Article Preview
TopIntroduction
Over the past decade, business intelligence (BI) has consistently been one of the top information technology/systems (IT/IS) developments of concern to top-ranking executives (Luftman & Ben-Zvi, 2011), but BI technologies and processes have a much longer history than the recent deluge of media attention would indicate. Business Intelligence was first described as the ability to understand how facts are interrelated (Luhn, 1958), and is now variously defined as technologies, methodologies, and processes for collecting, analyzing, and interpreting data to help managers make informed decisions (Wixom & Watson, 2010). The earliest implementations of business intelligence confirming to this umbrella definition occurred in the 1980s, when computing storage and power became affordable enough to allow business users to access data directly, bypassing IT personnel, and performing data analysis in real-time, though only with historical (not real-time) data (Walker, 2009). The goal of corporate business intelligence initiatives is to wield the power of data analysis as a real-time competitive weapon, deeply integrated with business processes across organizational boundaries (Raden, 2007).
The ability of organizations to provide business intelligence services is shaped by the firm's existing IT infrastructure, itself a function of its IT architecture. A firm's IT infrastructure describes existing capabilities offered as shared services across business functions (Weill & Vitale, 2002). IT architecture is a corporate-level plan to develop IT infrastructure in support of business objectives (Ross, 2003). IT architecture determines the range of a firm's IT capabilities, simultaneously enabling and limiting its IT services. Further, the development of a firm's IT architecture is a long and difficult process, influenced by previous IT architecture plans and current IT infrastructure investment.
Most companies are still early in the process of standardizing technologies, data, and processes, and facing the limitations imposed by their infrastructure on the development of IT architectures that will enable them to provide corporate-level BI services. Integrating the multiple incompatible applications resulting from previous investment is a difficult task. Additionally, heavy investment in enterprise resource planning (ERP) applications and data on data warehousing (DW) solutions, a significant portion of which has been done at the functional, not enterprise level, limits and complicates integration efforts. This arduous task requires standardization of data formats, data meaning, and business processes in order to proceed to the highest identified architecture stage in which shared services are provided across business units. The kind of BI services that enable successful competition based on analytics (Davenport, 2006) can only be provided as enterprise-wide shared services based on a common architecture.
This paper summarizes findings from interviews with four different companies1 in industries as varied as consumer health & pharmaceuticals, financial services, insurance, and hospice care management. The interviews provide insight into the range of issues faced by organizations in the provisioning of adequate infrastructure to support business intelligence services. The interview results also shed light on the ways in which semantic data exchange technologies, most notably the eXtensible Markup Language (XML), are being used to bridge compatibility challenges and implement localized solutions to provide BI services.