July 22, 2004 - Reflections on SDForum Agile Summit

On Wednesday I attended the SDForum Agile Summit which was a 4.5 hour program with two keynote speakers and three panels. Much of the information was familiar ground to me but the discussion on the need for courage made the trip worthwhile.

The keynote speakers I'd describe as an effective pairing of the guru and the convert. The part of the guru was played Ken Schwaber, Agile Alliance chairman and co-developer of SCRUM. The convert was the remarkably, surprisingly impressive Bob Shatz, VP of Development at Primavera.

Ken had two messages that stood out for me. The first was his timely, off-shoring related take on ROI and second was on the pain of implementing agile process.

For the first his root claim — and I think there's some truth to this — is that given the high failure rates of software projects companies off-shoring are looking to “fail for 1/2 the cost”. Aim high indeed. In setting up his alternative Ken pointed out a Standish Group finding that 60% of software functionality is rarely or never used. Ken's solution: adopt a methodology that will let you stop 40% sooner. He then combined this with the productivity argument. The industry average he said is about 2 function points per developer per month; he said the average increase he sees from the teams he works with is a 4-8 fold increase over their previous productivity. Which would you rather have? Failure at 1/2 the cost or success at an even smaller fraction? Seems like a no brainer when the guru is talking...

... but I don't see this argument as a winning one, at least not from the point of converting the industry. Several times I've heard people relate how agile processes are derived from lean manufacturing, and lots of time I've heard the question (and in fact this was the topic of one of the panels) “what will it take for businesses to embrace agile?” How come I've never seen anyone combine those to ask the obvious question “what did it take for business to embrace lean manufacturing?“ I haven't studied the issue but my naive answer is that they — and I'm thinking of the US auto industry — only made the change when they were absolutely forced to by the competition. Is it too much to suspect that history will repeat itself? Currently off-shoring is associated with lower costs; what will happen when it becomes associated with higher quality?

The good news? Us agile early adopters should have a competitive advantage until our success forces everyone else to play catch-up.

Ken's second message, the pain associated with implementing agile, resonated with me and my personal experience. The gist is that agile is centered on a philosophy of “inspect and adapt”, and while the process is simple on paper, and even simple to start using, the initial results hurt. In Ken's words, once you starting taking a close look at what's really going on “life gets very hard”. Suddenly your process surfaces everything that's wrong, things that have always been wrong, and it is easy to blame the process for these revelations. It takes courage and motivation to get past the organizational denial.

Appropriately it was both courage and a strong humanistic streak that radiated out of Bob Shatz as he told the story about the adoption of SCRUM at Primavera. Bob related how after the 3.0 release he was just completely unwilling to put his team through hell again, he wasn't going to sign people up for another death march, but didn't know what to do. He read something written by Ken, called him up, and ended up hiring him to help convert Primavera. The courage Bob described was two fold:

  1. he took the personal risk of making this change — at least in the beginning — “behind the back” of the company co-founders,
  2. he was willing to face the problems that came up and continue with change even as the process was painful.

I really liked his point that if you're going to adopt agile you've got to ready to “deal in the reality, because that's what agile is all about”.

Having taken his team through the entire 4.0 product development cycle, what did he seem proudest of? That they didn't work a single weekend, not a single day of mandatory overtime. (Also interesting was an issue he hopes to address for 5.0, the fact they spent their last two 30-day sprints fixing bugs. “Not very agile, is it?” So for the 5.0 release he has his team adopting Test-Driven Development and pair programming.)

I've seen a lot of discussion about the satisfaction for individual contributors in using agile techniques. Hearing Bob talk — hearing the pleasure and satisfaction in his voice — made me think we need more of these stories about how much more fun it is to manage and lead an agile development team.

Posted by Jeffrey Fredrick at July 22, 2004 04:11 PM

Trackback Pings

TrackBack URL for this entry:


Post a comment

Remember Me?