SOA ROI in Question?

Everyone is talking about this study from Nucleus Research stating that SOA is having limited success when considering ROI.

“Only 37% of enterprises have achieved a positive ROI from SOA deployments.”

Michael Krigsman followed up with his comments here, and I suspect many will have something to say about this.

Joe McKendrick also had some interesting comments in his blog.

“Frankly, I’m surprised… that the results are that high. About 37 percent — more than a third of companies — seeing positive ROI from SOA doesn’t seem too shabby, in my humble opinion. Remember, most companies are just starting to learn and implement SOA methodologies. Most don’t have key performance indicators that link SOA activities to business metrics. How many other types of corporate ventures have seen results like that while still just out of the starting gate?”

In looking at the study it appears that only one data point was looked at…developer productivity. As I’ve found time and time again, developer productivity is just a side benefit from SOA. Indeed the core benefit is agility, or the ability to have IT work better with business. Unfortunately, agility is not an easy metric to figure out, and thus many who are implementing SOA focus too much on reuse, and not enough on the larger more strategic picture. As Joe puts it:

“Uh oh…. So far, developer productivity is the only real tangible metric most SOA efforts have been able to demonstrate — it’s the most easy to capture, and the metric most obviously tied to an SOA effort. This may be where those 37 percent are getting their positive ROI.

But the promise of developer productivity boosts is not enough to get the business all excited about SOA. Enterprise-wide support for SOA hinges on the ability to demonstrate value to the business at large — more growth, revenue opportunities, and all that good stuff. These are still uncharted waters.”

The larger issue is that SOA, at the end of the day, is a systemic change in the way organizations approach enterprise architecture. Thus, the benefits will only be understood when the architecture has undergone that change. I suspect those surveyed were talking about small projects, and typically projects that just stood up Web services and called it a day. There is not much benefit in doing that unless you’re willing to take it to a logical conclusion, including a complete overhaul of your architecture and then measuring the value of agility and reuse at the end state.

To Joe’s point, I’m also surprise the success rate was 37 percent considering the tactical nature of the projects I’m seeing today, and the unwillingness of many enterprises to turn SOA into what it really is…an architecture. Architectures take careful planning, design, as well as implementation. Moreover, the ROI is years out, but still very much worth the effort.

As Krigsman puts it:

“The takeaway: if you’re considering a SOA project, meticulously plan ROI right from the start. Assuming your project rests on a solid business case, careful management will help protect that all-important business payback.”