We had 26% of respondents in our recent reader poll report they had a working SOA. That’s out of 300 non-software business end users. Obviously our readers are going to be the more advanced type, but we set a fairly high bar in that services had to conform to an actual enterprise architectural plan and can now run across heterogeneous systems. Another 33% of respondents reported they have operating Web services, undertaken legacy integration and adopted some basic standards. I imagine many of those users would answer “yes” to the question of whether they have an SOA, but we thought it was important to differentiate the two groups. We also broke down the SOA group, with 15% of the total survey respondents reporting they’ve moved into some advanced areas of SOA maturity. That would probably represent the true vanguard group.

A 2006 survey by eBizQ reported that 28% of respondents “have their SOAs deployed, with only 19% having more than 50 services up and running in the SOA production environment.”

Now I don’t know what qualifies as “deployed” in that survey and I suspect that means something less than enterprise-wide deployment. After all, once you get into huge corporate entities you rarely find anything that’s truly enterprise-wide. Also, I find using the number of services to be an awful measure for the maturity of an SOA. All a bunch of services might indicate is that you’ve built stovepipe Web services willy-nilly.

A recent survey in Data Strategy Journal reported 17% with an SOA “running full steam.”

Probably the real question to ask, and one I suspect very few could answer, is how much business are your services doing? Have you tied agility to revenue? That’s the real grail and it’s probably a small subset of the “I’ve got an SOA” crowd who can claim it.

It’s important to separate the fairly ancient (in IT terms) concepts behind SOA from what Schmelzer, Bloomberg and Hurwitz have been talking about in terms of competition. What they’ve been saying is that if you fall behind here, one or more of your competitors may be in a position to eat your lunch. A recession would create an opportunity for that scenario to play out in multiple sectors.

Maybe SOA-savvy won’t make a difference, but a slow economic year is just the thing to put a lot of tea leaf predictions to the test. We’ll probably know more about the value of SOA after a bad economic year than we would after a good one, which is where I was trying to drive with this blog entry.

I find it difficult to believe that even 5% of U.S. Fortune 1000 companies have anything that even closely resembles an enterprise-wide SOA given the level of effort such an undertaking requires. But that is exactly what I believe will be needed to achieve the level of demonstration sufficient for that “golden opportunity” the article mentions, to materialize.

With regard to the gushing references to Mssrs. Schmelzer and Bloomberg and Ms. Hurwitz, as somehow being prophetic, give me a break. While I respect and enjoy their writing on the subject, the concepts behind SOA go back at least 10, if not 20 years.

Overall, I’d suggest a little more researched fact and a lot less unsubstantiated opinion would make SOA Talk somewhere valuable to visit on a regular basis.

Christopher Casey