You’ve taken on a major BI effort, worked with business users to develop requirements and business rules, and built your first set of code. If you are not careful, you are about to enter the bottomless pit we call system testing. Do you know the warning signs? Test cycle after test cycle inches the project along. There are no signs of code validation on the horizon, just more endless…testing. Almost every BI development project that struggles can point to System Testing as the time when it was apparent that the project was in trouble. As QA, domain experts, designers and developers struggle to complete the testing, the frustrated project managers and sponsors keeping asking: “Why can’t we get this done?”
There are methods that can make your system testing a formality rather than a flop. What steps can we take to remove the uncertainties from system testing? How can system testing (and delivery) happen as planned? How can we rethink our test cycles to design and build with system testing in mind?
There are many showstoppers unrelated to testing that can land a BI project in big trouble. Let’s consider some of the biggest factors: incorrectly defined requirements, an under-performing environment, and poor design practices. Let’s assume your project has solid requirements and business rules supported by acceptable source data quality. Let’s also assume the environment is performing well, and that you have employed effective design practices well-suited to the nature of the data flow. These are landmines that have ground many projects to a halt. Not yours.
The next six parts of this series will discuss the methods to mitigate system testing risk and lay a firm foundation for iterative system testing:
Part 2: Build a functional “Canary” test data set
Part 3: Build large test data sets
Part 4: Build initial and incremental data sets
Part 5: The Case for Thorough System Testing
Part 6: Design testing to be evaluated automatically
Part 7: Pass functional canary testing before moving code to the test platform
These methods will enhance the other critical factors for BI project success: requirements, environment, and design practices. These six methods will help you stay out of the bottomless pit.
My next blogs will detail each of these methods.