Speed To Business 2

How does the advance of business process management mesh with current plans for consolidating enterprise application implementations?
Some question whether the ERP vision can embrace BPM, which is focused in many ways on ERP's opposite. Certainly, BPM is critical in streamlining ERP implementations and identifying processes inside the ERP realm that could be automated or in other ways made more efficient. A 2003 AMR Research note observes that "the real problem is that packaged applications have been designed from the inside out, focusing on internal process improvement."

The BPM vision is ultimately focused on more than automating routine processes. It's about aligning business goals with IT implementations to capture innovation, so that companies can push rapidly outward and seize competitive advantage as soon as possible. BPM systems must make it easier — not harder — for organizations to be agile, respond quickly to market changes, and even structurally grow or shrink to fit objectives. "Jobs that call for simply following recipes will become scarcer," said MIT professor Erik Brynjolfsson, quoted in MIT Enterprise Technology Review (March 10, 2004). "Demand for an innovation-driven workforce will continue to grow." Companies that employ innovators will depend on BPM to turn vision into reality.

BPM in Action: iUniverse

iUniverse is the largest publisher in the U.S., introducing 5,000 new titles within the last year through all distribution channels, including online stores. The company has dreams of introducing dramatic changes in how authors publish and market their work. To attain its ambitions, the company must be smart about how it manages business processes. Intelligent Enterprise recently communicated with Vernon Stinebaker, general manager of iUniverse (Shanghai) and vice president of iUniverse about the company's BPM experience, chiefly with Intalio's BPMS products.

IE: Companies are finding that they must move from an application-centric vision, which has led to data and processes trapped inside proprietary silos, to one that looks at business processes in an end-to-end fashion, crossing applications as necessary. Do you see the iUniverse system as exemplifying that trend?

Stinebaker: Our implementation currently spans more than 10 business processes that not only integrate our offices, which are located on two different continents, but also allow us transparent integration with our business partners, who, for their part, may be unaware of their participation in a business process. I think this very much reflects the effectiveness a BPM system (BPMS) can bring.

Our adoption of BPMS has been very much a strategic direction for iUniverse. We envisioned that we could leverage a BPMS to achieve significant improvement in the following areas:

  • Flexibility: the ability to quickly design, develop, and roll out business processes that would be readily adaptable to a broad range of tasks
  • Efficiency: the ability to decrease time to market
  • Effectiveness: the ability to decrease the resource required to design, develop, deploy, and maintain business processes.

We've found that two different levels of thinking are necessary to allow effective BPMS deployment. Both relate to your point about the problems with an "application-centric" vision.

First, most legacy systems are designed to be "fully integrated": that is, they attempt to provide a complete solution to a given requirement. They typically fail, creating a constant need for development, improvement, bug fixes, and so forth. However, the main difficulty is that most business processes span multiple requirements. For example, we not only need to accept an author's manuscript into our system, we also need to accept and validate payment. These activities obviously span multiple organizational departments, and from a traditional systems point of view, multiple applications. Looking at the business process and designing and implementing supporting systems as a process flow implemented through (hopefully reusable) components or services — as in Web service-oriented architectures — provides an improved paradigm for understanding, designing, and implementing the process. The overall system effectively "bridges" the silos and may ultimately eliminate them.

My second issue with the application-centric vision is that developers have grown accustomed to thinking of applications in fairly monolithic terms. While they may achieve some level of code reuse through code libraries or "cut and paste" techniques, the use of granular components that deliver a single service is an unfamiliar paradigm to many developers. Getting them, as well as other business and technical team members, to change from an application-centric mindset and think in terms of end-to-end business processes and reusable Web services takes education.

IE: Is the process modeling a major aspect of the iUniverse implementation?

Stinebaker: Yes. We considered the modeling capabilities as a key requirement during our evaluation of Intalio's product. We use the modeling capabilities on a broader basis as our roll-out continues.

IE: Is process automation the primary goal of the implementation? If so, are you looking at how you'll scale up the system in terms of complexity and greater numbers of processes and users?

Stinebaker: Our processes include a full spectrum of automation, from fully automated transactional, to fully automated periodic to combined manual and automated processes. So far, we haven't seen any scalability problems, and don't anticipate any. I should say that as a small to midsized enterprise, we may not have the scalability requirements that larger BPMS implementations may have. However, we have experienced no system failures or downtime in more than two years' production use.

David Stodder is editor-in-chief and editorial director of Intelligent Enterprise.