The Practice of Predictive Identification: Optimising for Organisational Needs

The advent of predictive policing systems demonstrates an increased interest in more novel forms of data processing for the purpose of crime control. This paper draws on interviews with police practitioners in the Netherlands and the UK to deconstruct the rationalities that are embedded within the turn to predictive identification. Debates on predictive policing have predominantly centred data in the analysis of the institutional and societal implication of prediction, linking its use to the premise of efficiency and accuracy and foregrounding issues around bias and discrimination. Yet, little is known about its actual practice. In policing, I find that studying data as practice surfaces new insights into the relationship between risk and the ways in which crime priorities are operationalised and the security mandate of the state is negotiated. Drawing on Harcourt’s (2008) observation that the desire to predict crime says more about the police than it does about a potential offender, I argue that predictive identification is not about prediction, nor about efficiency, but rather it is about optimisation. Here, datafication serves to overcome self-defined organisational challenges within the police.