Risk-Based Test Estimation

Test Estimation is a problematic area. Two main approaches were followed traditionally to estimate testing effort. Development Ratio method with testing effort as a fixed percentage of the development effort, and independent estimation techniques of the test requirements or WBS. Both approaches often apply adjustment factors to the resulting effort to reflect different technical and environmental parameters in order to increase estimation accuracy. However the old approaches share a common pitfall; the absence of direct correlation between the estimated effort and the criticality of the system under test in its operational context (risk of failure). Critical or sensitive test objects require more attention than the less critical or the more stable ones. The effort spent by testers on every test object got to be function of the risk the end users might incur should this object fail. This paper is an industry report based on practical experience that aims at combining the benefits of the traditional approaches with consideration of product risks. It provides a method that draws a product risk profile and devises a risk-based effort calculation formula and metrics scheme where effort is proportionate to the risk level of each object under test.