Five Steps to Optimizing BI and Data Warehouse Performance

The pressure is on business intelligence and data warehousing professionals to handle ever-higher data volumes and ever-more-complex queries while reducing decision latency. Follow this five-step approach to identify key business drivers, optimize system performance, guide new technology deployments, improve responsiveness, and invest for future performance demands.
4. Reduce the time it takes to remedy unsatisfactory performance and implement information change requests. Make it a goal for your organization to improve your users' experience by addressing problems and implementing information change requests rapidly. The largest percentage (44%) of participants said it takes their organization about the same amount of time to remedy unsatisfactory performance as it did 12 months ago, and 5 percent said they are getting slower at doing this. Organizations at the two highest levels of maturity are reducing their response time; those at the Strategic level are faster (38%) and those at the top Innovative level are much faster (13%) than a year ago.

The majority of organizations are able to respond to unsatisfactory performance within hours (38%), but for 26 percent it takes days and for 15 percent weeks. In addition, most organizations (39%) take one to five days to implement information change requests, and 37 percent take six to 30 days. Slow implementation of these requests is not just a performance problem; it can become a data quality problem when users share wrong or outdated information through reports and applications.

A critical factor in improving response time is to know when performance demands will be at their highest so you can plan resources accordingly. Nearly three-quarters (73%) of participants said it is important to identify periods of peak performance demand; 59 percent said that peak periods come at the end of the month, and 42 percent said they occur at least weekly. Use information-based insight to anticipate spikes in performance so you can both address problems rapidly and analyze trends to guide deployment of new technology resources.

5. Assess your organization's maturity and invest for improvement. Ventana Research measures organizational maturity using a four-level scale. In this benchmark research, most participating organizations rank at the middle levels, with 38 percent at the third-highest Strategic level and 29 percent one step down at the Advanced level. Advanced organizations are reactive in their approach to remedying performance and are less confident than those at higher levels in the quality of their information sources for performance optimization (37% said they are fair). Strategic organizations are active in their approach to managing and tuning BI and data warehouse systems; they can monitor and tune some systems but have more limited insight into future needs than organizations at the Innovative level (17%).

Strategic and Innovative organizations are taking the most steps to collaborate with business users to analyze performance and determine what actions to take. At the Strategic level we find the largest percentage of organizations that have deployed BI or data warehouse appliances (38%). Innovative organizations have the most information sources for performance optimization, are the most confident in their data quality, and are leveraging them with workload analysis and tuning tools to gain insight. Use these results to assess your own maturity and to determine where you can apply improvements in terms of your people, processes, information and technology.

David Stodder is vice president and research director - information management and IT performance management at Ventana Research. He was previously editor-in-chief and editorial director of Intelligent Enterprise.

The complete "Optimizing BI and Data Warehouse Performance" report, which is available for purchase from Ventana Research, offers extensive detail on its benchmark research, including 15 charts and graphs, on the demographics and people, process, information and technology maturity of survey participants.

Editor's Choice
Sara Peters, Editor-in-Chief, InformationWeek / Network Computing
John Edwards, Technology Journalist & Author
John Edwards, Technology Journalist & Author
James M. Connolly, Contributing Editor and Writer