The Appeal of Model-Driven Approaches
Model-driven architecture enables a platform-independent model of business functionality to be created independently of the underlying technology. Importantly, the modeling environment is geared toward non-technical business people. This decoupling of business logic and technology lets business people get in on the act of application/process definition since the models are understandable and don’t require inclusion of the technical implementation details. But model-driven apps aren’t just powerful because business people can create a models; they’re powerful because those models can actually be translated directly into executable code, sometimes with only minor technical modifications.
BPM suites (BPMS) are a prime example of a model-driven application; graphical process models are typically created by business analysts using a standardized notational format such as Business Process Modeling Notation (BPMN). That model, similar to a flowchart view of the process, is easily understood by business users, yet it contains enough information to support a relatively painless and swift conversion into an executable process in the BPMS engine. If integration with systems are required, an IT person will likely be involved to connect specific tasks in the process through Web service calls — often just matching of input and output parameters — but generally little or no code needs to written in order to create an executable process.
At Allianz of America, for example, the business owns the process and maps it down to a specific level of detail before turning it over to IT people who "move it over to the BPM tool" says Tim Rofling, IT Director of Allianz of America Shared Services. Relying on an interative methodology and focusing on small projects with a potentially big impact, the company has managed to implement an impressive number of processes, each within a 60- to 90-day timeframe. Examples include securities application processing, money processing for applying premiums and life insurance underwriting. In a new (insurance) product implementation, Rofling says cycle times to implement new products were reduced from more than 50 days down to 12 days largely because the model-driven approach removed IT development bottlenecks.
Moving to a model-driven approach changes the entire application development cycle: not only do business and IT people collaborate on modeling the business processes, the same model is then used to monitor and manage the processes in real time once they're in production.
BPM and SaaS
Software-as-a-service (SaaS) is no longer seen as risky technology. In fact, it's downright mainstream to the many organizations using systems such as Salesforce.com to manage their confidential customer information. SaaS is used for a number of reasons: to reduce the up-front cost of systems and the IT infrastructure required to support them, to pilot the use of a system or technology without making a major investment, or to bypass the long cycle time of a new system implementation within a large organization.
A few BPMS offerings are starting to appear via a SaaS model, with more expected during 2008. In some cases, these are pre-packaged applications built on a BPMS platform. For example, Enkata offers a contact center application based on Lombardi's BPM suite while Lawactive has legal applications built on Metastorm-powered workflows. SaaS-based BPMS platforms are also becoming available, such as Appian Anywhere and Fujitsu Interstage.
Although it’s unlikely that most large organizations will use a SaaS-based BPMS platform in the near future, the attractive pricing structure allows small- and midsized-businesses to take advantage of technology that might not have been previously within their financial grasp.
BPM and the Economy
Given the increasingly bleak economic projections for 2008, Gartner and other analysts are predicting a slowdown in IT spending in most categories (although they're still expecting nominal growth in IT spending overall). Bucking this trend, however, will be expenditures on BPM and related technologies, which will be significantly above average thanks to interest in running businesses more effectively and efficiently.
The predecessor technologies to BPM, such as human-facing workflow, became popular in previous economic downturns for precisely the same reason: automating tasks and reducing handoffs within business processes can reduce the headcount required to complete those processes. In fact, until 2002 when the drive for compliance struck most large organizations, improving productivity — either for the purpose of reducing headcount or increasing capacity — was the primary driver behind most BPM implementations. The past three to four years of economic growth have shifted the focus of many BPM implementations to providing greater agility and visibility into business processes, but increased efficiency is usually assumed to be underlying benefit of every deployment. As belts tighten, the interest in BPM will swing back to productivity improvement.
Sandy Kemsley is an independent systems architect specializing in business process management, Enterprise 2.0, enterprise architecture and business intelligence. Write her at [email protected]