New ways of organising freelance work: How algorithmic management unfolds across the tight- and loose-control platforms Uber and M-Turk

Extended Abstract – Submission Reshaping Work in the Platform Economy Conference New ways of organising freelance work: How algorithmic management unfolds across the tight- and loose-control platforms Uber and M-Turk

Mareike Möhlnann, Warwick Business School, June 2018

Algorithms are transforming work and its organisation more than technological advances have ever done before. Previous research suggests that algorithm-powered management holds great promise, through increasing efficiency, productivity and flexibility (Brynjolfsson and McAfee, 2014; Faraj et al., 2018). However, applying algorithms to digital work may result in overbearing control, manipulation and monitoring, lack of transparency due to black-box execution, and constant surveillance of workers (Faraj et al., 2018; Newell and Marabelli, 2015).

Research attention has only recently turned to “algorithmic management” (Lee et al., 2015; Möhlmann and Zalmanson, 2017), referring to a novel way of organising work in which algorithms not only mediate work processes and provide decision support, but also take over the supervision of workers previously performed by human managers. Existing research addressing algorithmic management has focused mainly on how it unfolds on platforms such as Uber. However, there is little knowledge of how different forms of platform governance create distinct work relationships, and how these affect the ways in which algorithmic management unfolds on different types of platform. Previous research has not addressed algorithmic management in contexts with loose-control platforms, such as M-Turk. In contrast to tight-control platforms such as Uber, on loose-control platforms, a heterogeneous group of contractors offering their services via the platform takes over some governance responsibilities, for example by setting work tasks and terms for freelance workers (Constantiou et al., 2017). Consequently, the workers may have opportunities to control some work variables, such as deciding which contractor they would like to work for, or which task they would like to conduct.

This study unveils how different modes of platform governance create distinct working relationships that affect how algorithmic management unfolds across tight- and loose-control platforms. We drew on two independent data sources, and (Jones and Alony, 2008; Vaast et al., 2013). Computational methods were used to harvest data from both forums, using the complete datasets of both forums from their first entries until early January 2018. We followed a grounded theory approach, allowing us to analyse data in a systematic way (Gregory et al., 2015; Glaser and Strauss, 1967)

Our results reveal distinct governance types of algorithmic management. “Platform- dominated” algorithmic management is executed on tight-control platforms such as Uber, and is designed to respond to the interests of the platform provider. “Co-governance” algorithmic management is executed on loose-control platforms such as M-Turk, and is designed to respond to the responsibilities and interests of both platform provider and contractors. While workers on tight-control platforms express very negative feelings about black-box algorithms, these feelings are moderated on loose-control platforms by freelancers’ choice over some work variables. Freelance workers are able to implement reversed features of algorithmic management on loose- control platforms. They use digital tools that run algorithm-based searches or classifications, allowing them to employ some features conceptualised as algorithmic management in reverse. They implement these tools in order to evaluate contractors’ performance, and may even block those for whom they do not like working. On tight-control platforms, since workers are more controlled and have less choice about work variables, there is little room to implement reversed features of algorithmic management.

We extend the emergent literature on algorithmic management in several ways. First, we are the first to compare how algorithmic management unfolds in different platform governance contexts, specifically on tight- and loose-control platforms (Constantiou et al., 2017). We show that different ways of organising platform governance affect work relationships and how algorithmic management materialises on different platforms. Second, we are the first to extend the existing literature by theorising how algorithmic management affects the interplay between the “triad of relationships” (Hawlitschek et al., 2016), referring to the dynamics of interactions and work relationships between the platform provider, freelance workers and clients/contractors on different platforms.

Our findings contribute to the literature addressing the organisation of work based on algorithms (e.g., Faraj et al., 2018; Horton, 2017; Newell and Marabelli, 2015), as well as literature focusing on the organisation of work on digital platforms and in digital marketplaces (e.g., Chen and Horton, 2016; Deng et al., 2016; Orlikowski and Scott, 2015).

Brynjolfsson, E., McAfee, A., 2014. The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies. W.W. Norton & Company, New York, NY. Chen, D.L., Horton, J., 2016. Research note – Are online labor markets spot markets for tasks? A field experiment on the behavioral response to wage cuts. Inform. Syst. Res. 27(2), 403– 423, Constantiou, I., Marton, A., Tuunainen, V.K., 2017. Four models of sharing economy platforms. MIS Quart. Exec. 16(4), 231–251, 798/474. Deng, X., Joshi, K.D., Galliers, R.D., 2016. The duality of empowerment and marginalization in microtask crowdsourcing: giving voice to the less powerful through value sensitive design. MIS Quart. 40(2), 279–302, Faraj, S., Pachidi, S., Sayegh, K., 2018. Working and organizing in the age of the learning algorithm. Inform. Organ. 28(1), 62–70, Glaser, B.G., Strauss, A.L., 1967. The Discovery of Grounded Theory: Strategies for Qualitative Research. Aldin, Chicago, IL. Gregory, R.W., Keil, M., Muntermann, J., Mähring, M., 2015. Paradoxes and the nature of ambidexterity in IT transformation programs. Inform. Syst. Res. 26(1), 57–80, Hawlitschek, F., Teubner, T., Adam, M.T.P., Borchers, N.S., Möhlmann, M., Weinhardt, C., 2016. Trust in the sharing economy: an experimental framework. Proceedings of the International Conference on Information Systems (ICIS 2016), Dublin, Ireland, 11–14 December, pp.1–14. Horton, J., 2017. The effects of algorithmic labor market recommendations: evidence from a field experiment. J. Labor Econ. 35(2), 345–385, Jones, M., Alony, I., 2008. Blogs: the new source of data analysis. J. Issues Informing Sci.

Inform. Tech. 5, 433–446. Lee, M.K., Kusbit, D., Metsky, E., Dabbish, L., 2015. Working with machines: The impact of algorithmic, data-driven management on human workers, in: Begole, B., Kim, J., Inkpen, K., Wood, W. (Eds.), Proceedings of the 33rd Annual ACM SIGCHI Conference. ACM Press, New York, NY, pp.1603–1612, Möhlmann, M., Zalmanson, L., 2017. Hands on the wheel: navigating algorithmic management and Uber drivers’ autonomy, in: Proceedings of the International Conference on Information Systems (ICIS 2017), Seoul, South Korea, 10–13 December. Newell, S., Marabelli M., 2015. Strategic opportunities (and challenges) of algorithmic decision- making: a call for action on the long-term societal effects of “datification”. J. Strategic Inform. Syst. 24, 3–14, Orlikowski, W.J., Scott, S.V., 2015. The algorithm and the crowd: considering the materiality of service innovation. MIS Quart. 39(1), 201–216, Vaast, E., Davidson, E.J., Mattson, T., 2013. Talking about technology: the emergence of a new actor category through new media. MIS Quart. 37(4), 1069–1092,