Beyond algorithmic management: labour control in platformbased food

Beyond algorithmic management: labour control in platformbased food- delivery Benjamin Herr, MA (university assistant) benjamin.herr@univie.ac.at University of Vienna

Domain: sociology & humanities

Hitherto research on location-based platform work (Schmidt 2017) focused on working conditions (Stewart/Stanford 2017), the outsourcing of risks and costs (Srnicek 2016; earlier: Frade/Darmon 2005), the labor market at specific platforms (Hall/Kruger 2018) or ratings as a form of control (BEIS 2018; Nosko/Tadelis 2015; Rosenblat et al. 2017).

Research on platformbased food-delivery highlights the important role of algorithmic management (Lee et al. 2015) in these work settings (for example Cant 2018). Apart from that, findings related to platformbased food-delivery are scarce, and to a large extend anecdotal (CIPD 2017).

Food-delivery platforms consciously influence the work organisation, especially fleet operation and dispatching (McKinsey & Company 2016; de Stefano 2016). A core managerial problem is the ‚indeterminacy of labour‘ (Smith 2006), i.e. to ensure an exploitable result and performance out of the purchase of labor time. The early labor process debate provides a rich set of analytical tools to engage with this aspect of working life (Braverman 1974; Brighton Labour Process Group 1977; Friedman 1977; Edwards 1979). The overall research interest is informed by this stream of thought. It aims to get better insights in the strategies management applies to control labour in platformbased food-delivery. Following research questions are guiding the text:

• What role plays technical control in the work organisation of platformbased food delivery?

•What impact have direct forms of control in the respective labor process?

•What is the relationship between data-driven forms of work organisation and forms of control?

In order to provide empirically funded answers to these questions a case-study design was applied. The observed company is a leading food-delivery platform that operates in various European cities. The author worked for four months in one of their branches as a food-deliverer. Based on the analysis of fieldnotes and lived experiences of the labour process a guideline was constructed. This guideline set the basis for subsequent semi-structured interviews with 16 salaried workers and freelancers of the branch. In addition, a second participatory observation was conducted at a traditional bike messenger company (which took 7 months).

The findings suggest that automation plays a decisive role in the organisation of the labour process. To break down the aspects of automation in this work setting: Dispatching processes in the respective case have been increasingly digitized. The algorithm sets space and time of the order. Instructions are standardised via automated phone calls. Parts of the performance - the pick-up and the drop-off - are standardised as well: workers have to confirm each step through a protocol on their app. Through the GPS-tracking workers constantly produce performance data during their shift. This data results in so-called performance reviews.

However, these facets can hardly be interpreted as a means of control, at least in the observed case. In order to do so, these predominantly data-driven forms of work organisation have to be embedded into particular social relations.

This is what the second section of the findings points towards: Performance data has to be actively and personally communicated through supervisors. These supervisors themselves are embedded into a system of competition via international team rankings. Navigating through the city is too complex as that it could be compulsory standardised. The upholding of a start-up feeling through various kinds of narratives is actively created by the management in order to control feelings, emotions and belonging. The messenger mythos was used as a discursive fragment to successfully silence demands. Measures of the performance review are the materialisation of the contestation between capital and labour on this respective shop-floor level.

Altogether, the findings suggest a view on platformbased food-delivery that goes beyond notions of a data-driven, automated and algorithmic controlled workplace. It calls for a more nuanced understanding on how living labor is governed in these new business models. By identifying the social interfaces of automated forms of control it opens paths for future research on the intersection of labor research, activism and social innovation. This presentation contributes to the conference in two ways: firstly by challenging the notion of technical control in this context and highlighting concepts that are already in place, it points to the continuities concerning labour control in this sector. Secondly, by comparing the form of work organisation in platform based food-delivery with the nature of work in traditional bike messenger services it contributes to the comparison of similarities and differences between two fields in the bike delivery sector and thereby points to the new facets of control in this particular work setting.