Earlier today the National Labor Relations Board announced the results of the vote on whether workers at the Amazon warehouse in Bessemer, Ala., would join a union. The vote was 738 in favor to 1,798 against. It’s bad news, but it doesn’t mean workers in future Amazon campaigns won’t or can’t win. They can. The results were not surprising, however, for reasons that have more to do with the approach used in the campaign itself than any other factor.

The stories of horrific working conditions at Amazon are well-known. Long before the campaign at Bessemer, anyone paying even scant attention would be aware that workers toil at such a grueling pace that they resort to urinating in bottles so as not to get disciplined for taking too much time to use the facilities, which the company calls “time off task.” Christian Smalls was fired a year ago for speaking publicly about people not getting personal protective equipment in his Amazon facility, in bright-blue state New York. Jennifer Bates, the Amazon employee from the Bessemer warehouse, delivered testimony to Congress that would make your stomach turn. Workers at Amazon desperately need to unionize, in Alabama, Germany—and any other place where the high-tech, futuristic employer with medieval attitudes about employees sets up a job site of any kind. With conditions so bad, what explains the defeat in Bessemer?

Three factors weigh heavily in any unionization election: the outrageously vicious behavior of employers—some of it illegal, most fully legal—including harassing and intimidating workers, and telling bold lies (which, outside of countries with openly repressive governments, is unique to the United States); the strategies and tactics used in the campaign by the organizers; and the broader social-political context in which the union election is being held.

Union Busting

Given Amazon’s total efficiency in delivering packages, and its new dominance in Hollywood as a key film and TV series producer and financier, it’s not difficult to imagine that its union busting operation is top of the line. There is nothing new about the ruthless nature of employer campaigns to defeat unions. If you need a refresher, read Confessions of a Union Buster, by Martin Jay Levitt, published in 1988. The book is written by a former hired hand of employers. It’s filled with swagger, as it should be, given how many campaigns Levitt helped destroy. In it, he tells the reader, “Union busting is a field populated by bullies and built on deceit. A campaign against a union is an assault on individuals and a war against the truth. As such, it is a war without honor. The only way to bust a union is to lie, distort, manipulate, threaten, and always, always attack. Each ‘union prevention’ campaign, as the wars are called, turns on a combined strategy of disinformation and personal assault.”

One perusal of Levitt’s book—which should be required for all union organizers and activists—and you get a sense of just how deeply stacked the deck is against workers trying to form unions in the United States (and, increasingly, the world, as union busting is now a hot service-industry export). His book, the Amazon campaign, and just about every union election since the Reagan era are proof enough that to stand any chance of reversing the diminishing fortunes of America’s workers, HR 842, the Protecting the Right to Organize [PRO] Act of 2021, which just passed the House, is desperately needed.

Support for unions today is at record highs, and support for big business is at historic lows. Sadly, popular support for any proposal has little if nothing to do with legislation getting approved by Congress. Given the history of failed attempts at progressive labor law change under Democrat-controlled administrations—even with majorities in both houses—passage of the PRO Act seems like a long shot. But in spite of the many obstacles thrown in the path of workers trying to form unions, lifting up the strategies and tactics that have led to the most successful outcomes is crucial. To accept the defeat of hard-to-win unionization campaigns is to accept a very bleak future. To stand a chance at winning the hardest campaigns, the best methods must be deployed from the earliest days of a campaign and followed throughout the election. Wishful thinking and inexperienced hunches have no place in a campaign against an employer as sophisticated and well-resourced as Amazon.

The Many Warning Signs in the Campaign

Inaccurate list of workers. From the get-go, the campaign in Bessemer had what many experienced organizers recognized as nearly fatal flaws. The first of these was a widely inaccurate assumption about how many employees worked in the warehouse. When the union filed the official paperwork with the NLRB to hold the election on November 20, 2020—a time when few were paying attention to anything other than the United States presidential election—the Retail, Wholesale and Department Store Union (RWDSU) assumed that there were 1,500 workers at the warehouse. Not long after the RWDSU filed, Amazon responded to the NLRB that there were approximately 5,800 workers in the warehouse. Once the union took this first step in the process—the official act of filing for the unionization election—it triggered the truly byzantine legal process that governs union elections in the United States. Amazon’s lawyers argued that if the union believed only 1,500 employees were eligible to vote in the election, the union wouldn’t have what is called a sufficient “showing of interest,” which requires 30 percent of all employees’ having signed authorization cards stating that they wanted to hold a union election.

