Roger Sessions, a world-renowned expert on IT architectures, has been out here recently to drink our world-renowned flat whites and to work with some Auckland organisations on simplifying their architectures.
First, about those flat whites: when Roger first came to NZ to meet with Fronde architects back in 2007, he went on a journey of experimentation with our coffee to see if he could find anything that matched his beloved Starbucks Doppio Macchiato. Fortunately for New Zealand, our flat whites blew him away and ever since then he’s been coming back to have more.
On his latest visit for Microsoft, covered in Computerworld, he brings the not-entirely-startling news that complex architectures cost money. And then he startles us by adding numbers to that: $NZ5.4billion in opportunity costs due to project failures caused by overly complex architectures. He’s also written a white paper about it (PDF).
That’s a big number. And it’s an annual cost!
His logic does make sense: every failed project has both direct costs (the money that was spent on the project, which is basically thrown away) and indirect costs (the benefits that will never be realised, the operational costs that continue to rise unchecked because the project didn’t deliver, and the opportunity cost of using the project funding to do something else). Projects developing highly complex solutions are far more likely to fail than projects delivering simplicity. In his whitepaper he uses publicly-available statistics (including Standish Group’s CHAOS report) to put a likely dollar figure on those costs.
Roger’s solution is simple: simplicity. But that’s hard, too – actually spending the time up front to simplify, simplify, simplify, and knowing what to do to actually get less complexity. That’s where Roger’s SIP methodology comes in – check out his website for a lot more detail.
We agree with Roger - we’ve been using his SIP approach to design solution architectures for our clients since we first met him in 2007. We’ve seen it deliver architectures that were easier and cheaper to build, to test and to maintain, and we’ll continue to use SIP wherever we are faced with complex requirements.
