases, but it operates by a logic that does not suit the task of organizing employers for whom place is much less important than it used to be. To tackle a company like Amazon, organizers will have to find the pain points in its goods circulatory system and think about themselves not primarily as organizing individual workplaces but about disrupting the companies’ operational flows to the end of organizing the workers who make them possible.

If we look at important logistical nodes from the standpoint of a traditional union, that is, how many new union members at a particular facility might be produced from a certain organizing expenditure, then some of the time such places will look like they are not worth the hassle. From this perspective, Blanc is probably right to say that logistics is not as “central” as manufacturing was in the 1930s.

If we look at such nodes in terms of what it is possible to leverage from their disruption, however, then the proposition looks altogether different. The problem then is not that opportunities to “wound capital,” in historian John Womack’s phrasing, are less crucial than they were in the 1930s, but rather that we lack the strategies and institutions necessary to meet their networked structure and scale; that the key targets today are not too small and diffuse, as Blanc would have it, but simply difficult to reckon with given existing organizational purviews.

To stick with the CIO moment, I believe the issue of organizational purview was the fundamental problem with the American Federation of Labor (AFL) in the 1930s. The AFL had many problems, but by the 1930s, its top leaders very much wanted to meet the moment in organizing the basic industries. The key thing that separated the AFL’s John Frey from the CIO’s John L. Lewis was not whether the basic industries should be organized but how they should be.

The AFL had persisted for half a century at that point on the principle that the best means of ensuring solidarity organizationally was to unite workers by craft. Thus, their strategy was to issue federal organizing charters to capture all workers in a plant on the front end who would then be hived off into different crafts on the back end. James Matles, director of organizing for the United Electrical, Radio, and Machine Workers, complained at one point that the AFL wanted to split up the workers in one General Electric plant into seventeen different unions.

In retrospect, it seems obvious that industrial, wall-to-wall unionism was the path forward and that the craft orientation of the AFL was hopelessly outmoded. But at the time, with the prior failures of industrial unionism efforts in mind, it was an unthinkable waste of resources for the AFL leaders. They were dead wrong in their assessment of the situation. And the essential problem again was one of institutional separation and limited organizational purview.

The idea that demands could be made regarding the entirety of the practices of a company or industry (rather than on behalf of different trades) by workers of disparate job classifications, races, and ethnicities just was not a viable option for the AFL.

From the standpoint of strategic chokepoint analysis, I believe that the labor movement today is somewhat like the AFL in the 1930s. Now, as then, a structural transformation in the economy has rendered the traditional modes of organizing not obsolete by any means (even craft unionism is still alive and well today), but inadequate to take advantage of the new geographies and chokepoints of production and distribution.

Disrupting Flows

What would a labor movement geared toward disrupting flows instead of organizing individual places look like? It would not be radically transformed in practice, but it would have a different organizing orientation and a willingness to experiment with new tactics. In what follows, I review four features of this orientation.

Targeting key nodes at which disruptive labor actions have maximal system-wide impacts
There were many flaws in the OUR Walmart campaign initiated in 2011, but the key one was that it went after Walmart stores instead of its distribution centers. This lesson must be internalized by any union that wants to take on the big retailers or parcel companies. Since then, there is also increasing sophistication and differentiation within distribution center networks to contend with.

To my mind, the basic criteria for choosing a good distributional node to target are twofold:
1. The node serves a critical role in the larger distributional system
2. Labor is capable of impacting it through disruptive action

The Amazon Fulfillment Centers, where orders are selected and boxed up, are a good example of number 1 but not number 2: their efficiency and size are absolutely central not just to Amazon’s distributional capacities but also to their overall business model. But these facilities are tremendously difficult to disrupt, given the redundancy Amazon has built into its system.
By contrast, the Amazon Delivery Stations, where packages are picked up to be delivered to your doorstep, are a good example of number 2 but not number 1: last-mile delivery is the most costly segment of a package’s journey, and it is very labor-intensive. Delivery stations also tend to be much smaller than other distributional nodes, making them easier organizing lifts.

But for that very reason, shutting one down would not have broader impacts on Amazon’s network. A strike at a delivery station would be visible in a particular metro region, but it would be contained, and Amazon would raise throughput at other nearby delivery stations to compensate.

As I have argued elsewhere, the sortation centers, the “middle-mile” facilities where packages are sorted by zip code, check both boxes. In the case of the sortation centers, they are rooted near metro areas to serve a set number of delivery stations, and so are place-based in a way that fulfillment centers are not. In the traditional “hub-and-spoke” logic of parcel operations, in which a number of smaller spoke facilities route packages through central hub facilities, they are more crucial as distributional “hubs” than the delivery station “spokes” to Amazon’s overall operation, and they are fewer in number than either the fulfillment centers or the delivery stations.

There are other facilities in Amazon’s system that potentially satisfy both numbers 1 and 2 (I am thinking in particular of their new network of national inbound cross docks, which serve as key storage facilities for Supply Chain by Amazon), and perhaps the sortation centers are not the targets I am making them out to be, given aspects of their operation of which I am currently unaware. The important point for my present purposes is that a labor movement interested in disrupting flows will be locked in on this question of which nodes tick both boxes.

