Study measures the elusive ROI of SOA -- through reuse

Building an SOA component will cost 20% more than a traditional application the first time around.

A couple of weeks back in this blogspace, I looked at the emerging debate over the value, if any, of reuse, noting that many organizations had the challenge of coaxing departments to share or reuse their code in the first place.

Building an SOA component will cost 20% more than a traditional application the first time around

Assuming that departments do agree to buy into the reuse aspect, a report out of LogicLibrary puts some hard numbers to the whole reuse scenario, in an SOA context, in a study of its early adopter customers. The conclusion: an SOA-based service costs more up front to build and test and deploy than one-time use software, but will deliver a payback almost as soon as it is reused the first time around. (A PDF of the report is available here.)

SOA-based reuse is also far more cost-effective than traditional component-based reuse, according to Jeffrey Poulin and Alan Himler, the study's authors. Building components for an SOA requires only about a 20% additional investment over development for one-time use. Building traditional reusable software components can cost up to 50% more than a one-time piece of software.

Once a service is developed for reuse within an SOA context, it is also more likely to actually get reused, the study finds. The level of reuse in a typical SOA application was about 25%, compared to 10% of traditional component-based applications.

Thus, SOA-based services are cheaper to build the first time around than traditional applications, and are more than twice as likely to actually get reused. Poulin and Himler calculate that SOA-based services show a positive ROI after only 1.3 uses, compared to an ROI after 1.7 uses of a component in traditional development.