Savings realized by adopting Agile (Scrum and XP) over Waterfall

Question

I was wondering if someone can share their exeprience or point to a study on cost savings realized by adopting Agile (using Srum and XP), over traditional Waterfall.

Answer 1

In my opinion, cost savings is only a part of the equation. Return on investment is a better measure since it takes into account both costs and returns, with the time value of money. Being conscious about your costs is important, but it is even more important to develop software that brings in value. There's only a certain limit you can cut your costs down to (zero) but there's practically no upper bound for returns.

Agile helps you on the returns side as well, but since this question is about costs, let's not go there.

Biggest cost driver for software development is employee salaries, which in turn is driven by project effort. Therefore the key to cost saving is minimizing the effort required to get the job done.

Some quick points related to agile costs vs. waterfall costs:

• Cost of change does not grow as rapidly as with waterfall. It is cheaper to introduce changes relatively late in the project. This is achieved by constant refactoring, test automation safety net, continuous integration and splitting the work into stories allocated into iterations.

• Defects can be found and fixed earlier due to frequent and early releases, continuous integration and test automation. It's the cost of change curve again.

• Projects can be terminated early while still bringing in some value with the most important stuff ready.

• Working in short iterations helps keep developers focused on the important stuff. This mititages student syndrome by having frequent deadlines and Parkinson's law with short time-boxed iterations. There's less room for slacking off.

• When the project is running smoothly at a steady pace, people are not as likely to burn out, switch jobs etc., all of which is costly.

• Feedback from previous iterations helps keep the project more predictable, thus preventing cost overruns. You have actual data on the project's performance, not just plans based on wishful thinking.

Some further reading on software project economics:

• Steve Tockey: Return on Software: Maximizing the Return on Your Software Investment
• Walker Royce et al.: The Economics of Iterative Software Development: Steering Toward Better Business Results

Answer 2

There is a paper by David Rico comparing the ROI of different methods, agile and non-agile. The paper is basing the comparison on ROI data found in studies of different method implementations.

Answer 3

One of the most significant cost savings associated with Agile (and to a lesser extent other Iterative methodologies) is that you fail fast. If a project was never meant to be, you'll know a lot sooner. The benefits of this should not be underestimated.