The Pitfalls of Early Web Services Adoption

Viewpoint: July 16, 2003

Web services technology is delivering real business value today – that’s the good news. Early adopters are generally not getting the balance right between near-term business impact and long-term architectural direction – that’s the bad news. In fact, unless early movers are careful, they may encounter diminishing returns from the technology rather than harnessing the potential for increasing returns.

Based on extensive experience with early adopters of Web services technology, I see some clear patterns of adoption. I’ve discussed some of these patterns in earlier blog entries (see “Where Will Web Services Be Deployed? – Part 1” and “Where Will Web Services Be Deployed? – Part 2”). This time around, I’d like to focus on another pattern – one that is cause both for optimism and for concern.

Web services are being adopted along two parallel paths within the enterprise. The IT department is driving the first path. This path is largely restricted to prototypes and experiments designed to learn about the capabilities of the technology – very few initiatives in fact involve production deployments of the technology. In a few cases, the IT department has become deeply focused on designing a service-oriented architecture (SOA) – seeking to develop a detailed blueprint of what this architecture would look like across the enterprise before launching any specific deployments of Web services technology.

Non-technology line executives are driving the second adoption path for Web services. These are executives with specific business problems. They somehow heard about Web services technology. Somehow, they also became convinced that this technology could address their business problem and developed the courage to spearhead production deployments of the technology. It is this second path driven by non-technology line executives that accounts for most of the production deployments of Web services technology to date. (By the way, this is a pattern that extends well beyond Web services and that will have profound implications for technology providers, but that’s the subject of a future rant.) In fact, the pace of adoption along this path is surprising many of the early analysts who expected a slower ramp up.

This rapid pace of adoption by line executives is encouraging for a couple of reasons. First, it confirms that the economic value proposition of the technology is compelling – pragmatic line executives are becoming early adopters because they clearly see that the potential business impact outweighs the early adoption risks (for a more detailed discussion of this economic value proposition, see Chapter 4 of my book, Out of the Box: Strategies for Achieving Profits Today and Growth Tomorrow through Web Services. Second, it highlights one of the strengths of Web services technology – it can be adopted incrementally and targeted against very specific business needs.

But, as is often the case, strengths can also be weaknesses. Since Web services can be adopted incrementally, you don’t need to think in broad architectural terms about the technology in order to get real business value from it today. You don’t need to, and so many executives, especially non-technology line executives, do not. They are very pragmatically focused on near-term business impact. They are moving opportunistically, addressing one business problem at a time.

There are two problems with this approach. First, from a business perspective, the early adoption of Web service technology is largely ad hoc and opportunistic. That means the early deployments of the technology are targeted against real business needs, but there is no assurance that these deployments are targeted against the areas with highest potential for near-term business impact. There are few, if any, examples of companies systematically surveying the highest impact business areas for Web services deployment. The result is that near-term business impact is generally sub-optimized.

The second problem has to do with long-term potential. The long-term, and most significant, business value of Web services technology will only be realized when it is used to support the deployment of broader SOA architectures that extend across multiple enterprises. Opportunistic, one-off deployments of the technology may solve near-term business problems but, unless they are designed to be consistent with a broader architectural vision, they will contribute nothing towards longer-term value creation. In some cases, they may actually make it more difficult to implement a consistent architecture later.

Make no mistake about it. As useful as Web services technology is proving to be, it is only a modest step in the direction of SOA architectures. SOA architectures deliver a much higher flexibility to business operations based on three key elements: the service concept, the philosophy of loose coupling and the notion of a broker.

– The service concept transforms the way we think about software, making functionality available as a context independent, reusable service, accessible by anyone with appropriate authorization.
– The philosophy of loose coupling creates a more modular way of designing services so that they can be easily brought together to perform specific business functions.
– The notion of a broker is essential to reducing complexity at the end-points, particularly when dealing with long-lived, loosely coupled asynchronous transactions (for example, the execution of a securities trade or the booking of a travel itinerary with multiple travel providers). Brokers, especially in the form of service grids, help make loosely coupled connections both more efficient and more robust by federating shared enabling services. (For more about service grids, a concept that John Seely Brown and I have championed, see our jointly authored working paper “Service Grids: The Missing Link in Web Services” and the extended article “Service Grids: The Missing Layer in Web Services” in the December 2002 issue of Release 1.0)

Once deployed, SOA architectures create a powerful platform for increasing returns because network effects take hold as more and more services become available within the architecture. Equally importantly, SOA architectures are especially adept at supporting activities across multiple enterprises, given the focus on broker-mediated, loosely coupled services. As a result, the increasing returns potential of the architecture is not limited by the boundary of the enterprise – the value continues to increase as the architecture extends across multiple enterprises.

Good CIO’s understand the value of a consistent IT architecture. Done right, it can create enormous leverage where the sum of the parts becomes far greater than the individual parts themselves. But CIO’s can go too far in championing IT architectures. They are often tempted to block near-term initiatives until the broader architectural implications are completely understood. They can also become consumed in developing elaborate conceptual architectures that become increasingly detached from real business need. They may also have difficulty in structuring architectural migration initiatives in ways that can generate real business value along the way – the efforts can become expensive multi-year programs with few clear milestones along the way. At the extreme, architectural initiatives can become black holes, consuming time, money and people at an alarming rate.

– How can companies strike the right balance between the need for near-term business impact and the longer-term opportunities created by a broad-based SOA architecture? The FAST strategy approach I outlined in a previous blog entry (see “FAST Strategy”) holds the key to creating the right balance. FAST strategy approaches help companies to navigate through increasingly uncertain business environments. This same approach can be adopted to help companies migrate to new technology architectures. What key elements of the FAST strategy approach are useful in harnessing the full business potential of Web services technology?
– Focus – The FAST approach emphasizes the need for senior management to define a shared view of the long-term (5-10 year) direction of the business. In a similar manner, the CIO must take the lead in aligning the senior management team around a common view of the long-term IT architecture that will support the business. The FAST approach suggests that this view must contain both a high-level statement of the business need and the architecture best suited to meet the need. The key word here is “high-level”. Avoid the temptation to over-specify. The art is to define an architectural direction that is high level enough to accommodate many different permutations over time, but specific enough to guide choices in the near-term.
– Accelerate – The FAST approach requires that a company move along a parallel path, refining the view of long-term direction while at the same time launching waves of near-term (6 to 12 month) operating initiatives designed to accelerate movement towards the longer-term direction. In the case of Web services, this suggests that companies should systematically identify the highest impact opportunities for Web services deployment, both in terms of near-term business impact and ability to accelerate the implementation of SOA architectures. A few well-chosen initiatives are likely to have greater impact than a broad number of under-resourced initiatives.
– Strengthen – The FAST approach also requires executives to focus on organizational bottlenecks that prevent even faster movement and to launch select near-term (6-12 month) organizational initiatives designed to remove these bottlenecks. In the case of Web services, this might involve near-term skill-building efforts (don’t forget the value of learning through doing) or deployment of shared enabling services in service grids to shorten the lead-times of more specific Web services deployments.
– Tie it all together – In the FAST approach, executives rapidly iterate between two time horizons – a 5-10 year perspective on the business and 6-12 month waves of operating and organizational initiatives. The performance results of the near-term Web services deployments provide valuable learning and insight that can help to refine the 5-10 year view of the SOA architecture that will be most valuable for the business. Conversely, the long-term architectural view can help to make choices and shape deployments of the technology in the near-term.

Web services technology certainly will deliver near-term business impact. The challenge for executives is how to maximize this near-term impact while also building the foundations for the real economic prize: the longer-term business value creation opportunities created by the deployment of SOA architectures.