The new process tracks the fluctuating prices of mortgages, municipal bonds, corporate bonds and other fixed-income assets. The process is integrated with 25 touchpoints, including 10 accounting systems and pricing data sources such as Bloomberg and JP Morgan. Once all trades are completed by 5:30 p.m., the process compiles and validates pricing information for all assets and then loads that information into the accounting systems by 9 p.m. "If there's a high variance between yesterday's price and today's, we can escalate that to a manager who can investigate," explains fixed income technology director Sayee Bapatla.

The process has two dashboards. One monitors technology systems and verifies that systems are connected and processes are being carried out at defined intervals. The second dashboard, for back-office operations staff, displays green, yellow and red condition monitors reflecting progress toward the 9 p.m. deadline.

"A red light might show that certain pricing information still isn't available and that we're entering a dangerous timeframe," Bapatla says. "Depending on what they see on the dashboard, managers can turn to predefined contingency plans based on the asset class, the pricing sources and the seriousness of the problem."

Work on the pricing process began in May, and it's expected to roll out in July. Bapatla says the next process will support risk reporting, giving investment managers KPIs and dashboards so they can quickly understand exposure to interest rate changes, political turmoil, fluctuating currencies and other factors that influence asset pricing.

The BI Connection

As analysis needs get more sophisticated, the more likely external analytics, BI systems and BI professionals will need to be involved in process planning and support. Citigroup's risk reporting process, for example, was expected to be integrated with existing analytics engines built on an internal data warehouse. (At press time, the company announced it would sell its asset management business to Legg Mason, so there might be delays.)

Blue Rhino, too, has higher ambitions for a planning process for the company's nine production centers where tanks are filled, reconditioned and refilled. The process tracks data such as employee hours, overtime hours and production per employee and per shift. Once all facilities are using the process, the company will have a finger on the pulse of every production facility, and the goal will be to drive the cost per cylinder downward. For now, available intelligence is served up through a "gigantic HTML report," says Teddy Doohan, the company's BPM developer. "We've never tracked some of these numbers before, so we're trying to figure out where we need to provide detailed analysis."

Moving from operational data to deeper business performance insight has always been a challenge, and no less so in the context of BPM. Indeed, all the products we tested provided process metrics as well as basic reporting and analytics for data such as process execution times, the number of times a process has executed and more business-minded values such as process cost. Developing business metrics and KPIs, however, will require IT development and integration work, not to mention analysis and agreement on data correlations behind the metrics. Some products embed OLAP cubes that can be exposed to third-party analytic tools and BI suites. While some BPM vendors are developing prebuilt industry- and application-specific KPIs and process metrics, don't expect them to reinvent the wheel when it comes to sophisticated, general-purpose reporting and analytics. We expect to see more partnerships and integrations with leading analytic tools and BI suites.

Let the Continuous Improvement Begin

If advanced BAM and real-time decision support are the exception rather than the rule, rarer still are real-world examples of continuous process improvement. This state of affairs is ironic given that much of the marketing hype around BPM is about "changing processes on the fly" to respond to changing business conditions.

"Businesses all too often want the perfect model before they get started," says Gartner's Hill. "Change is inevitable, so you need to continuously improve processes." If users make changes at all, they tend to do so by instinct, experience and observation rather than by formal analysis.

All BPM system vendors talk up continuous improvement, yet some of their products present technical obstacles to real-time process change, says Delphi Group analyst Nathaniel Palmer. "If the BPM system requires you to recompile code to change a process, the change is not going to happen quickly," Palmer explains. "Most systems are moving toward a model-driven environment in which you have a BPM execution layer and process definitions expressed in XML [in the BPEL standard or otherwise]. This lets you make changes on the fly because you're not changing code."

At the very least, all of the BPM products we examined provide a mechanism for deploying new versions of processes without interrupting in-flight processes. Process simulation was more of a differentiator, as only six of the nine products we looked at offered this feature. Simulation is typically used to test processes before they're deployed, but a select few can also draw real-time metrics from the process engine to support closed-loop process optimization.

Even if relatively few users appear to be performing continuous process improvement, support for closed-loop feedback is important in any BPM system selection. Perhaps your BPM journey will start with the basic but important step of integrating systems and eliminating paperwork and manual process steps. And maybe you're now preoccupied with just getting to the consistent, documented processes required by the Sarbanes-Oxley Act. Eventually, though, as competition heats up or as business requirements change, you'll start relying upon the higher-level reports, dashboards, KPIs, simulation and process change features we examine in "IT Detours on the Road to BPM."