Labor leaders lament the external assaults on unions, but this is rarely extended to confronting unions’ internal deficiencies. When labor actions are successful, unions and their allies are pleased to see praise in print. When success is ambiguous, they tend to get defensive and stay clear of open public discussions. We shouldn’t, they proclaim, “wash labor’s dirty linen in public.” Negatives, they warn, will further “demoralize workers and encourage their enemies.”

Critics outside labor are dismissed as not understanding workers nor appreciating the scale and complexities of taking on the second-largest private employer in the United States. Moreover, critics may see the surface impact of a labor action but are in no position to grasp the behind-the-scenes qualitative developments.

These are legitimate counters, and they warrant some humility on the part of even friendly critics. Yet those who look to the working class to lead social change do the class no favor when, in the name of supporting workers, they pull their punches or patronizingly “cheerlead” and do their best to contain the internal debates so vital to moving the labor movement to a new level.

Leadership discussions behind closed doors won’t cut it; workers can only develop through actively participating in the back and forth of honest and sometimes hard debates. Nor will leaving it to unions’ monthly meetings be enough. Attendance is generally very low, the meetings often inhibit rather than encourage controversy, and addressing the larger issues demands spaces that include others in the labor movement. If the working class is to be fully informed and actively engaged, and if the broader labor movement and socialist left are to be brought into the Amazon struggle, there is no alternative but to allow the discussions to spill into the public domain.

Such openness may pose risks, but the greater risk lies in an unchallenged labor movement continuing to drift. Without confidence in the ability of workers to “get it” and rise to the occasion if appropriate structures and strategies are in place, the desired transformations in the labor movement and society can’t win out.

In this high-stakes context, the president of the Teamsters, Sean O’Brien, dramatically announced that if Amazon does not recognize the union and come to the bargaining table by December 15, it will face a “national strike.” Amazon, of course, ignored the invitation, and a few days after the deadline, the strike was launched. As a national strike, the action — a few exceptions aside — was a flop. It was neither national nor noticeably impacted Amazon operations. It did not win concessions from the employer, and the number of workers involved was disappointingly low.

But — and this is critical — while numbers and economic outcome matters, what matters more in developing an eventual winning strategy is whether the action contributed to raising working-class expectations and building working-class capacities. On this score, those involved at the base seem to have salvaged some positives. To the extent, however, that this proves true, it emerged in spite of, not because of, the strategy hurriedly proclaimed from above.

The Teamsters’ strategy seemed driven by getting media attention rather than the development of the deep worker base fundamental to forcing Amazon to accept a union. Why else suddenly declare a national action when the base for such a test of worker strength had clearly not yet been developed? Why else declare a “strike”, with its intimations of staying out until the goals have been achieved? Why take on the liabilities of a bravado measure such as a national strike when a substantive and creative protest could have achieved as much or more?

The first rumors that O’Brien was shooting for a decisive action before the end of the year had surfaced in the summer of 2024. Perhaps he and those around him saw this as a “structure test” to assess whether organizing Amazon was possible. Perhaps they were concerned to take advantage of a relatively sympathetic National Labor Relations Board (NLRB) before it was replaced by an expected Donald Trump victory. Or maybe this just reflected the strategic weight given by O’Brien to the media’s potential role in isolating Amazon and accelerating interest among workers in joining the union. We don’t know.

What we do know is that going in, the on-the-ground organizers knew full well that a national strike was premature. And based on the modest scale of the number of workers and sites the Teamster leadership targeted to go out, it seems that they too must have had misgivings about the degree of support. The ten thousand workers the Teamsters announced would be walking out was impressive for a protest, but at only 1.4 percent of Amazon’s 740,000 blue-collar workers, not awe-inspiring for a “national strike.”

The same is true of the number of sites to be shut down. Of the well over three hundred fulfillment centers (FCs) in the United States, only two were listed to strike. Of the over seven hundred delivery stations, only six were to walk out. Half of the eight strike sites were in California: the pivotal air hub in San Bernardino with its over one thousand workers and two delivery stations also in Southern California, plus another in the Bay Area.

