Right (on) time

Optimize business operations by delivering
timely BI throughout the organization.

For most companies, business intelligence (BI) is often in the hands of too few people and unavailable when it is needed to maximize business performance. Opportunity certainly exists to improve the timeliness of BI to help improve effectiveness. Imagine the impact on an organization’s bottom line if all employees would “row in the same direction” with every action they take to optimize costs and revenues.

Such a holistic vision goes beyond just giving people access to BI. By exploiting timely delivery of BI in the context of every business activity, a company can transition to implementing right-time business optimization. This is attained when people and systems continuously know the best action to take to keep a business running smoothly, with minimum risk and maximum profit.

Implementing this vision requires coordination of objectives and decision making across all levels of a company. Many organizations use BI and corporate performance management (CPM) to operate at the strategic and middle-management levels. However, BI is rarely available to support decision making at operational levels, which is where the bulk of employees are found. Yet operational employees (call center operators, counter staff, repair workers, etc.) are expected to somehow make the right decisions and be aware of events occurring around them. Given the thousands of small decisions made every day, a chasm exists between users of advanced analytic data and the front-line workers who need to make smart choices.

Constructing the proper framework

To solve these problems, contextual analytic insights need to be continuously delivered to all enterprise employees so they can optimize costs and revenues. Several popular IT initiatives must converge to make this vision of right-time business optimization happen. They include enterprise data governance, BI, CPM, business process management (BPM) and service-oriented architecture (SOA). Many organizations have already started these kinds of IT projects. However, they are often implemented in a stand-alone fashion, with insufficient understanding of their combined business benefit.

If data governance is the foundation that delivers data we can trust, then a CPM framework must be built upon that foundation to create the structure needed for business optimization. (See figure 1.) Ideally, strategic objectives from the top cascade down the corporate hierarchy, expanding into more specific and relevant objectives at each lower level until all employees know where they fit into the framework. While aligned objectives cascade down, measurements bubble up so managers can take action.

Using CPM software that leverages scorecards and dashboards, individual employees observe how they contribute to profitability. Several CPM products already have this multi-level strategy management capability, but too few companies exploit it. To be effective, however, CPM software must go beyond being integrated with just BI subsystems. To play a role in business optimization, CPM must also be integrated with event processing and business processes.


Click to enlarge

On-demand delivery

To take the next step, BI must be available on demand through SOA to processes, portals, desktop office suites and applications, as well as within CPM systems. By making BI available through Web services using SOA, critical analytic information can be delivered at the right time in the context of any process task being performed. By combining it with portal software and BPM, role-based analytic information delivery becomes possible when the user performs each process activity. BI to the masses finally becomes attainable, not through reporting tools but through SOA delivery of analytic information to every portal and Web site.

Once analytic information is available as a Web service, it is fairly easy to connect that service into multiple business processes. Good examples of analytic insights delivered this way include “next best offer” campaigns or dynamic price quotes. Whether in a portal, Web site or long-running workflow, BI Web services deliver insights on demand from facts in the data warehouse. Using BPM tools, it becomes possible to map BI Web services into specific operational process activities. These process orchestrations can then be executed to control how an enterprise uses BI during business operations. BPM tools do much more than orchestrate the execution of Web services, however. Their key use is to separate processes from underlying services to provide corporate agility on a scale never before seen.

Process-oriented BI in action

As an example of process-oriented business intelligence (BI), consider a corporate property insurance underwriter who has to make daily decisions on whether to underwrite agent quote requests for corporate and industrial properties. Similar decisions must be made on renewing existing policies. Many factors influence these decisions—fire risks, property features, previous claims, industrial hazards, similar property claims, etc. Using on-demand analytic insights during quote and renewal underwriting activities allows guided decisions to be made on writing policies and giving competitive quotes.

When the insurance company expands into the mid-market, it receives a substantial surge in property quote requests from new brokers and prospects. The insurer cannot afford to hire hundreds of underwriters to handle these requests. Therefore, the underwriting decision process must be automated to handle the volume of in-bound quote requests. In this case, the company could deploy rating and pricing software that leverages underwriting expertise by using rules engine software and analytic data. The software could price and approve many requests, reject some as too risky, and defer others to the actuarial staff. The same rating engine could guide an underwriter to the correct premium price to minimize risk or to recommend re-insurance if the risk level is uncomfortable.

—M.F. and D.G.

Management by exception

Right-time business optimization must also support management by exception. This is achieved using event processing—the automated monitoring of business events to enable responses to problems and opportunities as they occur in real time. In event processing, analytics are triggered to analyze an event or correlate collections of events. If the event is significant, alerts may be sent to employees, or a software process may be invoked to handle it. One example is the continuous monitoring of point-of-sale (POS) terminals against inventory in retail stores. This can be used to automatically optimize and align supply chains with actual demand. Another example is continuous monitoring of loan and mortgage payments so that if payments are missed, credit card limits for the delinquent customer can be immediately closed off. Note that both examples require real-time analytics to make the correct choices.

Scalable event processing requires data to be analyzed before it reaches any BI system—analyzing “data in motion” before it becomes “data at rest.” This emphasizes the importance of monitoring streaming data and holding data in memory while it is analyzed. Streaming data analysis needs access to the CPM performance indicators to understand events in the context of profit and loss, as well as provide drill-down into the details. This kind of dynamic business management is called business activity monitoring (BAM) or complex event processing (CEP).

"Contextual analytic insights need to be continuously delivered to all employees so they can optimize costs and revenues."

The big picture

Within the CPM framework, if workers have a role-based scorecard in the employee portal, they can manage their responsibilities in the context of company strategy and objectives. (See figure 2.) Each worker has personal performance metrics calculated by the CPM software. This is basic to CPM subsystems—that is, until you add real-time event alerts and recommendations into the employee’s scorecard. Problems and opportunities in the employee’s domain that have to be addressed immediately appear side by side with performance metrics and analytic recommendations. Individual workers can finally see “the big picture”—at least within their roles.


Click to enlarge

Taking this one step further, by integrating BPM and event processing into the CPM scorecards through SOA, you can measure processes versus outcomes. With CPM the objectives and measures are understood. By associating individual steps with scorecard results, you can observe how decisions affect the “red light, green light” status of an objective’s fulfillment. By capturing the recommended decisions and the actual choices made, and knowing the thread of execution that led there, a business process analyst can determine whether the process needs to be changed. Using BPM simulation tools, an analyst can then test changes before implementing them. To do this, BI, event processing and BPM subsystems must be integrated with CPM metrics to deliver role-based, right-time business optimization at all levels of the enterprise.

Some assembly required

The ingredients needed to implement right-time business optimization are installed in many IT organizations. Indeed, many parts are already in production—but not all together. And, of course, who sets out to build a vision? Instead, it’s better to focus on projects that provide incremental business value.

Right-time business optimization provides a snapshot of how several major trends and architectures will converge into a widely accepted design pattern. But as one pundit put it, “The future has a way of arriving unannounced.” Some smart IT architects and business process experts are already implementing right-time business optimization. Hopefully, they are not your competitors.

Your Comment:
Your Rating:


9/8/2009 1:46:16 PM
— Anonymous
Fuzzy Logix