Digital Valuation: Lessons in relevance from the prototyping of a recommendation app

Abstract

This article describes the use of a prototype recommendation app to explore how users are included and/or excluded in categories of various kinds of ‘People Like You’. In the study, interviews with users of the prototype app indicate that the experience of receiving personalized recommendations isroutinely evaluated in terms of relevance, that is, as either of interest to them or as beside the point, as accurate or inaccurate, with accuracy often understood as recognition of their context(s). We build on the interviews to develop an analysis which suggests that the capacity of recommendation systems to make relevant recommendations is a function of the parallel projections – from the app on one side and users on the other – that are made as part of an interaction order. In developing this analysis, we reflect on the implications of the interaction order for the inclusion and exclusion of users in categories or kinds of people. We highlight the importance of the temporal formatting of interaction as a continuous present for the relation between belonging and belongings, and thus for the creation of a datasset (Beauvisage and Mellett 2020).