IBM's Kareem Yusuf linked to an interesting document on his blog today.
He was commenting on a question he was asked by a customer in Europe earlier this week - how do I get started?!
This ties in with what I was saying earlier this week about quick wins. I was arguing that building an SOA for SOA's sake was unlikely to meet with success: you have to be solving real business problems.
Kareem's link is a great example of this in reality: five IBM customers who found that a service-oriented approach to solving a real business problem delivered value - and quickly.