Technology has driven us to a fascinating moment in the history of labor relations.
At the very high end of the labor market, workers have never been more empowered, thanks to the internet. Open communication tools and forums inside of companies—and the increased liquidity of the job market—have given workers at companies like Google more power than ever. They are aggressively using their power to play a role in shaping corporate policy and strategy on issues like China and military contracting—with seeming success.
On the flip side, at the low end of the labor market, technology has helped tip the historical balance of power against workers in favor of employers. The rise of on-demand labor markets, coupled with the ability of employers to micro-target incentives in real-time, has undercut workers’ ability to assert real power on platforms like Uber and Lyft.
While the shifts at both the high and low ends pose interesting challenges to the status quo, I—like many—am more concerned about the dynamics at the low end of the market, and how to create a healthy path forward for less-skilled workers.
The question is how to create a healthy path toward greater worker empowerment around labor marketplaces.
It is unreasonable to rely on the benevolence of consumers or corporate platforms. Their interests are simply at odds with workers. Government may have a role to play, but their incentive structure is complicated and the process is too slow to keep up with nimble platforms, as we are seeing play out in real time with California’s AB5.
I believe that workers and labor advocates must become more sophisticated in the tactics they use to advocate for themselves.
Labor needs to learn to fight fire with fire.
In terms of what that practically means, I believe that labor organizers need to update their playbook for strikes and work stoppages. They need to move from 20th century–style mass actions toward technology-driven, highly targeted, micro labor actions that look more like guerrilla warfare, and less like open-field battles (where they will always lose).
The Problem: Why Traditional Strikes Fail in the Context of Modern Labor Marketplaces
Traditional labor strikes are no longer effective, due to the impact of technology on the nature of work.
First, there are far too many eligible workers to possibly coordinate effective work stoppages. In a traditional labor setting, there is a finite labor supply that is able to do a given job in a given factory. In a labor marketplace setting like Uber and Lyft, technology has been deployed to standardize and simplify and focus the human job so that nearly anyone can do it with minimal training.
Thirty years ago, driving a taxicab in London was a high-skilled job. But with GPS, maps, electronic dispatch, etc., nearly anyone is now able to do the job. Expanding the eligible labor supply for a given job has benefits for everyone. But it does make it very difficult for labor to run a successful strike because there are always plenty of people who can cross the picket line in a pinch.
Second, even setting aside the expanded labor pool, the fact that labor platforms tend toward variable employment versus full employment of each person makes it far harder to run an effective strike. If all your workers are fully needed all the time to run a factory, then the impact of a minority of them stopping work is large. If, however, your workers can “flex” up or down the number of hours they work, then your existing team can absorb certain shocks to the system. The workers who continue to work have a far easier time simply picking up the slack, which makes it hard for strikes or stoppages to have impact.
Third, beyond labor pool considerations, there is the powerful management tool of variable, personalized and private incentives. Historically, people were paid a wage, so you might in theory offer people more money to break a strike, but you would do so broadly. The fact that labor marketplaces can dynamically offer workers bonuses, incentives, etc., nearly instantly, personally and privately (where what you offer is not known by everyone), ensures platforms can maintain a supply of workers. People might be willing to execute a work stoppage in a fixed world, but with dynamic incentives the reality is that everyone has a price at which they will come back to work.
The final notable reason that strikes are so difficult with the modern labor platform is that the workforce doesn’t work in a single location, but is geographically isolated and siloed. In a factory strike, the striking workers can simply stand outside the office. It is hard as a worker to cross through that physical line. Conversely, in the case of labor markets where working involves turning on an app in private, it is far harder to feel bad about not coordinating with fellow workers. The physics of reality matter.
The net of all these changes is that it really is impossible to have effective strikes and for workers to coordinate on withholding their labor collectively.
The Solution: Data-Driven ‘Guerilla’ Work Stoppages
So, if you accept my premise that it is important for workers to have the ability to threaten to not work so they can stand up for themselves, what can be done?
My view is that labor organizers to date have not thought hard enough about what modern data-driven strategies should be for coordinating meaningful work stoppages. Here, for illustrative purposes, is how workers should organize in the case of on-demand ride-hailing work platforms:
First, labor organizers should get all workers on a given platform to sign up and install a location- based worker app. This app should offer communication functionality broadly to workers, etc., but most importantly it should be able to identify the location of all drivers as they work.
Second, rather than attempt to coordinate mass strikes, the organizers should take a data-driven approach to spontaneously launching “micro-strikes” where they randomly and unexpectedly “black out” service at key venues, events and times. Rather than a general strike, the organizers should randomly take an airport like SFO “offline” every once in a while, or make pickup at a baseball game stop out of the blue, etc.
Third, and very importantly, those being asked to strike need to be compensated for lost wages. In a mass action it might be possible to get a group of people to forgo wages together. But with micro-actions where a specific worker may be asked to stop work randomly, you have to compensate them for their lost income.
This should be feasible. When you do the math, and think about driver income net of expenses, it actually shouldn’t be that expensive to black out key events and moments in different cities. It is also highly trackable from an adherence perspective. Using GPS, it is easy to know if the drivers have complied with the request to strike and if you should pay them out. Would the platforms respond with incentives to get people to drive? Yes, they will, but I believe that you could get good (if not perfect) compliance if you made people whole on their expected income. And I bet you could gamify the reward if one of the strikers randomly gets a $1,000 bonus, to encourage participation.
Workers could use a system like this to negotiate for better treatment and more compensation from the platforms by making it clear they can hurt the consumer brand and experience at if the platforms do not play ball. If I knew that Uber and Lyft might be randomly taken completely “offline” by a lightning-strike after a ballgame, I would likely start arranging other transportation options. Exacting deep and unexpected pain on consumers at specific times and places is far more effective than what might come from more broad-based labor engagements, that might make pickups a little slower, or prices a little higher broadly for a period of time.
Funding Lightning Strike Platforms
What this strategy does not address is where to raise the capital to deploy these tactics.
The best answer would be if workers contributed a small subscription fee. If one million Uber drivers contributed even $1–$2 a month to a fund, that would create a budget for dozens of strategic work stoppages that could have a real impact if focused well. This still might be a hard sell.
There is an argument that nonprofits could fund something like this. Again, since lightning strikes only cost tens of thousands of dollars, a little donor money goes a long way.
A possible answer—with moral challenges—is for hedge funds or companies to fund these kinds of labor actions. You could imagine all sorts of weird profit incentives in a world where people could locally influence the supply of on-demand workers for a given platform or task.
Finally, of course, there is the old adage that the point of having a bazooka versus a squirt gun is that you never have to use the bazooka. There is a world in which you never really need to use these lightning strike tactics. Instead, simply by building a user base and raising a war chest you encourage a more balanced dynamic between employers and workers. RIght now, it feels like workers don’t even have a squirt gun. Ideally it would cost very little to run a platform like this because the threat of its use is sufficient.
The Worker Collectives of The Future
In my mind, it is hard to imagine the resurgence of traditional looking unions, at least in the U.S. Their past is too checkered, their structures are too open to corruption and their tactics are generally too antiquated for the future of work.
What I can imagine is that each worker is a member/user of one or more worker collectives/tribes that coordinates highly targeted actions against platforms to help them negotiate well with the corporate platforms themselves.
The on-demand platforms like Uber and Lyft—which I suspect will become more common over time—will of course hate a future that looks like this. However, the nice part is that the more steps they take to avoid classifying workers as employees, the more they will have difficulty arguing that tools like this should not exist.
Ultimately, I am quite convinced that tools like this in the future will exist and help workers. The only question in my mind is who executes the work-platform lightning strike model first.