According to a Forrester research report from 2006, 44% of smaller companies (fewer than 1,000 employees), said that deploying Service Oriented Architecture is a high or critical priority. Among SOA’s more natural constituency, big companies (40,000 or more staffers), a hefty 67% reported they would be using SOA by the end of 2006.
But as SOA’s inevitability grows, a question is still asked – skeptically – by cost-conscious IT managers: What is the return on investment for SOA?
However, this isn’t the most effective question to ask, says Forrester analyst Randy Heffner. As he wrote in his report about how businesses pay for SOA, “Any attempt to assign a specific ROI to SOA should be viewed with heavy skepticism.”
On its surface, that statement sounds like corporate heresy. Not assign an ROI? What element of the enterprise budget escapes the all-important ROI discussion?
But Service Oriented Architecture needs to be viewed in a larger context. “Because you should not be selling SOA. You should be selling the solutions that you’re building,” Heffner tells Datamation.
Nuts and Bolts vs. a Complete Machine
SOA is, in essence, the approach, the platform, the process. It can’t necessarily be cost justified as would, say, a new piece of machinery.
In the same way that the cost of nuts and bolts are not assigned a specific percentage of a new car’s value, SOA cannot be looked at as a precisely separate element.
When you try and say, “Overall, industry wide, what’s the general ROI of SOA? It’s like saying, what’s the ROI of nuts and bolts?” Heffner says.
“I can tell you the ROI of a car, but the pieces and parts of it?” Because SOA is the approach, not the solution, cost justifying it in the context of a quarterly budget is tough – and perhaps even counter-effective.
“In some sense, there are ways you look at SOA and it’s just best practices. Should you have good strong design for your system? Well, yes. Just like I want good strong design for my car.”
Not that SOA doesn’t contribute to the bottom line. There are myriad ways it can do so. As Heffner notes, SOA can enable your company to access major new markets; it can help create better customer loyalty; and it may reduce application maintenance costs. The possible benefits are vast.
But, again, many of these benefits do not have a specific dollar value because they are part of larger processes. What part of the fatter bottom line is due to SOA, and what part is due to, for instance, the new product launch that SOA made easier?
A Tale of Two CIOs
There continues to be significant misunderstanding of the value of SOA in the enterprise. Short term – but critical – factors like budget and deadlines can push companies away from the more involved planning needed to begin deploying SOA.
For instance, Heffner points to two very different attitudes voiced by two CIOs at a recent Forrester conference. One executive, the CIO of Humana, had grasped the larger value of SOA and come to terms with its related expenses. When asked about SOA, he said, “It’s just the cost of doing business.” But the other CIO, who apparently had yet to explore SOA, said, “Yeah, that’s like the latest in the reuse thing, right, like objects and components?”
His comment, Heffner says, showed a complete lack of understanding of the strategic use of SOA, about how it can add business value. It’s this narrow attitude that steers some companies away from adopting Service Oriented Architecture.
“If you don’t understand some of the value, and think of it as a way to achieve reuse and drive down a few categories of IT costs, then you may think it’s better just to do something else,” he says.
“People can be short-sighted and shoot themselves in the foot all the time.”
A Strategic Bet – Or a Sure Thing
Clearly, the concept of justifying SOA as a necessary strategic move rather than a specific profit generator will be a stretch for some managers.
“Executives are always happier with hard dollar justifications,” Heffner notes. “But executives also make many decisions based on strategic justification.”
In some cases, investing staff time and budget dollars in SOA may be a kind of “strategic bet.” And even the savviest strategic bet may require a leap of faith.
Heffner points to Apple’s decision to invest in the iPod and iTunes. “There came a time when it was a strategic bet. Somebody could have all the data behind them, ‘We think this will sell this much or that.’ But at some point it’s just pure speculation. And they had to say, ‘Here you go – we’re going to do this.’”
On the other hand, in some cases a SOA deployment can indeed be tied to a specific business requirement.
“There’s a logistics firm that I talked to that the core enabler, from a business justification standpoint, for doing SOA was in going to the executive management of the board and saying, ‘The core to our business model is flexibility.'” However, the firm found that its IT infrastructure was not built for flexibility. Therefore, they decided: “We need to restructure around the idea of flexibility. SOA is the key thing that will help us do that.”
This cost justification is probably more straight-forward than that used by many firms mulling a SOA deployment, Heffner notes.
Getting SOA “for Free”
There are some lucky businesses that get SOA “for free.”
These are enterprises who have opted for a vendor, like SAP, who is building SOA capability into its platforms. If that company stays with SAP and upgrades, it will automatically gain SOA functionality – without a lot of hand wringing about justifying the cost.
This doesn’t mean these fortunate firms have utilized SOA to its full potential. “They’ll have SOA infrastructure,” Heffner says. Truly making the most of this capability is another matter.
Still, this “free” method is a good choice. “It’s reasonable if you can get done what you need to for your evolution toward SOA with tools that your vendors provide.”
One Word: Evolution
If Heffner were to sum up the key principles for companies to keep in mind as they adopt SOA, he would use one word: “evolution.”
As firms navigate the myriad challenges of SOA deployment, “Everything comes back to: How are we evolving our applications? How are we evolving our platform? How are we evolving with the market’s maturity to do SOA? How are we evolving with our own maturity, our processes, our governance?”
All these questions don’t need to answered in the initial plan. Instead, it’s “a step at a time, judged by the business value that we accomplish each step of the way.”