“Organizing isn’t reading Das Kapital,” proclaimed labor organizer Emma Kinema in a 2022 interview with Reboot. That is: when faced with poor working conditions, theory alone won’t solve your problems. A union, however, might.
In today’s essay, writer and software engineer Josh Erb tells the first-person story of the Mapbox Union effort: what inspired him to organize, the opposition the union faced, and what he learned about power and solidarity along the way.
— Jasmine Sun, Reboot Editorial Board
Nearest Neighbors
By Josh Erb
The first step is courage
It takes guts to say “union” to a coworker and mean it. The first time I said it to another person, I was sitting on a bench in Lafayette Square in Washington, D.C. just North of the White House. It was a brisk, sunny afternoon in the Spring of 2019.
At the time, the word carried no weight. I was talking to a coworker who had just quit a software engineering job at our company, Mapbox, effective immediately. We had been talking about how powerless they had felt when Mapbox effectively forced them to move from New York City to Washington, D.C. How it shouldn't be so easy for leadership to wake up and change so many people's lives. Employees should be able to push back on those decisions. I said the word "union" to fill in what felt like the natural conclusion to the thought.
Several months prior to that conversation, and shortly after a large round of funding, Mapbox leadership seemingly decided they weren't comfortable allowing “junior” engineers to work remotely. When I first joined the company in 2018, there were no engineering levels. The mandated relocation of junior engineers came shortly after defining the category. It is harder—went the apparent rationale—to cultivate engineering talent when it is not in the same city as management. All junior engineers now had a choice: move to one of the primary cities Mapbox had offices in, or accept a severance package.
What would a union do?
When people ask me why we decided to form a union at Mapbox, I have trouble giving a consistent answer. If you ask a dozen of us that question you would get thirteen different answers. It’s common to think a union is a group of people who have rallied around a single goal and agree on the best way to get from point A to point B. In practice, however, a union is rarely so coherent and its goals are never so static. It is the process of collecting all of our different concerns and values and weaving them together into a shared vision of the future we want to build. Our union was the vehicle to a better, more certain future. It was never the final destination.
In a similar vein, my own motivation has always been hard for me to pin down. By all accounts, Mapbox was the best job I've ever had. Mapbox built its reputation by being a developer-friendly underdog that provided better performance and more flexibility for map and location services than the sleepy, long-established companies that still used terms like "Geospatial Information Systems" in their product names. By the time I joined in 2018, its products were being used by rideshare companies, social media apps, weather apps, data analytics firms, and news outlets. More concretely, it gave me career opportunities, paid me well, and the benefits were better than any job I had before or since.
But there were cracks, of course. Forced relocations and repeated, unexpected layoffs. In these moments the curtain would slip and the stark power imbalance of my work life would be briefly thrown into sharp relief. No workplace is a utopia. The longer you stay, the harder it is not to see the cracks.
When discussing tech industry unions, the first question is usually 'why do you need one?' Unions are a tool for workers with limited individual power to build and exert influence on those above them. In tech, workers have far more power and wealth than their blue collar counterparts. Still, the tech industry is not solely made up of well-paid software engineers and product designers. Even if it were, these jobs are not uniquely stable. The recent spate of tech layoffs has taught us that even high-prestige development jobs can easily be cut. And even if you trust your leadership, you are one hostile takeover away from having the rug pulled out from under you. As an individual employee, you have no recourse beyond hoping you can find a new job before savings run out and bills come due.
My vantage point changed in 2020. Participating in the protests against police brutality in D.C., I experienced first-hand the righteous anger and catharsis of mass protest. I saw the potential for well-coordinated groups of people to break through complacency and create space for the possibility of change.
For the first time, I saw with clear eyes the connection between the private and the public. Between the cycles of tragedy and protest and the long periods of complacency. How could we hope to effect change at a national or even municipal level, if we couldn't even make progress in the place we spent the majority of our waking lives?