In a sign that might have seemed encouraging to the union organizers, they were able, between late November and mid-December, to gather enough additional workers’ signatures to meet the minimum 30 percent threshold to hold an election, even of the much larger number of workers Amazon said were eligible (to achieve 30 percent of 1,500 employees, the threshold would be 450 cards signed, but 1,740 were needed to reach that threshold for 5,800). In fact, according to The New York Times, the organizers had gathered a total of 2,000 authorization cards by late December. Less-experienced organizers are often confused when workers sign an authorization card to hold an election—and then vote no. Experienced organizers never frame the question as, “Do you want the right to vote whether or not to have a union?” We ask them to commit to vote yes and to sign a petition saying so when they sign the authorization card for the election. These are very different questions, and they get you very different results in the end. But the showing of interest set the stage for the next step in the cumbersome, clunky process: the official NLRB hearing that determines whether an election will be held and, if so, on what terms. That hearing took place on December 20. From then until late January, with the country’s attention closely focused on the insurrection at the Capitol, Amazon was engaging in its own attack on democracy—that supposedly guaranteed to people inside the workplace.

Handling discussions about union dues. The next warning signs came in February, when Amazon launched www.doitwithoutdues.com, a website describing all the things workers could do with the money they would otherwise pay in dues to a union. Amazon concurrently posted a hashtag on Twitter. The ploy backfired. Pro-union activists around the country took to the platform to tweet one clever response after another, making the reactions to Amazon almost as much of an obsession as Twitter jokes about the ship stuck in the Suez Canal. To organizers, however funny the national pushback on Twitter was—giving people a digitized platform to show their disapproval of Amazon—a deeper cause for concern was revealed in the official response from the RWDSU. Its national president, Stuart Appelbaum, and other campaign surrogates went on an offensive to prove Amazon’s leadership was lying. “Amazon is trying to make dues the issue, even though people don’t have to pay dues,” Appelbaum told The Washington Post. Similar messaging dominated the coverage in response to the entirely predictable union-busting message about dues, with one union official telling NPR, “As some workers point out, Alabama’s ‘right to work’ laws say employees can opt out of paying union dues.”

Although the union response was accurate—workers don’t have to pay dues in a right-to-work state—successful campaign organizers never suggest workers can opt out of paying dues. Just the opposite. Organizers skilled in hard conversations will have first done heavy inoculation about the dues issue long before the employer does, because it is entirely predictable that there will be signs everywhere—in bathrooms, lunchrooms, break rooms, beside time clocks, you name it—suggesting you get more from the company than you do by paying dues to a union. A more nuanced response is one in which the organizers ask workers why the company suddenly wants to discuss how workers spend their own money. The organizers can then help the worker understand that paying dues is essential to build the power required to take on monstrous employers like Amazon.

The semantics and messaging raised concerns beyond the dues conversation. In pro-union placard after pro-union placard, messages proclaimed such slogans as “The union is on your side.” In the many videos flowing out of Bessemer on social media, activists and organizers regularly talk about “the union,” as if a union is something other than the workers who are trying to form one. A better slogan would have been, “When workers unite, real change happens,” or anything that didn’t make “the union” sound like a building name or street address. Plant gate as focus, no house calls. In the vast majority of successful campaigns, how and where conversations with workers take place is crucial. On a web search engine, if you enter “Amazon changes traffic light pattern in Alabama,” the results show dozens of stories, highlighting one of many tactics Amazon deployed to frustrate the activists and organizers in the campaign. While nefarious, it’s completely within the norm of hard unionization fights in the United States. On Twitter, when the story first hit, people who had experienced the same thing took to social media to say, “Yeah, that happened in northern Ohio, too—in our election where the company dominates town politics.” None of these tactics are surprising with one read of Confessions of a Union Buster. What was concerning to experienced organizers, however, was the realization that the majority of the face-to-face contacts with workers were happening at the plant gate.