Winning over the technicians
In all distribution operations, there is a layer of workers above the baseline pickers, stowers, packers, and package handlers but below management: the technicians. At Amazon, they are reliability maintenance engineering (RME) workers; at UPS, they’re “specialists.” The names will differ from company to company, but these are more skilled workers who fix belts and robots, control package flow, and generally keep the operations humming. They are paid better than the average warehouse worker, but not so well that they could be considered a form of “labor aristocracy.” In the last few years, warehouse automation has taken off, and the importance of these technicians to logistical efficiency is only going to grow.

As with any skilled workers, technicians can be bought off, or they can develop a craft union mentality. But they can also find common cause with all workers at the company, as the tool and die makers at GM plants once joined forces with more unskilled assembly workers. In the latter case, they have a special role to play, given their intimate knowledge of how operations can be disrupted.

Organizing like workplace fissuring is a legal fiction
At many key logistical nodes and clusters, there is aggressive subcontracting, as at Walmart’s facility in Elwood, Illinois.

This workplace “fissuring” described by David Weil, Gerald Davis, and many others has been devastating for unions, in that it makes little sense to try to organize the fly-by-night sub-subcontractors depicted below. But the technical fissuring that has stultified the labor movement obscures an ever greater functional integration: parent-company subcontractors often have access to real-time point of sale data and gear their entire operation to parent company order fulfillment.

Of the many legal fictions under which the labor movement must suffer, the idea that workers at a subcontractor that is so integrated into a parent company’s operations so as to appear in practice as one of its own departments are not functionally employees of that parent company is one of the more egregious — and the labor movement is going to make little headway into such facilities if it does not draw inspiration from recent history to work around the constraints of labor law.

The National Labor Relations Board recently ruled that Amazon is a “joint employer” of its delivery service partners’ drivers, a major win for the Teamsters Delivery Station organizing efforts. In Donald Trump’s second presidential term, the International Brotherhood of Teamsters (IBT) is going to have to keep organizing with the joint employers in their sights, though most likely without the backing of the National Labor Relations Board.

Leveraging the power of already-organized workers to spur new organizing
As organizer Carey Dall has persuasively argued, taking advantage of logistical vulnerabilities today must include internal organizing in longshore, rail, and trucking unions, all of which could exercise enormous structural power to the end of spurring new organizing. A recent example from the ports is illustrative: in 2021, mechanics at P&B Intermodal in Tacoma, Washington, went on strike for recognition, and International Longshore and Warehouse Union (ILWU) Local 23 supported the strike by refusing to deliver goods to the yard. The union was recognized six hours after it was demanded.

Other port-adjacent intermodal yard workers and on-port drayage drivers would all benefit from similar leverage. And though it would take a great organizational revolution within the longshore unions to transform them from craft unions of longshore workers into industrial unions of logistics workers, the International Longshoremen’s Association (ILA) and the ILWU, with accurate knowledge of where containers are going from the ports, could once again “march inland” into other parts of logistics networks beyond the ports, as they did in the 1930s.

Amazon Inbound Cross Docks and Walmart Import Distribution Centers are both located close to ports to deal with inbound goods, and workers at either such facility would be greatly aided by support from the longshore workers in something like a recognitional picket. The same basic point holds for railworkers at intermodal terminals. The power is there if those muscles are exercised once again.

Strategic Disruption, Today

The labor movement is not inexperienced in any of the features just mentioned, just as the labor movement of the 1930s had plenty of experience outside of the craft model. It is a question of emphasis and will, not complete overhaul, though pursuing some of the things just mentioned to their logical end might mean some true novelties in practice. As an example, if organizers really set their sights on the key distributional nodes, what about workers organizing at stores and spoke facilities? Could they be organized to travel upstream in the distributional network to picket a feeder facility rather than only take action at their immediate workplace?

As an example, one relevant to present efforts: Starbucks runs five domestic green bean storage and roasting plants and has five regional distribution centers, each of which supplies roughly three thousand stores with coffee and other supplies. Could Starbucks store workers support the organizing efforts of distribution center workers, who in turn could leverage their structural power to win gains for store workers?

Similar examples could be drawn for any of the large retailers. Home Depot has a large rapid distribution center and internal freight consolidation facility at CenterPoint Intermodal Center (CIC) in Joliet, Illinois — together these facilities take up about 2.2 million square feet, roughly the size of the twenty Home Depot stores in the Chicagoland area. If nothing were to move out of Home Depot’s CIC facilities, what would the company be willing to concede to get things moving again? Regional card-check agreements?

Industrial concentration in large factories required an industrial approach; highly integrated networks of facilities require in turn a more “networked” approach.

This is obviously blue sky thinking, but it points in the direction of the kinds of actions that a labor movement interested in flow disruption would pursue. If we are going to go after the big targets, as the labor movement did in the 1930s, we must also figure out ways to disrupt their operations as a means to leverage union recognition.

Today that task involves different spatial dynamics than it did for the CIO. Sometimes it means tackling massive single-employer targets, as in the case of the promising campaign at Amazon’s KCVG air hub. But sometimes it involves fragmented and precarious workforces across multiple companies, where the attraction of the target is partially an organized workforce in one particular location but more importantly the leverage over a network of facilities that the organization of that location brings. Industrial concentration in large factories required an industrial approach; highly integrated networks of facilities require in turn a more “networked” approach.

What this entails, including how precisely such leverage is gained and what kinds of demands can be productively fulfilled by certain entities within those networks, is a tremendously complicated matter. But organized labor is going to have a difficult time making much headway in tackling the corporate behemoths of the day without thinking along these lines.