The Potential of Portfolio Analysis in Guiding Software Decisions

Developing a complex software system involves decisions about how to allocate a limited resource budget among a collection of costly software alternatives (such as technologies or analysis techniques) that have uncertain future benefits. Very little quantitative guidance is currently available to make these decisions. We suggest that these allocation problems are naturally viewed in the powerful portfolio selection framework of financial investment theory. We view each software activity as an investment opportunity (or security), the benefit from the activity as the return on the investment, and the allocation problem as one of selecting the "optimal" portfolio of securities.