Predictive Test Planning to Improve System Quality

Penny McVay shares how her team approached improving the quality of a large policy-writing application for a global insurance carrier. The application has many pieces and parts, thousands of lines of code are changed monthly, and the business depends on a stable application. To mitigate risk, the team investigated the question of how to predict where testing needed to focus. The regression test suite the organization had built was difficult to maintain, identified few defects, and took hours of effort to run. QA needed not only to optimize the regression suite but also to determine how to mitigate the most risk. To accomplish this, the team collected historical data to predict significant impact areas including what is being written in production, what transactions are being exercised in production, and what defects are occurring in today’s production environment. From this information, the QA team rewrote the regression suite to eliminate waste and provide more value to the business customers.