As the evidence sections show, there is data to support the assertions that the waterfall is more failure-prone for software-intensive projects and that disciplined IID with evolutionary requirements and design is correlated with lower rates of failure and defects, and higher rates of productivity.
Thus, a business case can be made based on several factors, including:
This analysis draws special attention to reduced failure rates. Projects fail for many reasons, but evidence [e.g., Thomas01] has shown that waterfall practices rather than IID were associated with the most significant failure factors.
Figure 6.7 shows research on failure rates [Standish00] across 35,000 projects. On average, 23% of projects failed and were cancelled before completion in 2000.
Figure 6.7. success rates
In addition, the average cost of these projects is:
Taking the average project cost for medium-sized companies, $1.1 million USD (let's round to $1 million), and the average year 2000 failure rate of 23%, if the organization attempts 20 projects in a year (a $20 million budget), it loses (conservatively rounding down) $4 million USD on four failed projects.
If modestly adopting IID leads to a slightly reduced failure rate of 17% (vs. 23%) then for every 20 projects attempted, one more succeeds and roughly $1 million is saved.
In this case, even a substantial investment in education and consulting expertise to transition to IID pays off handsomely. If a medium sized company with 10 projects per year and an annual $10 million project budget invested $100,000 in IID skills transfer, then on average in two years one more $1 million project will be successful. Assuming a two-year analysis and a 10% discount rate, this investment gives a (rounding) NPV of $700,000 with an IRR of 200%, not to mention reclaiming the lost opportunity cost of putting that $1 million project to work.