Cost effectiveness of software quality assurance

The authors address the economic aspects of quality assurance through an error cost model that determines the impact of quality assurance activities on overall development cost and on the number of residual errors. This model is based on the DOD-STD-2167A development cycle and contrasts the cost of finding errors in each development phase with the cost incurred if the error is not found until a later phase. Using data from a wide range of sources, the mode demonstrates that an early investment in error prevention and discovery through quality assurance standards, reviews, audits, and inspections removes errors at a significantly lower cost than finding and eliminating errors during later integration test phases. Furthermore, improving error discovery and correction in early development phases significantly reduces the number of errors that must be found in the later testing phases, enabling this testing to be more effective and thereby reducing the residual errors in the delivered software product. The model also provides a framework for characterizing a particular software development project and for comparing it with other projects