Rock solid test cases start with usage statistics

Real test cases with better coverage & lower test efforts.


“Does anyone else need this or can it be dumped?”

This far reaching question arose while relocating.


The whole idea of “HeatMap” – an illustration of data usage in SAP BW – originated from Andreas Förger at Randstad. This solution component was realized during a meeting with Gartner’s Thomas Oestereich in our Heidelberg offices with Datavardians Branislav Cik and Martin Zlaty tweaking details. Before I begin, I’d like to extend a big thank you to the collaborators of this team effort for sharing your ideas which resulted in the creation of an exceptional solution.


Originally the idea was to develop usage statistics as a basis for effective data management in SAP BW to automatically classify based on real user behavior data.  As a result the system itself answers the question: “Does anyone else need this or can it be dumped?” A closer look reveals that for this meta-intelligence other important case applications apply in particular SAP security and test management, respectively test automation.



What are the most used and complicated processes in my system? To illustrate this realistically lets take our ten highest order volume clients. Once identified, there are two questions:

1.)  Are there test cases where I use this data?

2.)  If yes, what is this customer’s most complex operation? Has this been tested? Even a customer request with 126 positions? If yes, congratulations. You’re part of a 5% elite according to a current survey ( of 215 German SAP users for SAP Test Management.


Let’s go one step further:

We’ll pick three key users from purchasing. HeatMap helps me answer the following important questions:

1.)  How do they use this system? Did usage alter from initial planning?

2.)  Which are the favorite reports and transactions?

3.)  Does usage still reflect current test cases in terms of complexity and volume?

4.)  Dead weight: Are there functions not being used at all? This pertains particularly to customized applications now standardized.



Heatmap provides important indicators: average run time of top 50 report / queries in the system. How does this vary depending on time of day, location or user selection? Is acceleration possible with aggregates or indexes?


These examples enable us to lay ground for effective usage of data and programs. Possibilities range from architectural questions to operational aspects such as  testing or archiving.

Leave a Reply

Your email address will not be published. Required fields are marked *