Sandy is my fellow blogger- she blogs at column2 but I met here when we both blogged over on ebizQ (my blog is here). She presented on “Business Process Management with Business Rules and Business Intelligence” and started by pointing out that she has long felt that business rules are essential for business process management and that she continues to be surprised how few BPM users seem to get this.
Enterprise business processes are key – things like order to cash – a chain of activities producing a business result. Usually spans functional silo and is required to manage the white space between these functional groups. BPM is a management discipline for improving cross-functional business processes. She discussed Gartner’s view of a BPMS and pointed out that this view includes Business Rules Management. Sandy feels that best of breed might well work better than trying to embed this in your BPMS.
Sandy started by discussing some basics of BPM. The process lifecycle is model->automate->monitor->optimize->(repeat). Increasingly these are better integrated and automated in a modern BPMS. Not always how it really happens but that’s the basic idea. The goals of BPM used to be primarily efficiency (fewer people, less handoffs, integrating systems). Compliance was the next big driver (after Enron, Worldcom et al) with a focus on standardization of processes. Now the key areas of focus are agility and visibility. Trend was from efficiency to compliance to a focus on competitiveness.
Business rules have a key role to play in BPM. The use of business rules to handle routing decisions, work assignment etc drives agility. Only then can you change rules without changing the process. This is the key reason Sandy prefers separated business rules as it allows for agility without requiring a change to the process map. If you only use a BPMS then this kind of change will often force an unnecessary change to the process map. Like me she feels that the rules in a decision will likely change much more often than the overall process map. Also critical to applying rules consistently across multiple processes and other applications.
Sandy’s experience is that most BPM customers are using the built-in rules functions of their BPMS and as a result have only fairly simple capabilities. They also risk having their rules bound into a specific process step (making reuse hard) and a particular process instance (preventing in-flight change). This is a particular issue for long-running processes and is something I debated with Jesper of BEA in this series of posts – 1, 2, 3, 4, 5
. Sandy also pointed out that a change to a process is potentially high risk so even a change to an embedded rule must be tested and deployed very carefully. In contrast, an externalized rule approach allows those rules to be changed in a more controlled, more localized way. This in turn makes it practical for the business to own these rule changes. Changes are shared across multiple process and non-process consumers and can be changed in a long running process “in-flight”. Sandy pointed out that the major BPMS vendors increasingly offer integration with the top BRMS vendors as well as offering loosely-coupled access to web services allowing that to be used to link any BRMS that could deploy in this way (which is almost all of them).
Sandy went on to talk about business intelligence and BPM. She has a fairly broad view of BI – everything from reporting to data mining to predictive analytics and real-time dashboards. BI in this sense allows visibility into the performance of a process and can contribute to process agility. Visibility comes from monitoring and reporting, analysis and optimization. Agility can be generated where the insight triggers different processes or steps based on prior decisions – more event-based decision-making or complex event processing (CEP). It can also identify that a particular transaction should route to a manual decision rather than an automated. It can also inform the person making a decision as to what decisions were made in the past and how that worked.
As with business rules, some BI capability is embedded in BPMS. However this is pretty limited in particular in terms of scope – only process data can be managed. More sophisticated customers, which is most, use a full BI tool by sending the BPM statistics into their main data/DW/BI environment. This allows both more reporting and integration of multiple data sources beyond the BPMS etc. Often not that much more expensive as most companies have a BI platform already.
BPM+Business Rules+Business Intelligence can deliver self-learning processes using results to modify rules and using data to improve or even replace manual decision making. Ultimately you can generate true decision management from rules and BI. BPM and rules turns policy into action, BPM+BI provides insight into your processes, BPM+BR+BI gives you decision management and self-optimizing processes. Very cool.
Check out Sandy’s blog at column2.com