Create a System, Not Just an Interface
Once the business architecture is defined, you can build the infrastructure to display and manage the data that will populate the dashboards and scorecards. Keep in mind that a performance dashboard is a full-fledged business information system. It consists of a layered set of analytic applications that run on a common set of BI services, including security, metadata repository, query engines and reporting engines. As shown in the "Architecture of a Performance Dashboard" diagram (at right), this service-oriented BI architecture in turn runs on a data management infrastructure that integrates and delivers information and insights to users on demand.
Starting with the crucial top layer for information delivery, a performance dashboard uses a set of layered applications built on common BI services. Data infrastructure supports the dashboard by integrating disparate information in a timely manner.
Ideally, a performance dashboard takes advantage of investments you've made in information architecture, data warehousing, data integration, reporting, analysis and planning applications. The dashboard is how you are creating tangible business value out of this infrastructure. Alas, too often not all performance dashboards have the infrastructure to deliver real insights and long-term value. Products with glitzy graphical interfaces have tantalized many organizations, leading them to implement pseudo performance dashboards that are:
» Too Flat. A quick and popular way to create performance dashboards is to use Microsoft Excel, PowerPoint and more advanced charting packages. These applications often look fancy, but they don't provide enough data or analytical capabilities to allow users to explore the root causes, even if they're highlighted by fancy graphical indicators.
» Too Manual. Some firms rely too heavily on manual methods to update performance dashboards that contain lots of information. Skilled business analysts may have to spend several days a week collecting and massaging information and putting it in a dashboard or scorecard. Sometimes manual data collection is valid when the data doesn't exist in any system, when there isn't much data to load or when the dashboard fills a temporary business requirement or serves as a prototype for a future application. You want to free analysts of mundane tasks.
» Too Isolated. Some dashboards or scorecards contain data and metrics that don't align with the rest of the organization, leading to confusion. You will encounter managers that want to use the latest dashboard technology simply to automate their Excel reports. Don't deploy dashboards and scorecards in a vacuum, or you'll just create problems down the road.
Do You Build or Buy?
Two years ago, there weren't many commercially available dashboard or scorecard products worth purchasing. Today, however, established vendors and a slew of start-ups are battling in the marketplace with performance-dashboard products. Many BI vendors provide dashboard tools that display charts in a portal-like framework and scorecard tools that map and display strategic objectives and metrics. You also will encounter vendors that specialize in discrete areas of visualization, process management, performance management, Balanced Scorecards and portals. Providers of ERP applications also are active, touting their greater experience with the application data sources. They either resell or offer their own dashboards and scorecards.
Leading, lagging, and diagnostic metrics from the linchpinthat connects a performance dashboard's business and technical architectures. Executive, employee and other stakeholder requirements as well as overall strategic objectives determine which elements of the two architectures must be aligned by the metrics linchpin.
Look for products that will help you clearly and concisely communicate key strategies and goals to all employees. Performance dashboards and scorecards are part of a layered set of analytical applications running on a common set of BI services; they let users measure, monitor and manage the processes and activities for which they are accountable. Don't forget that the system is dependent on a robust data management architecture that delivers actionable information to users on demand.
Dashboards and scorecards should translate the organization's strategy into objectives, metrics, initiatives and tasks customized to each group and individual in the organization. Don't settle for pretenders. Done right, dashboards and scorecards fulfill BI's ultimate potential: to become the critical tool for strategic decision making.
Wayne W. Eckerson is the director of Research and Services for the Data Warehousing Institute and the author of Performance Dashboards: Measuring, Monitoring, and Managing Your Business (John Wiley & Sons, 2005). Write to him at [email protected].