Susan Leigh Star (1991) captures a central concern of much politically sensitive academic practice: how to make that, which is rendered invisible, visible. There is a politics to that which is unseen, in which forms of what she calls “social control” become tied to the erasure of the practical activities–the forms of work–that go into making and stabilising the domains of the visible. 1 We follow the spirit of Star’s call for a sociology of the invisible, but our focus is not on human labour, but the labour of machines: the automated, unseen, digital work undertaken by ‘trackers’(other terms include ‘bugs’,‘pixels’,‘tags’). These are online data gathering tools, many provided by third party providers, activated when a user visits a particular website. They form part of what has been called the ‘invisible web’. Many of us are familiar with the way in which cookies are placed in our browsers. Cookies, however, are only one of a range of tracking devices that are deployed by tracking and advertising companies. When they are activated, content provided by a third party is loaded onto the website in question (even if it remains unseen; in some cases, this content can even import new trackers, so called ‘4th party content’). In the case we will consider in this chapter, the data gathered by these devices are being turned towards a quite specific form of algorithmic calculation (Gerlitz and Helmond, 2013; Helmond, 2013).