Paper short abstract:

This paper relies on an eight-month fieldwork to describe how consumer digital data assembled in data intensive financial business. By taking a pragmatic approach, we describe customer datasets as the outcome of multiple moments of qualification and valuation.

Paper long abstract:

This paper describes how consumer datasets are produced and mobilized in a data intensive finance retail startup. More concretely, the paper focus on unpacking the myriad of qualification and valuation practices and devices that take part in the assembling of a customer database.

The paper takes an STS inspired pragmatic approach were customer "digital data" is understood as a practical accomplishment involving a careful orchestration and manipulation trough expert practices and devices. This approach is empirically unfolded by presenting the outcomes of an eight-month experimental ethnography of the process of manufacturing and analyzing a transactional dataset from a low finance retail small company in Chile. During this fieldwork, we actively engaged in the process of creating, modifying and data mining a big set of customer digital transactional data. We describe two interrelated process. The first process deals mainly with the practical activity of making relations; the use of SQL queries for creating and modifying links between preexistent data that result in the creation of new entities. The second process relate to the practical operation of testing and valuing consumer data. We focus on how valuation practices and devices involves orchestrating different types of trials were the value(s) of a given entity or collective in the dataset is realized.

The paper conclude arguing that customer datasets might be understood as the outcome of multiple moments of qualification and valuation that relies on the interplay of different digital valuation devices and infrastructures.