Another two were in New York: a relatively large distribution center in Queens and the mammoth Staten Island fulfillment center, the first Amazon certified site in North America. Staten Island’s workforce of close to 5,500 represented more than half the total numbers that were to walk out nationally. The remaining two strike sites were delivery stations in Atlanta and Skokie (a half hour from Chicago).

As it turned out, even of the ten thousand that were to strike, only a small portion actually walked out. According to an investigation by Luis Feliz Leon, barely six hundred workers across the United States answered the strike call. (Including all forms of participation, the numbers would be at least double, perhaps even a bit higher, but this does not reverse the dominant story.) At the high-profile Staten Island facility, where workers faced perhaps the most aggressive Amazon response, estimates put the number at no more than two hundred to four hundred workers, meaning that well over 90 percent of workers stayed on the job.

Nonunionized workers in the United States (though not in Canada) have the legal right to take “concerted action,” which includes the right to collectively withhold labor. The Teamsters built on this by raising Amazon strike pay to cover the straight-time wages lost in striking. Amazon, determined to undermine this tactic, responded by warning workers they could still be fired if not at work. How could Amazon do this? A brief, but somewhat complex detour to answer this question tellingly exposes the limits of laws in the absence of the power to enforce them.

Amazon workers accumulate annual unpaid time off (UPT) to use as desired. If their time away from work exceeds these days, the workers are deemed absent and vulnerable to dismissal. The catch lies in the fact that, for all of Amazon’s technical expertise, the company has not programmed its computers to distinguish between regular time off and the separate time off for a protected legal strike. As the end of the year approached, workers had used up most of their UPT time, and so, if they went on strike, these extra days off would — wrongly — move their time off accounts into negative territory. Although Amazon was obligated to correct this (see NLRB JD09-25), the lapse in time leaves workers uncertain about where they stand, affecting their readiness to go out on an ostensibly legal strike.

Any firings will no doubt be challenged at the NLRB, but unlike corporations who can get immediate injunctions to protect their private property, workers can’t get an injunction to stop Amazon from using administrative mechanisms to undermine confidence in their legally protected right to protest and strike. So, workers’ fears trumped the generous Teamster strike pay.

All this aside, the Teamster offer of strike pay at the level of full wages is itself a questionable tactic. It suggests that workers can strike without worrying about lost income, which is clearly unsustainable. It sets a precedent for workers expecting similar strike pay next time a strike is being considered. And it can alienate other Teamsters whose dues are subsidizing the unionization of Amazon but whose own average strike pay is only about a third of what Amazon workers were getting.

Against those who dismiss the successes of the national strike, the organizers most closely involved emphasize valuable achievements. The strike made more workers aware that a union drive was going on and demonstrated that the Teamsters had their back. The Teamsters provided additional full-time organizers, booked off members from various Teamster locals to man picket lines, and financed the high-income support. New contacts were made (in this instance, the media focus had some value). Many activists matured into leaders, and leaders further developed their organizing and tactical skills. Young activists outside the Teamsters, especially from Democratic Socialists of America (DSA), provided picket line support and forged stronger links with the Amazon workers and organizers.

The coming months will tell us more about the balance of negatives and positives, but one immediate observation, highlighted by some organizers, points to the unevenness of the outcomes. The level of past organizing, even if slower and less dramatic than fanciful targets aimed at the media, seemed decisive. Where the previous level of organizing was most developed and Teamster locals or DSA chapters were active, such as in Southern California and in Queens, the unionization drive seems to be emerging stronger. Where the previous organizing was still rudimentary, the negatives seem to have carried more weight.

A central element in the withering of union strength since at least the 1980s has been the postwar tendency to prioritize wages and monetary benefits over working conditions and — closely related — limiting collective action to contract expirations. Power within the unions consequently shifted from the workplace to the union officials leading bargaining every three to five years. Worker participation narrowed to voting for or against tentative agreements and occasional stints on picket lines with few workers going on strike more than once in their lifetime.

This is hardly favorable ground for developing worker strength and leadership. It leads to the common view of unions as insurance agencies in which workers pay dues in exchange for benefits delivered by a separate body: “the union.” This is a process that invites both a high degree of apathy and a tendency toward bureaucratization. In the case of Amazon, moreover, announcing a strike deadline weeks or months in advance gives the corporation, with its powerful logistical capacities, time to prepare for subverting work stoppages. This points to the importance of thinking through other options.

