Adding Agility to BPM with BAM

A business process is a series of connected events that begins with an input and ends with an output. Your business process management (BPM) application uses roles, resources, business rules and more to achieve increases in process speed, efficiency and effectiveness. But these results alone may not be enough. Your customers also want consistency in the output; adaptability to changes; and always “faster, better, cheaper.” To achieve maximum return to the business, don’t settle for a “black box” process that is concerned only with inputs and outputs. To achieve consistency in performance, you need to establish standards that are continually measured and reported in a timely manner. Your process must be adaptable in both its execution and its design as requirements change. The process should build in and facilitate a capability for continuous improvement.

The key to achieving these additional goals is process visibility. If the process is visible, it can be measured. And, as Peter Drucker has stated, if it can be measured, it can be managed. In this case, it means management both from an execution perspective and a design perspective. Business activity monitoring (BAM) provides the ability to do this, thereby maximizing the value of your BPM/case management investment.

Visibility alone, however, won’t achieve the optimum performance from a BPM/case management application. Visibility must be matched by agility—that is, the ability to rapidly understand changing circumstances and initiate appropriate responses.

What is BAM?
As a process or case is executed, it initiates activities or is acted upon. The BPM engine orchestrating this process reads the instance metadata, recognizes roles and resources, invokes and responds to business rules, creating a stream of data that parallels and describes the process. Using the model in the process design template, BAM organizes the resulting data into a flow of automated activities and human intervention tasks, with their associated data and documents. The information captured describes changes to the application’s business state in the following ways:

• What is the change that occurred;
• When did the change occur;
• Why was the change triggered;
• Who is responsible for the change;
• Which business objects were affected; and
• Where is the process going next?

Because the process data is streamed on a continuous basis as it occurs, BAM processes and presents this data in time slices that are near real time in terms of latency. Capturing the process data in this way enables considerable flexibility and provides visibility and insight into the process, both as it occurs (individual instances or cases) and as aggregated trends and variances over time—variances that provide opportunities for improvement in business operations.

Real-time Tracking of Process Execution
Because BAM provides a repository of near real-time information, it can report on process performance in this same near real-time context, allowing the business to tweak, intervene or remediate exceptions surfaced by BAM reports. The volume of data in an organization with tens of thousands of process transactions per day would, in the past, have precluded such responsiveness. BAM, by using business constructs such as key performance indicators (KPIs), service level agreements (SLAs) and graphical visual management dashboards, can manage according to exception by highlighting the areas in need of attention and trigger remediation to bottlenecks and other error conditions. When a business rule flags an exception, a notification alert is sent to the process owner and departmental supervisors while there is still time to correct or repair the situation. If human intervention is preferred, the alert can detail the error conditions in an activity or a process and recommend a path of resolution. The preprocessed, hierarchical nature of BAM reports allows a manager to drill down through aggregated reports to identify root causes of process issues or failures.

Example: Real-time Tracking
In a case management application using EMC Documentum xCelerated Composition Platform (xCP) in a county court system, the court clerk observes in an operational dashboard that the average case duration for the past week is now 32 days, which is in the critical zone. The clerk drills down into the average duration of the component activities and discovers that all are within normal limits except scheduling. Drilling down farther into the scheduling activity shows that the number of new cases to be scheduled each week has increased steadily for the past month. The clerk creates an additional scheduling queue and re-allocates personnel accordingly.

In contrast to business intelligence (BI) and other reporting tools, the visibility of BAM promotes agility in responding to changing circumstances by identifying problems as they are occurring in an interactive format that enables rapid intervention. For case management in particular, where the process route or duration may not be known in advance, BAM visibility can be critical to accountability and adherence to SLAs. BAM events can even trigger alternate process routes, in effect making the process self-healing in those circumstances.

Typical exceptions reported in BAM include:

• SLA. A loan-origination process exceeded the maximum duration; the due date occurs in three days;
• Comparative trends. A loan amount exceeds the monthly average;
• Workload balancing. The processing queue is at 70% capacity;
• Performance management. A task was not assigned for processing, exceeding the KPI duration;
• Automated corrective measures. If the status is “RED,” trigger a review task for the supervisor; and
• Cross-functional. More than five exceptions have been detected in department A over the past hour.

Improving process design
True process management requires not just maintenance of the process—that is, intervention and remediation—but also continuous process improvement. BAM can contribute to this effort in a number of ways:

• Root-cause analysis. The continuous improvement methodology uses the “five whys” concept (asking “why” after each answer to a question) to uncover root causes. Because BAM data is collected at a granular level, and then progressively aggregated according to a hierarchical model of the process, the reverse process of drilling down through the hierarchy can serve to identify and isolate root cause;
• Trend analysis. BAM can aggregate and store data by standardized time intervals. Displaying performance graphs over time can reveal trends;
• Process variation. Continuous improvement requires that first a process must be stable—that is, its variation must remain within predefined limits (e.g., Six Sigma). BAM can monitor and display this data, called a run chart, to provide this basis information in near real time;
• Continuous process optimization. When a process is initially designed, certain assumptions are undoubtedly made regarding workload, resources, constraints, etc. With luck, the assumptions are correct, but more often than not, the design process becomes iterative, reflecting changing business needs and deepening understanding of business activities. When BAM is in place, BAM historical data can replace previous assumptions, allowing an analyst to apply validated data in order to optimize performance or adapt to environmental changes. In this manner, not only is process execution more agile, but process redesign becomes faster and more precise; and
• Segmentation. Segmentation refers to analysis by case type, resources consumed, provider type, etc. The ability to slice and dice by process variable can provide valuable insights into opportunities for improvement.

Example: Continuous Process Optimization
A contract management process was designed with the assumption that only 20% of the contracts processed would have a value that required CFO approval. Subsequent Documentum xCP BAM reports revealed a bottleneck at the CFO, because 90% of the contracts were routed this way. The BAM reports further revealed that all the contracts were approved at that level, because, at that volume, the CFO actually relied on the subordinates’ prior approval. The step was eliminated, thereby streamlining the process at no risk to the organization.