The inconvenience of organizing
A union is not something magically conjured by law or leadership's recognition. The union exists from the moment two or more coworkers agree to talk earnestly about their working conditions and, crucially, start coordinating to shape them.
Even before we publicly announced our contract campaign at Mapbox, the value of our union was evident. We were able to advise and encourage coworkers who had come to the U.S. on work visas and discovered they were being drastically underpaid, helping them aggressively negotiate for higher pay. We coordinated our questions for all hands meetings and pressured leadership to enforce their stated policies for the ethical use of our services. There was a moment of powerful commiseration when several team members realized that they had been put on “performance improvement plans” (PIPS) because manager rollover had led to confusion about their day-to-day work. We weren’t just bonding over shared frustrations, we were taking action to address them. We had built a dependable network of trust and support.
As the campaign gained momentum and the possibility of formal recognition from Mapbox was becoming tangible, I was committing more hours to the campaign each day than should have been humanly possible. On top of my work responsibilities, I made time to check in with coworkers about the campaign every week. Our organizing committee met regularly to assess our progress and adjust plans accordingly. Solidarity could be grueling, inconvenient work. You cannot automate its maintenance or assume it will continue to exist and grow of its own accord. At the end of each day I was drained, amazed that I had anything left to give. Then, awed by the miracle of how energized I felt when I woke up each day to do it all again.
We publicly announced our campaign for formal recognition by sending an email to leadership in June of 2021. There was an effervescent sense of joy when we made the announcement. But we only held onto this feeling for a week. The following Tuesday, leadership made it clear that they would not be voluntarily recognizing us. Our CEO announced that he was setting aside all other priorities to focus on addressing this "union issue." The counter campaign kicked into high gear. Leadership hired an anti-union law firm and a labor consultant to help “educate us” on the dangers of unionizing. Our calendars filled up with meetings where leadership and managers talked to us about how scared they were for the future of our company. After a month of intense pushback, coworkers started avoiding us. Colleagues once strongly in support began to accept the story that leadership was telling them.
One conversation in particular sticks out. I was checking in with a coworker I had worked with for years. They agreed that many of the problems at the company were caused by the action—or inaction—of leadership. They had been all-in for the union at the outset. But several months later, over a fraught FaceTime call, they told me they were going to vote ‘no’. They weren’t naive, they still saw that leadership was full of shit, and that they were arguing for their interests, not ours. But the campaign had made every aspect of work a slog. It was hard to think about complex systems architecture with all of our managers and international colleagues doom-posting on Slack at all hours. It was hard to write good code when the workday was full of meetings about how awful our union might make everything. Everything had just become too hard.
Modern technology is characterized by an obsession with convenience. Buy a product with one click and have it delivered to your door within a day. Tap a button and have a private car appear at your location in minutes. This mentality isn't exclusive to product offerings. It seeps into almost every aspect of working in the tech industry. Venture capitalists write about “founder mode,” which boils down to founders only being focused on the “big picture” and unbothered by the tedium of any daily administration work. Middle management embraces the philosophy of “Agile” to reduce the friction for shipping new features and delighting their end users. Software engineers are no exception. We talk about “velocity,” that is, how quickly we finish discrete development tasks. Slow velocity on a team is an indication that there is too much friction in the workflow. Convenience and its endless pursuit is the air you breathe at a tech company.
As organizers, we weren’t immune. Throughout the campaign, the impulse to find a more convenient method of doing the work washed over us like waves. There had to be a better method of identifying where coworkers stood, encouraging people to get involved in the effort, or making decisions as an organizing committee. We talked enviously about European labor law. How in Germany, simply demonstrating that a majority of workers were in favor of unionizing legally obligated employers to recognize them. We fantasized about the PRO Act being enacted and a card check sparing us the headache of a potentially bitter fight for recognition.