The starting point for one such option is Amazon’s core promise of reliable, rapid delivery. To accomplish this, Amazon has structured its operations around relatively distinct regional clusters of facilities. Building worker power regionally is consequently possible and more practical than waiting for national success. Possible because a regional capacity to disrupt cannot easily be overcome by Amazon. Practical because organizing in key centers like New York or Chicago is doable, while waiting to organize the whole country raises the likelihood that by the time the final sites are being unionized, the base at the first sites unionized years earlier may have faded away.

A regional focus opens the door to creative strategies that can reduce the corporate-union imbalance of power. The United Auto Workers (UAW) demonstrated such potentials in the first auto bargaining round under President Shawn Fain. It set aside the union’s traditional model of striking the entire company with unannounced rotating strikes at specific sites. The Teamsters could similarly choose a region where the members are strongest, concentrate their resources there, then move on in a few months to a second region, then a third and so on.

Such rotating actions would give each region time to prepare, and it would stretch the momentum of the struggle across time as opposed to a single national strike with a limited duration. (Periodic national actions could still complement the regional battles.) The point is not to “bring Amazon to its knees” — rarely do unions develop and sustain that kind of power — but rather to prevent Amazon from conducting business as usual. Such sustained guerrilla warfare would aim at getting regional recognition of the union and concessions from Amazon as the condition for it being “allowed” to get on with its business and avoid the frustrations and alienation of its customers.

Within a region, the unionization of any site is, of course, welcome. The strategic issue, however, is which sites are priorities. In this regard, the fulfillment centers are pivotal. Finding choke points like delivery stations or sort stations may give workers some temporary leverage. But the work of a delivery station on strike can be transferred or subcontracted, and sort stations can be worked around by rerouting trucks.

Though the Teamsters have an understandable bias for unionizing drivers, the advantages of emphasizing the FCs are that, for one, they include by far the most workers and will also get the most attention from the rest of the labor movement. For another, they are the least likely to be closed because Amazon’s business model must keep it in the region, and the investments involved in setting up FCs are so costly. Furthermore, as Stephen Maher and Scott Aquanno have convincingly argued, the FCs are the foundation of Amazon’s technical and financial strength.

Within this strategy of identifying the region as the principal organizing space, rotating strikes across regions as a key tactic, and FCs as the primary organizing site, the workplace remains the decisive unit of confrontation. It is here that effective unionization must develop the workers’ capacities to creatively disrupt — not only when the collective agreement is over, but during the life of the agreement.

This cannot happen without widespread daily worker participation in contesting production rates, health and safety standards, work schedules, and management arbitrariness. It demands the deep organizing that builds working-class independence, confidence, and strength to challenge management on a regular basis. The disruptions that result could range from a walkout of a specific shift or a particular department, to shutting a facility for two hours, leaving Amazon uncertain about whether it is worth trying to work around the work stoppage. Or shutdowns in large urban centers could take place where, for example, hitting two or three FCs that handle large products would be especially hard to get around.

Out of this comes the possibility of not only winning union recognition but also a union of a particular kind — a union able perhaps to also inspire other unions to do the same.

As Amazon organizers, committees, and workers prepare for “next time,” some broader observations, gleaned from working with Amazon workers in Canada and rich discussions with Amazon comrades in the United States, are worth considering.

1) Unionizing the unorganized begins with getting the already organized on side

It is generally not too difficult for union leadership to pass resolutions committed to supporting a major unionization campaign. But if that commitment is thin, then the support will be fragile. Membership support will fade as the union drive drags on, leading to questions being raised about the campaign diverting attention from the needs of the current members. Unions leaders and organizers will start looking over their shoulders, and the organizing commitments will become tentative.

Moreover, already unionized workers in nearby locals are a potentially powerful resource to activate in a unionizing campaign. Unless these union members have been brought intensely on side, recruiting the hundreds of local union members needed to actively participate in the drive can’t materialize.