Like explaining, “Don’t worry. You don’t have to pay dues in Alabama,” working a campaign from the vantage point of the plant gate is another tactic that successful organizers never use. Why? Because the employer is watching. That holds for all employers, let alone Amazon‚—a company that actually develops surveillance systems. The last thing nervous workers want is to be seen near the place they work, talking with union supporters. Successful campaigns require house calls—unannounced physical visits to workers’ homes so the conversation can be had away from the company’s watchful eye. In an interview in The American Prospect, an organizer in the Amazon campaign explained that they were not house-calling, because of the Covid pandemic. The union’s communications director confirmed to me that there had been no house-calling because of Covid. But in a hard-to-win campaign, you should put on a mask, ring the doorbell, have your sanitizer dangling from your chest or in your hands so it’s obvious, and step back and engage the worker, socially distanced but securely.

This very issue, thinking Covid meant no door-knocking, also came up early in the Biden campaign, after Sanders withdrew and the pandemic was intensifying. And Biden erred early on, later making a course correction when he realized how tight the election was. Similarly, in articles about the hard-to-win Georgia Senate runoff races, voting-rights organizers were clear they had to pound the pavement, get in their cars, and make house calls to every voter face-to-face, despite the pandemic. They put on masks and visited voters by the tens of thousands. The most comprehensive academic research on successful unionization, by Kate Bronfenbrenner, director of labor research education at Cornell University, makes an irrefutable case for house-calling. Yet the Bessemer organizers declared that they were relying on “digital strategies.” The union also reports that, other than standing at plant gates, the primary way organizers made contact with workers was calls from other unions across the country, phoning Bessemer Amazon workers. There is no substitute for house-calling in a hard campaign, period.

One possible exception to the plant gate rule would have been if large numbers of actual Bessemer Amazon workers were the people standing at shift change at the plant gate. But that wasn’t the case. Instead, what workers saw was the paid staff of the union and outside supporters.

Absence of majority support. One of the most important tactics in hard unionization fights is what organizers call majority public structure tests. A majority public structure test is when a majority of workers who are eligible to vote in an upcoming union election, or who are voting to strike, sign a petition or take photos and produce a public poster, flyer, or website that displays their signature or faces, with a message stating their intent to vote yes. When asked why that wasn’t done in Bessemer, the union’s communications director told me it had to “protect the workforce” from being fired, so it didn’t want to do anything in public. Game over.

A common error made in hard union battles is for the effort’s leaders to think there’s something unique about their particular circumstances—the industry, group of workers, type of workers, region of the country, time in history, level of surveillance, and so forth—that justifies not following good organizing practices, like performing structure tests and eventually making those tests public once you reach a majority. When fear is running hard inside a facility—which it certainly was in the Amazon election—the only way to overcome it is by asking each pro-union worker to step out and declare themselves pro-union publicly. What “protects the workers” is when a majority of them take this action together, all at once. You are teaching collective power in the conversations and actions.

Structure tests are run over and over privately and quietly until a majority of workers are willing to sign. Once a majority signs, that is generally a serious indication that the campaign will be victorious. But you don’t stop with one public structure test. You keep going, because there’s generally an explosion of additional support once the hesitant coworkers who are holding back realize that, in fact, a majority of their coworkers are uniting. Majority structure tests prove that the people workers in a high-fear campaign trust the most—their own coworkers—are ready to declare for themselves that enough is enough.

Workers watching coworkers take a stand in large numbers is what wins, not rallies with out-of-state superstars, not famous football players, not famous actors and actresses, not even Bernie Sanders or the president of the United States. (Though President Biden’s video is still worth applauding for many reasons: future campaigns and the general legitimacy of unions, not the least.) When there are more outside supporters and staff being quoted and featured in a campaign than there are workers from the facility, that’s a clear sign that defeat is looming.

The Context of Bessemer

Much ado has been made of the history of Bessemer, and by extension, Birmingham, as a place that—although it is in red-as-cadmium-red Alabama, a state that feels stuck in Jim Crow, with among the lowest unionization rates in the country—is somehow an exception because of the town’s history of unionization. And what an amazing history! There are wonderful stories of workers’ struggles to unionize in the region, of Black workers uniting with one another and with white workers—sacrificing their lives at times—to forge unions in the mines that once dotted the landscape. There’s also been a lot of emphasis in the media coverage of the percentage of workers who are Black in the Amazon warehouse, suggesting that demography is destiny. If the latter were true, there would have been a solid victory in the Nissan election in Canton, Miss., in 2017, when the media also grossly overhyped the election and overplayed the factor of a Black worker majority. In that election, the vote was 38 percent yes, 62 percent no.

