Good morning. Yesterday, SOA was the end-all, be-all. But this is IT. These things change overnight.
Guess what? The SOA backlash began.
Yeah, well, that's what you're going to see in the headlines anyway. If you've been around IT circles more than a few months, you saw this coming the first time you even heard of service-oriented architecture.
According to Nucleus Research, an IT research firm, a mere 37 percent of companies report a positive return on investment for their SOA implementations.
That's right. The most dreaded of three-letter acronyms words for techies: ROI.
There are mitigating factors for the SOA ROI. First, most SOA projects are still in the early, project-level stages. In other words, people are still testing it out with small projects - and that's a long way from having a full blown loosely coupled architecture.
Even the Nucleus press release notes that, "further findings suggest that while SOA drives developer productivity, it often ends with the project or department-- limiting the ability for broad-based adoption, impacting return on investment (ROI)."
It's been said, and I'll say it again - the payoff won't necessarily come on one project. It'll come as you build out and are able to repurpose your services for other projects.
Gartner has even stated that "What's the ROI?" is the wrong question to ask.
So why can't everyone get that ROI isn't the point right now? I mean, everyone involved with SOA obviously gets it.
ZDNet blogger Michael Krigsman asked that very question of Ian Grant-Smith, chairman and CEO of BAP Global Solutions, a SOA solution provider. Grant-Smith's response was the polite version of, "Yeah, right":
Senior executives constantly hear about the business benefits of SOA, but they aren't seeing results....Promises are not enough and executives are right to be unhappy.
I'll let you read the full quote on ZDNet, but I think you get the point.
ROI is a problem. But I think the survey's further findings hold the key to unraveling this ROI dilemma. Consider this: "According to a survey of 106 enterprises conducted by Nucleus Research and KnowledgeStorm, fewer than four in 10 developers use SOA. This statistic is in stark contrast to the marketing messages of the industry's major players, which tout SOA as the "Holy Grail" for improving developer productivity."
Four in 10. That's a problem.
What's up with that? According to David O'Connell, a senior analyst at Nucleus Research, developers view themselves as "creators of code and applications and shun SOA because it forces them to reuse code generated by others."
Holy Cultural Crutches, Batman!
You know what would help with that? Training. But are organizations investing in training? Of course not. In fact, organizations are pretty cheap when it comes to investing in SOA. SOA consultant David Linthicum recently cited funding as one of the top five reasons SOA implementations fail.
Of course, that's a bit of a chicken-and-egg problem. Executives want to see the ROI before they give the FUNDS. Wasn't it the CIO in Greg the Architect who said, "I want this thing dripping with ROI?" Yes, it was. But satire aside, Linthicum argues you can't do sweeping changes within the IT infrastructure without money.
(I should note that some disagree. Nick Malick argues that SOA is cost neutral. Malick also works for Microsoft, which - and this is a wild guess on my part - probably has better stuff to start with than 95 percent of IT shops out there.)
So: Where does this leave SOA?
In the same place it held yesterday. Start small, find some low-hanging fruit and deliver real business value. Explain that this is a long-term commitment. Try a few more projects. At some point, it will be your judgment call as to whether SOA is hype. But to be fair, most organizations are nowhere near that point.