SOA Metrics, Projects and Organizations

In Measuring the agility of an SOA approach Nick Malik proposes a metric to verify the “agility” claims for SOA. While I applaud Nick’s efforts to create some objective measures for the business case around SOA, I think that Nick’s proposed methodology falls down in step 2.

Nick proposes you apply the agility metric to two projects – one SOA project and one non-SOA project – and then compare the results. Hopefully the SOA project would demonstrate a higher agility measure.

The basic problem that I see is that “SOA-ness” is not a property of a project, “SOA-ness” is a property of an organization.

SOA is more about the general organizational approach to delivering projects and involves things like:

• The organizational structure that defines and delivers IT projects (including the way projects are planned and funded).
• The technical standards, frameworks and best practices that have been put in place and refined over time.
• The SOA assets (Services) that have already been delivered and are available for re-use on any given project.

If we wanted to compare metrics from two different populations to determine if SOA truly does provide benefit, then we would need to compare at the level of organizations. Find two different organizations – one that has adopted an SOA approach for IT management and one that has not. Then find which is the most agile or which has the better ROI or whatever favourite metric you have.

In practice I don’t think this is realistic. Embarking on a true SOA approach requires significant organizational change and that requires something of a “leap of faith” for the early adopters. Over time, hopefully the benefits of SOA will become apparent. Until then, I keep my eyes open for studies which can demonstrate SOA benefits and best practices – but I won’t hold my breath because as with many of these things, “your mileage may vary”.