The study, conducted by the National Institute of Standards and Technology, also found that while not all errors can be removed, more than a third of these costs – an estimated $22.2bn – could be eliminated by an improved testing infrastructure that provides earlier and more effective identification and removal of software defects.
The study examined the software activities in several major industries, but focused on two groups: automotive and aerospace equipment manufacturers, and financial services providers and related electronic communications equipment manufacturers.
In the automotive and aerospace study, Nist found that about 60% had experienced ‘significant software errors’ in the previous year. The total cost from inadequate software testing was estimated to be $1.8bn.
In the financial services group losses caused by buggy software were estimated at $3.3bn.
The study also predicted that the worldwide market for software testing tools, which was $931m in 1999, would grow to more than $2.6bn by 2004.
Testing tools are still fairly primitive, the study found, and the lack of quality metrics leads most companies simply to count the number of defects that emerge when testing occurs, without any form of grading.
Nist funded the study, which was conducted by the Research Triangle Institute in North Carolina. The full report can be found here.