ROI Watch

Rangaswami says the crucial message for CIOs is that there is no need for them to spend much time scoping out the technology, but it is vital for them to pay close attention to the return on investment.

"So our advice is - don't wait to create a project for Web services: instead, intercept an existing project and see if you can put Web services into it. If there's a project today that you're doing, you just see how you can put Web services into the equation. This is much more of a practical approach, where you're not trying to create something just to fit it in, you're trying to see how it works within an existing project."

Initial Web services projects should be small, Rangaswami advises. And while few of the early adopter respondents reported failed projects, they were able to provide some other practical advice.

For one thing, projects should have predefined metrics of success. Projects should be discrete and promise potential cross-project benefits. And CIOs can gain by starting off by identifying critical data that has three common characteristics: it is widely distributed across the organisation, highly dynamic in nature, and inaccessible to those who need to use it.

"Many respondents suggested intercepting an ongoing project and adding a Web-services component to it, rather than planning a new one from the ground up," Rangaswami reiterates. "They also set up initial parameters for the projects that followed the enterprise's data model and budget constraints. And they shared lessons learned from these projects across the company."

He says the two-year road map the report provides notes in the early stages that CIOs should pay no attention to architecture or other issues. "Once you get your hands dirty and your feet wet and so forth, then the next step is to create a services-oriented architecture," he says.

Other advice comes direct from participants. For instance the biggest benefit of Web services does not occur the first time, but on recurring integration, says the vice president of technology at one Fortune 500 financial services company, who participated in the Sand Hill study. Web services allow the company to reuse the application interface in another integration setting, within a different business process, or with partners.

Web services helped this executive's company smooth the impact of a major merger. The two companies, rather than wait for a full-scale integration of their IT systems, wrapped their back-end applications with Web services interfaces and quickly created a portal front end that offered a single face for all their customers to access.

Moreover, Rangaswami says CIOs should never take Web services to the board until they have "got their heads around a few projects", for fear the board will start demanding new applications before they are ready. "Don't go to the board first, because then the board is going to think: 'Okay, here's the next real cool technology.' That's not going to work. You know, you're going to do this grandiose vision and spend a lot of money and that's not going to work for you."

Better by far to undertake a few projects, show some wins, learn from the experience and then put together your own road map to take to the board.

Although Rangaswami is reticent about offering case study examples (the ones in his own company's report are confidential, and he says most of the others out in the ether have come from vendors and so are less than objective), he says the financial services, insurance and automotive industries have been at the forefront of adoption of Web services.

And he offers this further advice to CIOs keen to explore Web services.

"Don't get carried away with the technology. Don't wait for more standards to emerge. We think that a lot of the standards that might come about in the future may not be as helpful. You play the waiting game always if you are waiting for standards. We think there are enough underlying standards like SOAP, that you can get started, so don't use that as an excuse to wait. There's enough there that you can use today. There's enough there in the technology to do very, very interesting projects today.

And when it comes to turning to outside organisations for help, he says while the systems integrators are focused and have started building practices around Web services, there are also smaller consulting firms and some interesting "offshore" organisations - in places like India and China - that are getting more adept at doing this work quickly. CIOs should have reasonable confidence in turning to those organisations for help, he says.