Applying KM lessons learned to business analytics
Traditional BI solutions have focused primarily on delivering information to decision makers whether at the executive or staff levels. Although there has been much progress in the speed, accuracy and presentation methods of delivering information to users, there has been little progress in extending true decision support functionality to the broader organizations. Functions such as workflow management, collaboration, what-if analysis, scenario planning and predictive analytics are only now beginning to find their way into mainstream BI solutions. Much work remains to be done by corporate IT departments, systems integrators and IT vendors in moving the business analytics market forward. All relevant parties should begin to incorporate some of the lessons learned from KM projects into business analytics projects.
Organizations operating in today's global economy are faced with unprecedented competitive and regulatory pressures and a heightened level of uncertainty. Driven by geopolitical trends and financial scandals, national and international regulatory agencies have enacted rules with far-reaching impact on the daily operations of organizations in all industries.
By most accounts, the uncertainty is more profound than in past decades, and is likely elevated by the amount of data and information available to today's decision makers. (IDC research indicates that over a quarter of large organizations expect their data warehouses to at least double over the next three years.) In that market environment, there is a trend toward continuously shrinking decision cycles where improved—meaning faster, more accurate, insightful and flexible—decision making commands a premium and serves as a basis for competitive advantage.
Yet, significant shortcomings persist in the ability of organizations to address the decision-making needs of its employees. For example, only 14 percent of managers in a recent IDC study were very confident with the statement that the reports developed in their organizations deliver relevant information to the right people at the right time. The current shortcomings in the quality of decision making and information delivery range from an inability of organizations to capture the right data and deliver it to the right people at the right time, to poor data quality, system complexity and the disconnect of business analytics from operational systems. Those shortcomings are due to:
- the lack of appropriate investment in business analytics solutions and
- the disconnect between the information delivery and decision support functionality of most of those solutions.
Although evidence of the benefits of business analytics increases every year, investment in the technology and specifically the software to support decision-making processes continues to lag behind money spent on software to process transactions. Business analytics software has been shown to provide significant returns. The "Financial Impact of Business Analytics," a study published in 2003 by IDC, showed that the average five-year ROI for business analytics projects was 430 percent and the average payback period was 1.6 years. However, if we look at today's IT spending levels (see chart, P.17 KMWorld January 2006), we see that for every dollar spent in 2004 on transaction processing applications or capturing and getting data into databases, only slightly more than 27 cents was spent on getting the data out for business analytics to support decision making and statutory reporting processes.
Most organizations claim that their data is an asset; many have built data warehouses to collect and store data. However, in some cases, the more apt metaphor should be data landfills. Many organizations have become efficient at capturing data, but much less capable of organizing, analyzing, extracting and delivering it from those data stores to enhance the overall decision-making quality. If data is indeed an asset, market research suggests that a large amount of it remains dormant and is not leveraged to its full potential.
Decision-centric business intelligence
Business intelligence (BI) software has in the past focused primarily on data analysis based on historical trends and batch data capture at certain intervals. Most traditional BI solutions focus only on information delivery, addressing the speed and accuracy of decision making. Traditional BI leaves off at the information delivery stage, paying insufficient attention to forming a problem statement, searching for candidate solutions (hypothesizing) and evaluating the likely outcomes (modeling). The danger is that a decision maker without information relevant to the decision at hand is likely to rely exclusively on intuition, a notoriously unreliable practice. To alleviate that shortcoming, IDC has stressed the importance of focusing on decision-centric business intelligence (DCBI) that extends traditional business intelligence in the following ways:
- adds collaborative support on top of access to information by individuals;
- adds support for decision-making processes, especially for repeatable, operational decisions;
- filters, monitors and delivers information based on relevance, enhancing the insight of decision makers; and
- employs advanced analytics for decision optimization.
The issue here is not only in complementing query and reporting with data mining or statistical tools. From a technology perspective, such advanced analytics are often employed in assessing and prioritizing alternative courses of action within the hypothesize and model steps of a given decision-making process. However, advanced analytics are not sufficient. DCBI adds the following requirements over traditional BI and over advanced analytics:
- Decision process automation. DCBI must be able to capture the decision-making processes of the best performers. The goal is to improve the overall quality and consistency of decision making.