Before investing in a thorough SOA effort, one must take into consideration the expected life span of the business processes one is implementing as well as how subject they are to change. The shorter the life cycle and the more subject to change, the less return one will get short term and mid term on their SOA investment.

Abstraction of the business processes and the benign implementation of the IT processes are key. SOA is a lot more than web services. It is a philosophy of automation. This philosophy, if strictly adhered to, can be an expensive way to implement a system. It is not for everyone all the time, but it can function as a guiding ideal.

What would be an example of a short business life cycle? If a company wants to try sell a new product line to see if it works out, there is no sense in building SOA castles in the air until the business owners know it is profitable to continue advertising and selling this product. It may sell well only until someone else invents a similar and cheaper widget. Likewise with providing a new service, the management must know it is going to commit to providing this business service before it is worth throwing a whole new scenario in the lap of a software architect.

In this same vein, one must ask how often is the mission statement of the business changed? Has technological change radically altered the business purpose over the last 10 years? Or does the business instead cater to basic human needs like food which won't change except for subcultural and popular trends? Also the complexity of the business processes should be examined. Can the processes be modeled or are they obscured by a hundred different mitigating circumstances?

An additional question which must be asked is just how well can the customer communicate their own business processes? Is there anyone for the architect to talk to who can give definitive answers? The answer may well be "no" in more cases than one thinks. Some businesses never get their act together as the managers in charge are merely reacting to circumstances and not acting. Even stunning and well funded SOA work can not make up for poor business leadership as this is akin to efficiently wrestling alligators when the real goal is to drain the swamp.