A recent spate of quixotic startups demonstrate convenience’s allure, promising to simplify the processes of organizing your workplace. Implying, by their existence, that the crisis of unions in the United States is not a disorganized workforce against an entrenched profit-maximizing status quo. Rather it's simply because the task of organizing isn’t frictionless enough. How are workers, who already have full time jobs, expected to also effectively remember to check in with their coworkers and keep track of who is willing to sign a card? Startups like Unit or Frank sell (this is the operative word) the dream that if we just had the right platform, or if our tools were slightly better designed, more workers might stand up together and demand their rightful seat at the table.
Most often these startups announce themselves with a roar, before fading with a whimper. The problem is that they are built on the same market incentives that unions disrupt. These companies—and the VCs who fund them—fail to understand that union drives do not succeed or fail based on the effectiveness of the tools they use. They succeed based on the strength of their members' relationships.
Convenience for whom?
Earlier this year, after Amazon warehouse workers successfully organized for a recognition vote in Coventry, England, the company launched a "one-click to quit the union" tool. Posters with QR codes were put up in break rooms and common areas. They encouraged workers to scan them and end the needless disruption of a union by simply clicking a button. Amazon had made it easier to cancel your union membership than to cancel your Amazon Prime subscription.
The urge to reduce friction is not inherently a negative force. But convenience is often wielded to fortify the complacency of the status quo. One-click purchases depend on relentlessly pushing warehouse workers and delivery drivers to their limits. A website with 99.99% uptime requires site reliability engineers being on call and ready to drop everything and resolve the issue, day or night.
When Mapbox leadership fought our union they appealed to our sense of convenience. Over a fraught two months, they made the relentless case for how inconvenient everyone's life would become if our union were allowed to continue, let alone be officially recognized. Leadership put their own twists on classic anti-union phrases like “a union would slow us down and make it harder for us to innovate,” “the union would make it hard to remove toxic people from your team,” or “we want to fix these structural problems you've highlighted, but it would be harder to implement solutions if we have to have everything approved by the union.” Leadership was effectively outlining the inconvenience of no longer being able to rule the company by decree.
Within a month of our announcement, the tensions within the company had reached a fever pitch. The messaging from company leadership was simple: forming a union made the already hard work of running a competitive startup too difficult. Two arguments from leadership ended up being especially effective at moving workers away from 'yes' and over to 'no'. For some, it was the claim that having a union would make it harder for your direct manager to support you in your career. For others, it was the claim that announcing the campaign had caused the company to lose a round of funding, and that if we had any hint of a union present, it would be difficult to receive future rounds of funding. If we couldn’t raise more funding, we would all find ourselves unemployed and our equity worthless—the stakes were existential.
In September 2021, after succumbing to pressure from Mapbox leadership to withdraw our NLRB petition and conduct our vote for recognition electronically via a third party adjudicator, we ultimately came up short. 81 votes in favor, 123 against. When we initially went public with the campaign, 148 of us had signed our names to the vision statement we sent to leadership asking for union recognition. The final result was crushing, but not unexpected. The anti-campaign had effectively eroded our trust in one another. The fact that we agreed to forego the protection of the NLRB for the expediency of an electronic vote was a sign in and of itself of how much our position had weakened in the final weeks. We knew that our network was patchy and brittle by the time we were allowed to vote.
Network disruption
In the popular imagination, union campaigns pop into existence and then end with either victory or defeat when the final votes are counted. This is far from the reality.
I stayed at Mapbox for four more months after we lost our vote. I was firm in my conviction that the vote didn't decide whether or not we were a union. We were a union the moment we started talking to each other and collectively pushing for change. Recognition and a contract only stamped that process. I threw around the term “minority union” and pointed at our labor siblings over at Alphabet.
