Linking DCM and analytics for maximum impact

Nathaniel Palmer describes himself as being “part of the community that has long promoted adaptive case management.”

That approach emphasizes getting decision-support help for people from a system--an adaptable system, says Palmer, who is executive director of the Workflow Management Coalition, a global industry group focusing on workflow, BPM and related issues.

“It’s the idea of getting guidance from a system to help you pick a course and having a feedback loop and a rule engine with situational or context awareness,” Palmer says. That implies that users can get not only information about a task, but also access to policies telling them what they can and can’t do.

Of course, getting an case management system to provide all that requires considerable thought and planning. There’s no simple plug-and-play solution for delivering that level of case management capability. So setting out on the path to adaptive case management (or, as it's also known, dynamic case management (DCM) means being ready to explore and engineer a solution in a step-by-step way.

CONSIDERATIONS AND COMPONENTS
Analytics allows for presentation and ongoing analysis of the information associated with a case. “The combination helps you make decisions, but it is making decisions at a data level that humans can’t always do,” Palmer says. That includes correlating data that a computer can then crunch faster and more easily than a human, or recognizing when to invoke rules or provide reminders rather than simply leaving it to a person to remember a policy.

From an analytics standpoint, one best practice is taking a case-driven development approach to linking analytics and case management, says Palmer. “That means starting with a test case, then beginning development in tight increments or intervals where you are figuring out whether your development meets your test case,” he says.

That step can be helped by applying test-driven development (TDD). "This approach means that, at an early stage, you can figure out what the various parameters are that you need in order to apply analytics and whether you have designed the underlying rules in a way that is consistent with the answers that are required,” he says.

Building shared case management and analytic functionality isn’t so much an integration task as development of a suite of abilities in which analytics are a critical component, Palmer continues. “I am also suggesting that analytics must be context-aware so you can apply it to what is in your case system,” he says.

DCM VS. BPM: SOME IMPORTANT DIFFERENCES
Feedback capability is a key component. Palmer cites the example of one of his clients, a federal government agency, which needed to conform to both external and internal rules. The external rules covered matters such as paperwork-filing deadlines. The internal rules involved issues as whether someone involved in a case was a minor or a non-citizen, factors that could involve invoking another policy. “So, in this example, you are applying these business rules and getting feedback,” he says.

The fundamental difference between using a case-management system (such as DCM) and a process-management system (like BPM) involves the payload, Palmer says. With BPM, what matters is the state and rules being applied, but it’s otherwise agnostic, other than regarding what is relevant to human users. The payload is typically the data necessary to provide the sequencing and control flow that defines the process.

On the other hand, case management isn’t bound by a predefined process, but by the policies and content inside the case and the user’s visibility into that. “It is being able to say what your state is. Having that is what the analytics provides,” he says. “The human user looks at all the files and says, ‘This is what is happening now.’”

MAINTAINING FLEXIBILITY
Palmer is often surprised to find that those implementing DCM choose to tie the interface to the underlying information source. That choice can hobble the goal of being dynamic and adaptable. “You want your interaction environment, where the user is living and working, to be independent of the sources you are inputting to or pulling from,” he recommends. “This is a basic loose-coupling principle. A lot of systems don’t allow that; they are very closely tied to a file system of some kind, whether data- or document-oriented.”

Failure in that regard can reduce the flexibility you need. He recommends aiming for analysis to be based holistically on what’s involved in the case, rather than on some predetermined location of a file. “You want it all to be part of a package,” he adds.

But that word “package” need not be taken too literally, in the view of Keith D. Swenson, co-author of "Mastering the Unpredictable: How Adaptive Case Management Will Revolutionize the Way That Knowledge Workers Get Things Done” (Meghan-Kiffer Press, 2010), and vice president of research and development at Fujitsu America. He says the package needs to be something you assemble yourself. “I don’t believe in the closed system. The important questions are about how your organization performs, and it is rather unrealistic to believe your organization will only use one technology,” he says.

Instead, he suggests harnessing the capabilities of existing applications as much as possible. The key ingredient will be some kind of standard to exchange log/event data. Swenson knows of two: Business Process Analytic Format (BPAF), an XML format from the Workflow Management Coalition initiative, and eXtensible Event Stream, better known as XES. Both involve looking at events and recording them, which then allows for analysis.

Finally, keep in mind that DCM and analytics can make everyone more aware of what’s going on—and the combination may reveal some unpleasant truths. “The really scary thing is that case management makes failures more visible,” Swenson says. “The bad news is, if you are in an organization that blames people for failure, people won’t want to use case management.” In such instances, successfully implementing case management may require a cultural makeover, he says: “You may need to change the culture completely toward being more open.”

READER FEEDBACK: As this feature notes, injecting analytics into case management can increase visibility into many activities--which, in turn, can generate user anxiety. Have you run into such cultural issues? ebizQ's staff would love to hear about how you addressed them. Please e-mail Site Editor Anne Stuart at editor@ebizq.net.



About the Author

Alan Earls, a journalist who specializes in writing about technology and business, is based in the Boston area.

More by Alan Earls, ebizQ Contributor