I had a great lunch discussion today regarding the impact of ownership of SOA within the business. As I have discussed multiple times, SOA is about business services, not technical services, but the reality of this point is lost in semantics.  The following illustrates the different perspectives that each have and how that impacts the overall SOA initiative.

Let’s look at two business capabilities in real estate: renting the building space (sales) and drawing up rental agreement (legal).  Each of these business capabilities are effectively services to the business.  In order to operate effectively, they need to share some basic data, such as property listings and renter information (CRM).  They also operate synchronously as part of a business process that results in the execution of a rental of a property.  Otherwise, they operate fairly autonomously from each other.

From the business’ perspective, at a minimum, they’d like one and only one of each of these capabilities, each delivering in a consistent manner. Sometimes, that is not possible. Perhaps, there needs to be two distinct sales teams due to regulatory or licensing issues. In this case, the business will still want the multiple sales teams all operating in a consistent manner, using the same forms, if possible, and interacting with the legal department to draw up the contracts in a consistent manner. The consistency is what enables the business to operate efficiently even in the face of redundancy.

In this particular case, I believe, while not optimal, it is perfectly okay for there to be redundancy of certain information systems in the delivery of each of these business capabilities as long as it allows the business function to deliver consistently and efficiently. For example, if each of these groups had selected a different portal product to provide access to the property listings, that would be acceptable as long as they were operating off the same listings service. Forcing these groups to consolidate on a single portal product is not going to provide any additional efficiencies for these groups or simplify their ability to deliver.

However, what we see happening in the industry, is that SOA efforts are being led by IT and not by the business leaders. Hence, IT views the problem from a systems perspective and sees the need to remove redundancy at the information systems level. So, instead of focusing on business capabilities and making sure that the business rules, processes and goals are being executed consistently and that redundant business capabilities are consolidated, they look at the redundant portals and identify them as a target for consolidation, as so they should.  ITs job is to keep IT costs down and are rewarded by simplifying the work of the IT department.  However, as I used to remind my developers, it’s not about what’s easier to code, it’s about making the jobs of the users easier, that’s why they pay us.

The result of IT forcing consolidation of portal products is some cost savings, but interruption and change for the business. The overall productivity losses within the business function due to the interruption is typically never identified, let alone measured. Users will need to be re-trained on the new portal, things won’t work initially and users will become frustrated and be less productive. All of this for the possibility of reducing the costs of managing one additional piece of software.

I say let sleeping dogs lie. There are plenty of real information management costs in most businesses that eliminating will offer a real benefit to the business’ bottom line. Clearly, both of the aforementioned business capabilities need to be operating off the same customer and listing data. There’s plenty of work to be done by IT without focusing on needless change.  Eliminating the additional software licensing costs are penny-wise, pound foolish when the real impact is analyzed and assessed.

With regard to SOA, IT will view SOA as an IT effort that results in consolidation of technical capabilities.  That’s their task and that’s how they are rewarded.  IT is not rewarded for improving productivity in payroll by improving the end of month process.  And, therein lies the flaw in the system.  We have become heavily reliant upon our data and applications to run our business, and therefore, IT has become a critical function of the business.  Yet, the business world rewards IT for running with as low of an overhead that can be comfortably be achieved without putting the business at significant risk.  If we really expect SOA to live up to its promise, it is going to have to become an effort that is undertaken by the line of business as they reorganize for streamlined operations, or, if we’re going to let IT own the SOA, then we are going to have reward IT for improving productivity and usability of the systems in a manner that is best for the users and not necessarily IT.  That is, we’re going to have to let them focus on actually aligning the information systems with the needs of the business and not force the business to live with the limitations of the systems.