The influence of mirroring on the creation and adoption of the Snapfix service is the focus for this sub collection of blog posts and I choose data that aids in my research question and increases the likelihood of choosing appropriate team structures within Snapfix to achieve our goals as efficiently as possible. I consider case studies describing the creation and adoption of any major SaaS service (typically Google Cloud, Microsoft Azure and AWS). My posts are supported by archived empirical data from four distinct sources.

Firstly, from the body of academic literature covering the development and adoption of SaaS and other forms of cloud computing by organizations. This is considered to be the most reliable in terms of accuracy and quality of data collected due to the academic rigour required of papers published in recognized academic journals (Willcocks, Whitley, & Avgerou, 2008¹). The papers are drawn from the major Information Systems, Organizational Science and Computer Science academic journals and conference papers with a small number sourced from other academic traditions. I examine title and abstracts to eliminate duplicates, white papers, editor’s commentaries, purely theoretical work, experiments (with one exception) and irrelevant work. A further review of abstracts eliminated papers that hold little or no descriptions of organizational or technical detail. Finally, a snowball search resulted in the inclusion of 6 additional papers. Out of an initial body of 163 papers, I include 44 academic papers in my analysis, including conference papers, that deal with cloud computing and organizational structure. These include 27 papers that discuss organizational adoption of cloud computing and 17 papers that discuss development of cloud computing. The majority of papers are case studies, but I include a limited number of systematic literature reviews (4) that I consider relevant and valuable to these posts. The volume of case studies is based on reaching saturation. That is, I stopped searching for original sources of data when I noticed repetition within the materials being gathered. There are more papers dealing with adoption compared to creation of cloud services due to the more fragmented spread of relevant data on the former.

Secondly, from other blog sources, industry articles, podcasts and conference talks which are more numerous and provide factual reports on decision making, organizational change and issues encountered during development and adoption of cloud services. This source of data is from leading software practitioners who are publicly recognized and influential within the software industry (Rainer & Williams, 2018²). Examples include articles from Martin Fowler, CSO of ThoughtWorks and Werner Vogels, CTO of Amazon. These industry sources contain accounts of software development of cloud services (8) and consumption of these services by different organizations and more general discussions on software architecture (4) from leading industry thinkers and practitioners. Non verifiable accounts of cloud development or adoption are rejected.

Thirdly, from official technical repositories of the leading cloud providers, to confirm technical details or to verify the veracity of other sources. These include Amazon Web Services (7) and Google Cloud Platform (1).

Fourthly, following an analysis of the majority of the above mentioned secondary sources, I collect additional primary data by conducting interviews (2) with industry leaders immersed in cloud computing to expand upon and clarify specific aspects of cloud computing that are not fully expressed in the secondary empirical data sources. In one such case, I interview my colleague Paul McCarthy who is the CEO of Snapfix. Paul describes the adoption of Snapfix by our customers.

Guided by Eaton et al., (2015)³ methodology, I had originally intended to limit my research to secondary sources of empirical data. However, I felt that a limited amount of field interviews of senior industry practitioners and my accounts of industry experience in designing and developing global scale cloud services would enhance this series of posts and allow me to fill in some specific gaps encountered while analysing the main body of secondary empirical data.

Next: Data Analysis