Infosys Microsoft Alliance and Solutions blog

« Calling Blueprint Command from .mht file | Main | Impersonation and Delegation - There is too much confusion! »

BI Dashboards: Best Practices and Design Ideas

Based on multiple dashboard design projects  I have worked on in past, here are some quick points about deploying  ‘dashboard’ like displays for tracking business performance indices (popularly called as ‘Key performance Index’ or KPIs).

As score-carding and performance management initiatives get traction in 2009 through implementations featuring Microsoft Performance Point Server (PPS), the use of scorecard like displays is expected to intensify further.

1.       ‘Dashboard’s have become very popular with business managers as they
         Help gain quick insights into data and trends
         Monitor performance indices and track leading as well as lagging indicators
         Convert the business data into a ‘picture’ to easily communicate the story.
         But dashboards mean different things to different people

2.       Dashboards  v/s Reports
       Dashboards, by the analogy they refer to, should incorporate ‘information display’ and ‘controls to manipulate the information display’.
       Quite often, BI Dashboards are just reports with value added features to filter data, track trends or spot performance deviations.
       That is ok, but it falls short of true potential of Dashboards as a close-loop management tool.

3.       Why Dashboards?  For business goals.
           Business users should have some key, concrete business goals driven by business context.
           Some common goals that drive the development of BI Dashboards are
          One view of the (business) world
          Management by exception
          Corporate Performance Management
          Connecting Performance Data (historical data) with Forecasting Data (projections)
           From Data to Decisions to Actions

4.       What to Show on Dashboards? 
       As few things as you can and need… …And more leading than lagging data!
       Dashboards are not reports. If you show more information than necessary, you are hurting your own efficiency of using this tool.
       Numbers by themselves have no is when they are presented in corporative context ( this quarter v/s last quarter) that the true picture emerges
       So, Edward Tufte likes to stress, help them compare!!

5.       What to Show on Dashboards? 
         Positive and negative performance deviations from plan
         Fastest rising KPIs that were previously in negative territory
         Fastest falling KPIs that were previously in positive territory
        Alerts about things that are out of tune
        Comments and notes about alerts and other deviation indicators to give extra context
        Pace and size of the business trends impacting the business
        Contributing factors behind trends

6.       How to Leverage Dashboards? 
        Connect forecasting data to historical data. Allow contextual simulations and ‘what-if’ scenarios to allow informed problem solving
        Allow saving of scenarios and sharing of scenarios to foster collaboration around Dashboards
        Allow navigation along business organization entities to get low level insights
       Add ‘controls’, not just the ‘displays’ - thereby allowing the user to modify date ranges and switch perspectives.

7.       How to Develop Dashboards? 
        Document and refine usage scenarios
        List and prioritize business goals
        Understand and model the relationship between historical and forecasting data
        Understand and factor in inter-relationship between corporate communication practices and this new business decision tool
        Understand the linkages between data tools, decision analysis tools and operational action tools. Feed key performance indicators that really matter - to get a pulse of the business.


I liked the KPI wheel from BrightPoint Consulting - it helps you gather your KPI and metric requirements for dashboard projects. Take a look at it at

What do you mean by "Connect forecasting data to historical data."? you mean physically connect the data set by lines or showing them at the same time in separate lines so that users can see the relationship between two?

I was planning on just using a lot of the pre-configured images to handle the majority of the heavy lifting as I have this linked to an MSDN account. I think I’d still need to upload a Sharepoint ISO, but not 100% sure about that. Also not sure if I should try to configure this as a single server or use the more correct configuration of multiple servers. Seeing that you have this step-by-step for a single server, I’m leaning that way.

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.

Subscribe to this blog's feed

Follow us on

Blogger Profiles

Infosys on Twitter