IT business cases and return on investment (ROI) analyses remain key tools for IT governance and investment decisions. Given this, how do organizations cost-justify investments in the relatively immature world of service-oriented architecture (SOA)?
When considering the business case for SOA, most organizations seek benefits such as business agility, faster time to market, increased asset reuse, lower software maintenance, and reduced integration expense or legacy system consolidation. All great goals, but these desired SOA outcomes may fall victim to business case analysis. Consider the much-desired "business agility." How does an organization know it has become more agile? As measured how?
With SOA, the business case and ROI analysis can be tricky. I like to treat SOA as a series of ROI or value thresholds through which an organization progresses on its journey to desired business outcomes. SOA requires sustained investment toward creating reusable, composable services, which are leveraged to deliver value as they are consumed by various developers, internal and external customers, and trading partners. Once you have a portfolio of reusable services in place, you can begin to accrue more value by leveraging those assets. You can orchestrate business processes using BPEL (business process execution language). You can remove process and information latency through real-time event services. Integration expenses will drop because Web services are "pre-integrated." And the ultimate business value you are after -- business agility, or IT consolidation, or faster time to market -- will begin to emerge.
In my opinion, the real value of SOA will come from innovations that leverage reusable, composable services as the substrate for new applications, new business models, and new performance plateaus. SOA will enable service-oriented innovation.
But for proof points, I advocate a more scientific method called an SOA Value Hypothesis. First, create a hypothesis of how your business will benefit from SOA. If business agility is the desired outcome, operationalize the term "agility" so it is measurable; e.g., business agility for company XYZ means achieving 20% faster time to market for new products while also reducing application development time by 25%. Agility will be achieved through orchestrating business processes based on reusable Web services.
From that measurable business goal, work backwards through the intermediate steps required to actually achieve agility from SOA, and determine how you can collect the necessary data to prove your SOA Value Hypothesis. How will you know you are more agile? How will you prove you have achieved faster time to market? Do you have the necessary baseline metrics? Will you use customer surveys to prove it?
The process of measuring your SOA Value Hypothesis is important. Remember, you are still dealing with IT skeptics who are suspicious of anything new. Once you have framed your SOA Value Hypothesis and determined the metrics to prove it, then you can structure pilot projects and proof of concepts to demonstrate it.
Business cases do have a purpose, but too often they are used to deny investment decisions rather than enable business value. Don't let your business case process derail your SOA initiatives. Develop your SOA Value Hypothesis and prove it to yourself as well as to your organization. Science will set you free.