Is there really such a thing as an "SOA business case," or is it a paradox?
In presenting a business case, the three letters "SOA" probably shouldn't even be mentioned. However, SOA inevitably results in a bump in costs (at least for the short-term) that needs to be explained and sold to the business.
SOA proponents need to act locally, but think globally
In this new article, Rich Seeley explores the challenges around building a business case for SOA development, and the consensus of the analysts he spoke with (including yours truly) is that SOA may be a behind-the-scenes enabler of what you are trying to accomplish, but no one should set out to acquire funding for SOA for SOA's sake.
As ZapThink's Jason Bloomberg put it: "The most important thing to keep in mind about a SOA business case is paradoxically that it might not look like an 'SOA' business case at all. It must have a clear description of one or more specific business problems, along with the proposed solutions to those problems, along with their costs and risks -- just like any business case."
That's great advice. Build your SOA business case on one business problem at a time. However, while acting locally, think globally.
The problem is, SOA, at least initially, costs money. The larger the organization, the more money it's going to cost. It actually may cost less to solve a business problem with conventional technology than to address it through a service-oriented architecture approach. Over the long run, as the SOA gets up and running and stable of services grows, economies of scale kick in. The second time around, as another business process or unit is introduced to the interface or service, it's a little cheaper, and so forth. But until then, there's some up-front investments required that will be reflected in per-project costs.
Jeff Schneider of MomentumSI outlines some of the key resources required for an SOA-based effort, while warning that shifting to SOA costs more up front, and that companies shouldn't expect to see results right away from their SOA investments.
While Jeff's costs estimates for SOA projects are targeted at larger organizations, there are some important components that need to go into any SOA effort: lots of training, governance teams, enterprise architecture teams, and deployment teams. For the short term, at least among large organizations, the costs of supporting these additional teams and programs could run into the hundreds of thousands of dollars.
As mentioned above, this investment in SOA needs to be built into ongoing business initiatives, and not treated as an "SOA budget." However, even with SOA subtly built into the business case, the short-term cost per project may be somewhat elevated -- which may have to be explained. Thus, the writer of the business case needs to get evangelistic, and be able to talk about the long-term savings and advantages SOA brings to the business. Business managers won't want to hear about the bits and bytes, however. They want to hear about agility and faster time to market. Big-picture thinking is required -- what's the impact on the enterprise as a whole?
To this point, Bradley Shimmin of Current Analysis made some good recommendations, as outlined in Rich's article: In any business case, don't talk about SOA and its thousands of acronyms, such as ESB, WSDLs, and everything else. "If I'm writing a business case, the first thing I want to understand is my environment and what's wrong with my environment, and what fixing it is going to do to make my environment better."