FloridaAndMinnesota

Jim Johnson and his team have spent the last few years going around various software projects and measuring them. They have measured quite a lot of projects actually, about 30 000. They measure all manner of things, and eventually put each of the projects into one of three different categories:

• Successful (on time, on budget, most of requested features working)
• Challenged (late, over budget, fewer than requested features)
• Impaired (a polite way of saying 'failed' - cancelled or unused)

After a few years of doing this, Jim and his team have reached an encouraging conclusion. Although still only about 28% of projects fall into the success category, the percentage is gradually improving year on year.

Their results are more dramatic, however, if you look at them in terms of the size of the project. It turns out that small projects (under $500 K) have a 68% success rate, whereas they have not seen a single successful project that cost more than $10 M.

So what has all this got to do with Florida and Minnesota then?

Well, Florida and Minnesota are actually very similar states, when seen from the population demographics point of view. They both have the same kinds of laws and both needed a software system to handle their child benefits (or some such). The point is, they both needed exactly the same software system, but they went about getting it in very different ways.

Florida began their project in 1990, and estimated that it would cost $32 M, employ 109 people and be ready in 1998. Well, if you remember, Jim already said that he hadn't seen any successful projects that cost over $10 M, so unsurprisingly, this one is no exception. So far they have spent $170 M, and now estimate that they will be ready in 2005!

Minnesota took a slightly different approach. They started in 1999, and estimated that the project would cost just $1.1 M, employ 8 people, and be ready in mid 2000. Well below the $10 M watershed, the project was a great success, and actually delivered early and on budget. Florida could have paid each of the 8 Minnesota developers $10 M and still saved money...

All this just goes to show how much better it is to employ a small, effective team than to try to control a huge juggernaut. It reminds me of a conversation I had with a guy in the bar one evening. He told me that five years ago he was a very important software architect working for a large multinational corporation, with 800 developers under him. Today he works for a smallish consultant firm, and spends his time coaching his client's software teams in XP.

What struck me was how much happier he is now. He said that before he used to be forever defending his beautiful UML diagrams to developers who always thought they knew better anyway. Now he works by sitting down beside people, encouraging and guiding them to produce the great designs. Much more fun.

So, it just goes to show that next time you're planning a software project, you should make sure you're dealing with a Minnesotan XP coach, not a frustrated architect from Florida :-)

-- EmilyBache

back to Xp2002Stories

KeithB has a strong suspicion that that Florida job was done by a company he used to work for.³host³³date³June 10, 2002³agent³Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 4.0)³FloridaAndMinnesota