Organizations large and small went about the work of business intelligence in sometimes novel and always changing ways in 2004. The top five trends for 2005 and beyond illustrate BI's increasingly central role in business, and point to both the advantages and difficulties the technology will present next year and, in some cases, for several years to come.
- Tackling the data quality challenge: Good data doesn't ensure good analysis by itself, but one thing's for sure -- bad data always yields bad analysis. BI professionals spent a hefty portion of 2004 struggling to define which data sets are most important and strip redundancy and poor information from their BI processes. Those are monumental challenges that by definition require input from staff both within and from outside IT, and the data quality issue still hasn't been resolved at most organizations. The coming year will likely see as many data quality problems as 2004 did.
- Consolidating on fewer BI apps: Full consolidation onto a single vendor's business intelligence tools is still almost unheard-of among big enterprises, but organizations of all sizes did try to reduce their BI vendor count this year. Business intelligence has evolved from a project-by-project practice into an enterprise-wide strategy, prodding companies to bring cohesion to the BI tools they use from department to department. Sometimes that means cutting down to two or three software packages from seven, eight or more. Next year will see further consolidation by enterprises that want to reduce vendor count further and by those just joining the consolidation trend.
- Analyzing unstructured data: Once a practice centered around structured data almost exclusively, business intelligence increasingly involves analyzing and reporting on structured or semi-structured data. Private and public sector IT professionals are pioneering the confluence of BI and content management now, and they'll make more progress on that front in 2005. E-mails, field reports, service records, other text-based information and paper documents are all finding their way into the analytical soup.
- Making analysis available to a broader range of users: The dawn of "operational" BI came earlier, but it grew more prominent in 2004 and shows no sign of diminishing as we head into the new year. BI is still practiced by CEOs through executive dashboards, but reports and analytical tools have now also filtered down to "line-of-business" workers and process managers who need quick information access to do their jobs effectively. In some cases that means only publishing pre-arranged reports to operational staff, but other organizations are handing full ad hoc query and reporting capability to such workers.
- Expanding the streams of analyzed data: Traditional operational and transactional data streams still form the dominant business intelligence information streams, and will continue to do so. But new kinds of operational data are entering BI and will expand their presence in the coming year. Radio frequency identification, a technology that quickly tracks the movement of everything from shipping palettes to beef cattle, will become a major BI data source in coming years. A handful of large organizations are already readying data warehouses to handle RFID data. Another relatively new technology, the global positioning system, has made "locational" analytics possible at some enterprises, giving insight into the movement of items and even people between locations.