Good luck in getting your CFO not to ask it, but it turns out that "what's the ROI?" is the wrong question to ask about service-oriented architecture initiatives. According to a Gartner analyst, any attempts to assign a specific ROI to SOA "should be viewed with heavy skepticism."
The analyst makes an interesting analogy in this Datamation article, comparing SOA to car parts. While it's possible to calculate the ROI of a car, it's next to impossible to do so for, say, the nuts and bolts used in a car's construction, he says. SOA has a similarly integral -- yet hard to quantify -- value on business projects.
SOA is an evolutionary process that will require a leap of faith -- not a move that many CFOs are comfortable making in these budget-conscious times. Companies that worry too much about SOA's short-term impact on budgets and project deadlines will likely miss out on the strategic value of SOA, the analyst says.
Indeed, recent research from Aberdeen Group shows that 47 percent of companies consider "unknown value for ROI and return on assets" a barrier to broad SOA adoption.
An evolutionary approach was key to the success of a three-year SOA project recently completed by Volvo Cars Belgium that tied together all of that country's dealerships and linked them back to HQ in Sweden. Volvo achieves its aim of improved customer service by tracking the cars through their lifecycle, from placement of the order. to delivery, to ongoing maintenance activities.
Customer approval ratings have increased under the new system and there have been other improvements as well, such as speedier delivery of parts needed for repairs.
Using SOA involved a lengthy, and certainly not easy, chain of events. The biggest chores were upgrading hardware and converting existing data into an SOA-friendly format. Once this was done, however, integration between individual dealerships and the mainframe-based corporate system was "so simple," says a Volvo executive quoted in this CIO Insight article.
Echoing the advice from Gartner, a ZapThink analyst says companies should forget hammering out a big budget for SOA. Instead, he says, create "a bunch of little budgets for projects that will use SOA."