A Socio-Technical Account of an Internet-Based total I conducted a total of 55 formal interviews during this period (Table A1). All 55 interviews were tape recorded and extensive research notes were taken. This practice ensured that everything said was preserved for analysis. Data Analysis Because this study was part of a larger research project, the data was systematically coded into as many theoretical themes and categories as possible. As the categories emerged and were refined, I began to evaluate how they related to one another and what the theoretical implications were. I used the conceptual elements of ANT to guide this process. I also used version 5 of ATLAS. ti to code and store these themes and categories at the textual and conceptual level (See Figure 2). ATLAS.ti was also used for the overall man- agement of the research project and its associated chronologically as the research process unfolded, and filed for easy access and retrieval. Using ATLAS.ti for easy cross-referencing assisted in maintaining a chain of evidence to support the case study conclusions (Muhr & Friese, 2004; Darke, Shanks, & Broadbent, 1998). These inter- views together with the large collection of rich, thick qualitative information from a number of sources played an important role in addressing the complexity of the context studied. This trian- gulation of data was important in counteracting any biases in the collection and analysis of data (Darke, Shanks, & Broadbent, 1998). THE HEALTHCARE INSURER CASE STUDY During the latter part of the 90s, a major multina- tional South African consumer-driven healthcare