2) Should unionizing Amazon be a union project or the project of the whole labor movement

D. Taylor, retired president of Unite Here has argued that taking on the “commanding heights of the economy” as with Amazon “will take not one union but a powerful coalition of unions, a force like the CIO [Congress of Industrial Organizations] in the 1930s.” The intent is right and admirable. It would be mind-blowing if the unionization of Amazon became a crusade of the labor movement. This is, however, easier said than done, especially in a union movement where raids and competition for dues is more common than concrete solidarity. Ditto where there have been virtually no serious discussions among unions, private or public, about concerted unionization efforts. And with the Teamsters taking over the unionization of Amazon, a joint effort seems closed.

Should Amazon get unionized, and a strike emerge with Amazon looking to bust the union from the start, it may become easier to generate concrete solidarity across unions. Hot cargo actions could spread across Amazon’s supply and delivery chain: from port workers, railworkers, and truck drivers to airport ground crews and mail carriers. But this would have to be planned for in advance to win workers over to showing solidarity and risking the uncertainties. This suggests a role for all unions immediately in support of the Teamster organizing drive.

3) Solidarity now

Unions should be setting up Amazon support committees in each local to carry out education around what is at stake in the battle for Amazon; getting contacts from workers, family members, and friends who might know someone working at Amazon; developing volunteers to support the organizing; and preparing for possible actions. Labor councils and state federations could encourage and supplement this support. All of this would contribute to Amazon workers seeing they are not alone.

In Canada, where the Teamsters have not yet established themselves as the union of record at Amazon, there is still the possibility — however distant — that a council of unions could join in supporting the unionization campaigns and establishing a cross-union solidarity fund to support organizing. If this should surprisingly occur, this itself would mark a historic turnaround in union cooperation.

The closure of Amazon’s Quebec operations has posed the question of how Canadian labor might respond. Amazon, its gobbledygook explanations aside, feared that Quebec’s labor laws would impose a settlement in a certified delivery station and set a precedent, inspiring workers throughout North America. Its response, the closing of all existing facilities in Quebec, was eased by the fact that much of the Quebec market was already being served from nearby Ontario.

One constructive labor response would be for the Quebec union involved (Confédération des syndicats nationaux) to declare it is going to join forces with unions in Ontario to help organize the runaway work there. Another would be to have all unions in Canada join a campaign to spread the strengths of Quebec’s labor law across Canada.

4) Amazon vs. the post office

Amazon had previously looked to the post office to deliver a significant number of its packages because of the post office’s lower fees and available infrastructure. As government austerity reduced government support for the post office, the post office was compelled to raise its fees, and as a result, Amazon pulled back from using it, threatening the post office’s financial viability. Trump has now gone further and raised the possibility of privatizing the post office, playing directly into Amazon’s hunger for taking over even more of the delivery sector.

There will be campaigns to defend the public postal services. But why not use the threat to go on the offensive and reverse this process? Why not defend the existence of the post office by arguing that it is Amazon — in the business of delivering packages to homes and with the power of a public utility but in private hands — who should be placed in the public domain and integrated into the post office where it would serve social goals and meet labor standards?

5) What can socialists entering unions do differently?

The organizing of Amazon has involved a good many post–Bernie Sanders young socialists committed to the labor movement. Many of them are now involved in the Teamster organizing. This poses a difficult, and so far, unanswered question: How do socialists and militants operate within unions that are heavily top-down and dense institutions?

The fact that the Teamster leadership determined the timing and strategy without much concern for the input of many of its best organizers highlights the difficulties. So, too, does the fact that the long-present Teamsters for a Democratic Union — with its courageous history — has ended up, both at United Parcel Service (UPS) and during this strike as an in-house opposition with little independence, further underscoring this dilemma.

An effective response demands that it be a larger organizational one rather than one based on scattered groups of committed workers doing what they can to transform the Teamsters. It necessitates an independent left organized across Amazon with a politicized training program, regular communication and strategic discussions by way of a newsletter, a capacity to recruit salts, and links to socialist supporters in other unions and in the community. Just posing these issues points to the existence of a coherent socialist party likely being a condition for working effectively within unions with the goal of a deep-seated revival of the labor movement.