Regression Testing is a very critical aspect of IT Operations in any Enterprise. For a Telecom Operator, this is vital because of the number of changes at breakneck speed that are needed on the IT Systems due to the volatility of the business. Generally, Regression Suits are designed to include the most vital tests on the Enterprise so that maximum business risk could be minimised due to changes in the systems in the least possible time. Conducting a fully comprehensive Regression Test would mean a lot of resources and a lot of time. This could be substantial even if the best automated systems are used for conducting the tests. A prudent mechanism is to find the most valuable transactions and tests these transactions during the Regression Testing. Typically, 20% of the most valuable transactions if tested should eliminate at least 80% risk of failure of the changes in the solution stack.
Generally, I have observed that a well designed Regression Test Suit tends to static as the value of the transactions do not change so dynamically. Due to this static nature, the Regression Test Suit fails to deliver adequate value as it misses out on the dynamic nature of the Business based on the Products in the Market. The below formula covers this dimension.
However, the main issue which could be resolved by the below formula is impact on the solution stack generated through simultaneous launch of products across segments of business. Generally, product structure across business segment are very different. However, when these diverse product structure are serviced by the same IT Solution Stack, there are bound to be impact of changes on products in one segment on the product in other segments. Even with a very expert Development Unit, such impacts can be very difficult to find during the design and development and functional testing time. The Regression Testing the best phase for determining these deficiencies. Thus, the Regression Test Suit being sensitive to this aspect of the Test Requirement is very critical.