Algorithmic management: Learning from Uber's woes
24 June 2019
By Mareike Möhlmann
Next time you relax into the back seat of an Uber after a night out, spare a thought for your driver.
They have spent the evening at the beck and call of an algorithm, which dictates who they pick up, what route they take and how much money they earn.
If they fail to obey the instructions imposed upon them, they risk getting bounced off the platform – and if they have a problem or want to dispute a decision, they are at the mercy of an automated phone line rather than a human being.
As organisations increasingly turn to digital platforms to deliver their services, the algorithmic management practices used by Uber are becoming more widespread. It’s not hard to see why organisations are keen to go down this route.
It allows them to respond quickly and efficiently to customer demand and to manage huge amounts of people (three million worldwide in Uber’s case) with very little management manpower.
Our research suggests, however, that not all in the automated garden is rosy. The way the ride-hailing company is managing its drivers may well be leading to significant cost savings, but it is also resulting in bad feeling and subversive behaviour among drivers, which is counter-productive and has the potential to cause real harm to the business.
My joint research with Lior Zalmanson, Ola Henfridsson, and Robert Gregory revealed that algorithmic management practices are causing tension and leading to fault lines opening up in the employer-employee relationship.
Based on a snapshot of Uber drivers in London and New York, we find, on the one hand, that self-employed Uber drivers have a degree of autonomy. They can decide when and for how long they work, giving them the flexibility to meet family commitments, juggle work or study, or even kick-start a fledgling business.
On the flip side, however, as soon as they log onto the app, drivers are effectively under surveillance, with their every move controlled and scrutinised by the platform’s algorithms and no wriggle room to diverge from instructions, even if what they are being asked to do is not in their best interests.
Lack of transparency is one of the main causes of driver discontent. The algorithms behind Uber’s platform are complex. Drivers struggle to understand how rides are allocated, how ratings are distributed and how earnings are calculated. This leads to accusations of unfairness and manipulation (Uber has previously admitted using behavioural science to nudge drivers into working longer and harder).
These feelings of dissatisfaction are compounded by the fact that drivers feel lonely, isolated and de-humanised. They have no contact with a human manager and typically don’t know other Uber drivers in their area. There are no colleagues to compare notes with, no-one to call on in times of trouble and no community to be part of.
Anyone who has come across a belligerent cabbie won’t be surprised to hear that drivers have responded to this by raging against the machine. Our research revealed that they were ‘gaming’ the system, finding clever ways to work around the algorithms that Uber uses to control them.
We found examples of drivers secretly colluding to organise mass ‘switch-offs’, for example, leading to a shortage of rides in certain areas and a subsequent price surge. Drivers were also finding ways to break free from the unpopular UberPOOL, which forces them to take multiple passengers who are heading in the same direction, even if it’s not economically beneficial.
There are lessons here for organisations who are developing digital platforms and want to avoid the kind of backlash Uber has experienced. For a start, companies can’t expect to position themselves as ‘partners’ with their employees if they persist in keeping them in the dark about the way algorithms work. Finding ways to get employees actively involved in designing algorithm-driven systems will do much to counter negative feelings and build more supportive and engaged workforces.
Adding a human element to the way people are managed will also help workers feel less like they are being treated as a machine. Uber has recognised this with the recent launch of ‘Greenlight hubs’, which offer walk-in support services for drivers. Developing formal employee communities, which give staff the chance to network and socialise, will also help to create a sense of belonging.
It’s impossible to say whether having these kind of measures in place would have helped Uber avoid its recent high-profile and on-going run-in with the regulators. At the time of writing, Uber is heading for the UK Supreme Court to challenge an employee tribunal ruling that it should give its drivers access to employment rights.
What is clear, however, is that current models of algorithmic management are tearing employers and employees apart, rather than bringing them together. More research is needed to understand how digital platforms and the algorithms that sit behind them can be redesigned to bring about a better balance and meet the needs and goals of both parties.
Möhlmann, M. and Henfridson, O. (2019): What People Hate About Being Managed by Algorithms, According to a Study of Uber Drivers, Harvard Business Review.
Möhlmann, M. and Zalmanson, L. (2017): Hands on the wheel: Navigating algorithmic management and Uber drivers' autonomy, proceedings of the International Conference on Information Systems (ICIS 2017), December 10-13, Seoul, South Korea.
Möhlmann, M. (2016): Digital Trust and Peer-to-Peer Collaborative Consumption Platforms: A Mediation Analysis, published as SSRN paper, available at http://ssrn.com/abstract=2813367.
Mareike Möhlmann is Assistant Professor of Information Systems & Management and teaches Digital Marketing Technology and Management on the suite of MSc Business courses and MSc Management of Information Systems & Digital Innovation. She also lectures on Design Thinking for Digital Innovation on the Undergraduate programme.
Follow Mareike Möhlmann on Twitter @mareike_online.
For more articles like this download Core magazine here.