Someone recently asked on xp-cinci mailing list:

• I need a pointer to an authoritative source containing some statistics (rough of course) of the cost savings of agile over waterfall. I’m not talking about just intial development of course, but something that includes the faster-delivery, cheaper-enhancement, etc, etc. Can anyone point me at something that will stand up to management scrutiny?*

Was an interesting question to me, so I did a little research on the subject, links below, but I’m not sure I’ve found any definitive answers.

• First I think basic education is usually in order. For that I point people to Wikipedia: AgileSoftwareDevelopment article to gain a quick overview of what Agile is, and how it compares to Waterfall. I think the Wikipedia entry presents a fair analysis, pro+con of Waterfall vs. Agile.
• I found a decent paper on results of agile methologies at Jim Highsmith’s website.
• If you want to go negative, point out how ridiculed Waterfall methodology is (Waterfall 2006), with comical contents such as The Glacial Methodology.
• Let their imagination run wild on the cost of premature estimation. Agile methdologies more effectively deal with the “Cone of Uncertainty” inherit in any project (graphic below).

Nice post over at TargetProcess on planning in a relatively agile way. Outline of steps:

• Identify users
• Define user goals
• Define user usage patterns
• Invent functional solution to satisfy user goals and usage patterns, and step-by-step outline of steps in each process
• Sketch UI’s on whiteboard and capture with digital camera
• Iterate UI a little to improve it
• Start development

This isn’t following a pure Agile methodology, but might be a useful transitional process away from a more rigid functional-spec-driven process.

While doing some research on running Agile projects with a geographically disperse team, I ran across this article by Martin Fowler which outlines how ThoughtWorks has been dealing with attempting Agile projects offshore. Great read!