If you scan the literature that the union has posted on its website, you see a long list of national endorsers of the workers—and a much smaller list of local groups supporting the workers’ efforts. The media often played up the faith-based aspect of the campaign, with key staff of the effort being faith leaders or people of faith themselves. But there was a near-total absence of Bessemer or local Birmingham faith organizations on the endorsement list of the campaign. Although news reports often highlighted that meetings started with prayers, there was an absence of major local faith leaders publicly supporting the workers. In successful campaigns, overt public support from area faith leaders is often key where religious affiliation is common among the workers.

In Detroit, before auto plants were organized, many Black people were opposed to the union. The reason, according to Dr. Steven Pitts, who runs the new podcast Black Work Talk: “Many of the prominent ministers in the Detroit area had good relationships with Henry Ford. As the Black migration from the South was at its peak, Black families settled into Detroit, found a church, and got a job at the Ford plant through their church leaders. It would take a decade of struggle between pro-union Black workers and their civic leaders before the dynamic swung to Detroit turning pro union.”

In Bessemer, local community groups that I reached by phone told me that this was the first union campaign they could remember when the union reached out for the first time so late in the campaign—as late as February. In other elections in the area, even the Tuscaloosa Mercedes auto parts plant election, where the plant was 50 miles away from the city—a lot farther than Bessemer from Birmingham—the unions tended to reach out long before the workers even went public in the campaign (which happened in October in Bessemer).

In the next few days, you’ll probably see a lot of messaging claiming that “even though the workers didn’t win, they really did win.” But they didn’t. And that is horribly unfortunate. The media, especially the genre of media called the labor media, should have never overhyped this campaign—or the Volkswagen campaign, or the Nissan campaign. In all three cases, the impending defeat was evident everywhere. When media folks prioritize clicks and followers over reality, it doesn’t help workers, and probably hurts them. The coverage heaped a mountain of unwarranted attention that might serve the media narrative behind the PRO Act, but overhyped campaigns also leave people feeling defeated. Sometimes, in fact, they feel so defeated that they withdraw and give up forever. This campaign likely should not have been run once the organizers realized how off their assessment was of how many workers were actually in the warehouse. There’s no justification for putting workers on what organizers call a “death march.”

For Bessemer workers, the likely next step is that the union will file a huge number of totally justified objections, or “Unfair Labor Practice” charges, against Amazon. They are likely to win the right to another election based on the illegal behavior of the employer. In the legendary Smithfield organizing campaign, where workers at the nation’s largest hog slaughterhouse won their union after a 16-year fight, on their third election attempt, the lesson people should have taken is that, yes, labor law is broken—but also, you don’t take shortcuts when you run a campaign. Many of the same limitations in the first round of Smithfield were true in the first run in Bessemer. It’s time we don’t take workers for granted, we don’t sell them short, and we don’t create scorched earth haphazardly.

The conditions most workers in the United States endure when trying to form a union make the recent actions by Georgia’s legislature to institute further voter suppression seem tame. If the Senate passes the PRO Act, there’s no question the unionization rate would increase quickly, which is one reason winning its passage in the near future seems oddly distant. Despite the nation’s having the most pro-union president in nearly 100 years, the Senate remains immovable on issues far less challenging than major labor law reform; it wouldn’t even accept a federally mandated $15-an-hour minimum wage. And progressives have been trying to pass labor law since Jimmy Carter’s presidency—without success.

Every worker in the Bessemer campaign deserved to win. And if the rules for unionization in the United States came close to being fair, they would have won. But the rules aren’t fair. Quite the opposite: They are outrageously unfair. What workers trying to form unions against immoral employers do deserve is the kind of effort that stands a chance of winning. There’s plenty of evidence of what works. Social media and shortcut digital approaches don’t work when fear and division is the central weapon.

Workers can win unions—and workers can strike and win. It is hard as hell, and to do that requires a no-shortcuts approach.