Establishment of automated regression testing at ABB: industrial experience report on 'avoiding the pitfalls'

Establishment of automated testing is a high risk as well as a high investment project. It is not uncommon to come across real-life stories about failed test automation attempts in the literature. In fact, it is often cited that failure rates for test automation projects are as high as - if not higher than - any other software development projects. What are the 'common pitfalls ' that all these projects fall into? Can they be avoided via well laid mitigation plans? How does the process of 'pitfall avoidance' relate to the process of 'risk management' in test automation projects? After two consecutive projects, ABB has established automated build functional regression testing of its aspect integrator platform (AIP). This paper reports on the experiences from these projects. To this end, it puts the emphasis on the importance of incorporating the 'common pitfalls' in the formal risk management throughout the automated testing establishment life cycle, and the effects of such an act on the execution and success of the projects. Further, it elaborates on the lessons learnt during practice.