Tuesday, February 16, 2010

Interactive Bubble Charts – a technical gimmick or an important reporting feature?

The market for business intelligence differentiates 5 styles of BI:
· Reporting
· Advanced Analysis & Ad hoc Reporting
· OLAP Analysis
· Scorecards & Dashboards
· Alerting & Proactive notification
While reporting, the classic discipline of BI, i.e. predefined operational reports with perfect layouts and ideal for printing, is widely used in financial services and also the ad hoc analysis, advanced analysis (including data mining) and OLAP reporting is getting more and more popular, dashboards still represent a relatively new discipline in financial services. This has partly to do with the fondness for Excel in the various departments, but is also contributed to the fact that users have not fully grasped the possibilities a dashboard can provide, which a standard report is not capable of.

However, the demand for dashboards is significantly picking up lately!
What are the reasons for this change in perception?
The idea of dashboards followed the study of decision support systems in the 1970s. With the propagation of the web in the late 1990s, the dashboards as we know them today began appearing.
In management information systems, a dashboard is an executive information system user interface that is designed to be easy to read. Dashboards may be laid out to track the flows inherent in the business processes that they monitor. Graphically, users may see the high-level processes and then drill down into lower level data. This level of detail is often buried deep within the corporate enterprise and otherwise unavailable to the senior executives. Dashboards are therefore business driven. Dashboards give a visual representation of performance measures. They give the ability to make more informed decisions based on collected business intelligence and align strategies and organizational goals (e.g. to visualize balanced scorecards).
These are all important factors and business users saw the value dashboards could add and really liked them. Nevertheless they often hesitated to implement them and stick to their two dimensional standard reports.
The financial crisis however made financial institutions rethink their current reporting strategy. Business as usual was just not good enough anymore. In the past the organizations invested heavily in risky financial instruments, always looking for the biggest return, regardless the inherent risk. Now, with the dramatic changes in the economy, financial services institutions are starting to think more strategically. They want to be cost efficient, want to streamline their business processes and are therefore willing to invest in best practices in order to be well positioned for the future.
That also includes dashboards because they come not only with the advantages mentioned above; they offer an additional important benefit: they save time over running multiple reports!

A standard report / grid can provide information in a two dimensional fashion. With OLAP reporting you have the ability to drill down by multiple dimensions but you can still see only a limit amount of information in one screen. With dashboards, this limitation is not longer a problem.
Let us assume a bank wants to get insight into their business segment performance. The key metrics they want to see over time for all their segments are:
· Operating Margin
· Return on Equity (ROE)
· Net Income
In a two-dimensional standard report we would build this as a grid, with periods in the column and the KPIs in the rows. We would then build a filter into this report so that we can select a segment. A comparison of all segments at the same time is not easily possible. We can also not easily identify trends over time, especially if we have more data (multiple segments with lots of periods) to analyze.
Of course we can improve this report by adding sums, variances, and even traffic lights to indicate trends but it is still not intuitive and takes time to consume.

Another option would be to build an OLAP report where the segment is a dimension that we can show on the report. With nesting, i.e. the visualization of multiple dimensions in the rows, we could get more information on the report. But it would be difficult for the business to digest the information and comparisons between segments over time.
Now, if we build the same scenario as an interactive bubble chart into a dashboard, we can easily incorporate all the desired functional at once.
Each bubble represents a different segment (differentiated by its color); its size outlines the net margin of the segment. The position of the bubble in the graph is determined by the x-axis (ROE in %) and the operating margin (in %) on the y-axis. This is easy to understand but it gets even better. With the interactivity of the bubble chart you can see how the size and position of each bubble is changing over time.

This resonates very well with the business users. They save a lot of time they had to spend on finding the information and can instead concentrate on their real job, analyzing the information and making informed decisions based on the findings.

Now, that the business has seen the potential and the wide range of use cases for dashboards they are asking for more. They would like to generate dashboards that incorporate all their relevant information they need to perform their daily business. With well defined dashboards that are concentrating on the information that is really relevant, the answer is simple.

MicroStrategy for example offers for this purpose a dashboard book, i.e. a set of dashboards that are contained in one file, available for distribution (via email or on a mobile devise). The business already loves it.

No comments:

Post a Comment