But leadership did everything they could to ensure that the final vote also marked the end of our collective efforts. Severance packages were offered and taken. Leadership committed flagrant unfair labor practices (ULPs) by firing members of the core organizing group illegally, as the NLRB eventually ruled. The strategy worked to undermine us, but it came at a cost. Mapbox had once had a vibrant, thriving work culture. With every day that passed the company felt more and more like an abandoned mall. Even those who didn’t vote ‘yes’ were heading for the exits.
In January 2022, I finally admitted to myself that I was burnt out and running on fumes. Our organizing committee was still limping ahead, meeting regularly and talking to the coworkers that remained and would answer our calls. A few months after my final day, there was a layoff event and the remaining members of the committee were let go.
On more pessimistic days, I look back at the Mapbox campaign and I see parallels to the failings of our current political moment. Why nominally popular initiatives like healthcare reform or addressing climate change continuously fail. People agree that there are problems that they want to solve, but the moment that action becomes inconvenient, things fall apart.
Jane MacAlevey begins her book A Collective Bargain by saying, candidly, that “Unions are such a pain in the ass. Really.” My experience with building and sustaining one inclines me to agree. But I've also come to see the inconvenience as a feature to be embraced, not a bug to be fixed. Some of the most fundamental and fulfilling parts of organizing are the inefficient ones. Having an hours-long conversation about the rough patch a coworker is going through. Attending training sessions to improve your organizing skills on a Saturday afternoon. Planning and hosting a union event outside of work hours where workers from all over the company share what motivated them to join and what they hope to accomplish.
The inconvenience isn't the main point, of course. We threw ourselves into unionizing because we wanted to materially change our relationship to the work that we loved. We hoped to protect ourselves from the whiplash of a company where leadership has unilateral power to upend our lives based on a bad market report or a funder's whim. To cultivate a workplace where we could be sure that our coworkers were being treated fairly, our products were being used ethically, and that our jobs would be there for us tomorrow. Fundamentally, it was a question of voice and agency.
Making decisions collectively will always be the least convenient way to make progress. It will be tedious, exhausting, and we will daydream about simpler ways that we might bring it about. But the fact remains, if we want to build a future that benefits more than just the bottom line, we will need to embrace the inconvenience of building it.
Josh Erb (he/him) is a writer and software engineer. You can find him on his website and on bluesky.
Reboot publishes free essays on tech, humanity, and power every week. If you want to keep up with the community, subscribe below ⚡️
🌀 microdoses
This essay only scratches the surface of everything that happened in the Mapbox Union campaign. Organizers will be releasing an oral history podcast that goes much deeper. Sign up here to be notified when it’s available.
- is totally right about that Scott Alexander AI art survey: “People prefer AI art because people prefer bad art.”
I very much want to read this book of Selected Amazon Reviews.
And so are salmon hats 🐟
💝 closing note
A Reboot gift guide is in the works… if you have favorite artifacts (physical, digital, cognitive) to recommend, do let us know!
Gratefully,
— Jasmine & Reboot team
The point about friction is poignant.
I’ve spent most of my career in tech, but my activist time is spent predominantly on housing policy. The org where I help out, Yimby Action, focuses on grassroots organizing and the attitude you allude to - that we just need to /one weird trick/ our way out of our problems - is something we encounter quite frequently from folks in industry.
I think the fundamental disconnect is that doing politics in a democracy means building consensus and engaging in the democratic process as a bloc. Building relationships, organizations, and institutions all help with that by reducing the cost of collective action, but the underlying task is always to convince people of a) a policy position and b) a political strategy.
If your only mental model of working with other humans assumes rigid power hierarchies, the assignment is much more about gathering information, making decisions, and disseminating orders. Challenges that are much more tractable via product design since it kinda assume away a lot of individual agency.
Anyway, that was all a meandering way of saying I think your call out wrt to attitudes is onto something and reflects something deeper in how some of us conceptualize group effort.
Beautifully written - thank you for your openness and vulnerability. Hoping that your story will inspire many more to consider the real aims of "frictionless" technologies. There is no forward progress without the friction between the tire and the road!