Dashboards, whether in your car or part of an application suite, are meant to provide users with quick visibility to vital stats and metrics so you can make smart decisions.
That said, not all dashboards are created equal. In cars and in front of the computer screen you can often hear people saying “what am I looking at.”
Cluttered BI dashboards happen when BI teams just get to work building them without carefully considering what information they should deliver to the user and in what form. Often, the result for the user is information overload.
According to analytics services firm SearchBusinessAnalytics.com, their numerous planning interviews with clients indicate that executives typically look for summarized data on their BI dashboards. Users are fine with capsulized version of the information as long as they are able to drill down further when they refer more detailed information.
Here are some tips on how to create a BI dashboard that can effectively deliver relevant, timely and easy to understand metrics:
Limit a dashboard page into no more than four to six metrics. Put metrics for each subject area or business unit on a separate page or tab. Executive dashboards generally provide a global view but have multiple pages and tabs.
Don’t drill down to deep. Three levels are usually enough. Beyond that, if executives have more questions they can call in the data analyst or someone from the Bit or IT team.
Use charts and bar graphs, bullet list and “sparkline” graphs for visual aids. Pie charts are not always advisable because the human eye can only effectively compare slices that are right next to each other. Gauges are effective in getting the attention of the user, but present little information and take up too much screen space.
Dashboards should contain filters so that users can further customize the data they are viewing, by narrowing or expanding the scope of information being presented. Consider adding an alert feature which calls the attention of the user when a pre-defined criterion is reached.
Be prepared to alter the dashboard you present. As users begin to work on the dashboard they will likely realize they need visibility on other factors or don’t need data on a certain subject or demands of the business have changed.
BI dashboard building may be an ongoing project but good design principles will generally stay the same.
Sponsor: IBM Canada Ltd
Increasing Profitability with Analytics in Midsize Companies
Complete with several mini-case studies, this Nucleus Research Note dispels the misperceptions surrounding Analytics, revealing how small and midsize companies have cost-effectively implemented and deployed Analytics solutions and gained benefits.