by Alex Rosenblat
Uber has long claimed it’s a technology company, not a transportation company, and an intermediary that connects supply (drivers) with demand (passengers). The language Uber uses communicates a strong message of distance between itself and its relationship to drivers: Uber classifies drivers as independent contractors, labels them “driver-partners”, and promotes them as entrepreneurs, although the company faces legal challenges over issues of worker misclassification. Uber relies on the politics of platforms to elude responsibility as a traditional employer, as well as regulatory regimes designed to govern traditional taxi businesses. The terminology Uber uses fosters a certain promise about the freedom of automated systems for organizing work that credits workers with a lot of autonomy and independence.
In a 9-month empirical research study I did with Luke Stark (NYU), we discovered that Uber exerts significant — albeit indirect — control over how and when drivers work through a system of semi-automated and algorithmic management. Some of these findings are discussed in this blogpost.
Plouffe’s Augmented Workers
David Plouffe, President Obama’s former 2008 campaign manager, was recruited to Uber as a Senior Vice President of Policy and Strategy in September 2014. At a Tim O’Reilly Future of Work (for business) conference in Dec. 2015, Plouffe took the stage, and thanked an Uber driver for “working on the platform” (not for Uber) shortly after announcing that drivers are, in fact, augmented workers (like the rest of us…). This language emphasizes that “humans” are not at the centerpiece of Uber’s innovation: it’s the technology that matters. This public display of technological dominance is re-iterated in Uber’s contracts with its drivers.
Uber’s legal contract with U.S.-based drivers states that Uber Technologies Inc. “provides lead generation to independent providers of rideshare or peer-to-peer (collectively, “P2P”) passenger transportation services.” In its contract with UK-based drivers (updated Oct. 20, 2015), drivers are described as customers (i.e. consumers of the app, similar to passengers).
Both frames imply that drivers are not in a traditional employment relationship with Uber, and that the platform is a neutral, software intermediary that merely connects drivers and passengers. Any negative consequences are thus construed as a feature of connectivity, rather than as a function of power relations. As our research illustrates, employment structures and hierarchies do emerge through Uber’s software platform through automated management.
Uber’s rhetoric and logic creates the context for policy developments around a new labor classification that accommodates intermediary employer-platforms in the on-demand economy, like Uber and Lyft.
Uber’s scaleable model can “enhance the efficiency of the operation of the labor market” according to Seth D. Harris (Cornell), a former U.S. Deputy Secretary of Labor & Princeton labor economist Alan Krueger (whom Uber funded for an Uber study of the labor market).
Krueger is on the academic speaker circuit to promote a new worker classification he developed with Harris out of the Brookings Institution, labelled “Independent Worker,” as opposed to “Independent Contractor.” Their report, “Modernizing Labor Laws for Twenty-First Century Work” supports the idea that platforms, like Uber’s, are mere intermediaries and highlights that these intermediaries function as such because they control very few functions. They write, for instance, “The intermediaries’ apps allow independent workers to select which customers they would like to serve. The intermediary does not assign the customer to the independent worker; rather, the independent worker chooses or declines to serve the customer (sometimes within broadly defined limits)” (p. 9).
However, as we found out in our research project, that’s not what work looks like for Uber drivers.
Autonomy or Control?
Uber is a passenger-facing company: it offers a relatively seamless way to achieve a mostly homogenous experience for getting from A to B in hundreds of cities and over 60 countries. Yet, as I detail below, this successful standardization is in tension with entrepreneurship. Drivers must submit to a system that molds their interactions, controls their behavior, sets and changes pay rates unilaterally without consulting their “partners,” and is generally structured to minimize the power of driver voices.
As Luke Stark and I observe in our research report (p. 4), Uber’s system enforces blind acceptance of passengers by drivers, who are not shown the passenger’s destination or how much they could earn on the fare. Drivers risk “deactivation” (an Uber word meaning to be suspended or removed permanently from the system) for canceling unprofitable fares, such as short, minimum fare rides.
The Uber system requires drivers to maintain a low cancellation rate, such as 5% in San Francisco (as of July 2015), and a high acceptance rate, such as 80% or 90%. Drivers absorb the risk, produced through information asymmetry, of unknown and unprofitable fares, even though Uber promotes the idea that they are entrepreneurs who are knowingly investing in such risk.
Drivers have to maintain a high star rating to remain active on the Uber platform, such as 4.6/5. They are highly encouraged to subscribe to a long list of etiquette of how they should behave in order to elicit high ratings from passengers, who are prompted to rate them at the end of each trip. This code of conduct is couched in the language of a recommendation, such as: “Riders give the best ratings to drivers who…” followed by items like “Riders prefer for drivers not to promote other businesses during the trip.” The power of the rating system is its ability to mediate Uber’s power indirectly: for example, Uber provides training videos (relevant part 6:19-6:51) to drivers suggesting that 5-star drivers provide phone charging cords or bottled water to passengers.
If drivers opt not to follow Uber’s suggested etiquette, they risk lower ratings and deactivation. In this way, the rating system automates the management process of keeping workers in line with certain performance targets and behaviors. The automation of quality control over a large, disaggregated workforce is not to be confused with independence for workers.
The Algorithmic Manager
Uber uses surge pricing to organize drivers to work at a particular place at a particular time, without guaranteeing the validity of the surge incentive if they do follow it. Surge is produced through an algorithmic assessment of supply and demand and is subject to constant dynamism. Uber sends real-time and predictive alerts to drivers about “high demand” and drivers interpret that messaging to mean “surge will happen.”
Many drivers expressed frustration and enthusiasm alike for surge pricing, because its very dynamism is characteristically fickle and opaque. Drivers noted that they would sometimes converge en masse at a surging area, and find that supply was no longer too low — the surge would disappear. Some drivers reported experimenting with trying to game these algorithms, and many developed responses to surge pricing based on their experience with its duration, reliability, and potential reward in their respective locations.
Besides, the rate that drivers are paid is based on the passenger’s location, not their own. Even when they travel to an active surge zone, they risk receiving passengers at lower or higher surge than is initially advertised, or getting fares from outside the surge zone.
Drivers are penalized if they try to assert a preference for surging fares over non-surging fares. For example, drivers will be locked out of the system for varying periods of time, like 10 minutes, 30 minutes, etc. for declining too many rides. They also get warnings of deactivation for “manipulating” surge.
Uber drivers are “free” to login or log-out to work at will, but their ability to make choices that benefit their own interests, such as accepting higher-fare passengers, is severely limited. That power dynamic reflects the divide between drivers and Uber more broadly.
Where Do We Go From Here?
The “independent worker” category provokes policy debates about the right way to address worker protections and rights in an on-demand economy. But, as Sara Horowitz of the Freelancer’s Union observes, the organization of work through technology evolves quickly and iteratively, and policy may not adequately foresee or forestall the consequences or implications of large-scale, on-demand labor models. Independent contractors, independent workers, or platform-entrepreneurs would be better served by the additional creation of representative-intermediaries who can challenge and negotiate the hierarchies or employment structures that emerge through software platforms.
Part of this blog is excerpted from our research draft.
Alex Rosenblat is a researcher and technical writer at Data & Society. Her areas of research include socio-technical systems, the intersection of technology and labor, and the social and civil rights implications of emergent technologies. She currently examines the relationship between semi-automated systems and labor management, with a particular focus on the on-demand economy. She holds an MA in Sociology from Queen’s University in Canada, and a BA in History from McGill University.