Faced with pressure for more sophisticated decisions, shorter windows of competitive advantage and increased business decision complexity, organizations are looking for ways to streamline decision making. Enterprise Decision Management (EDM) combines rules-based systems with analytical models to automate decision-making capabilities. While certain products can be plugged into an organization’s existing decision-making process, others pose more of an implementation challenge.

Where business intelligence has succeeded in collecting information and delivering it to individuals, it comes up short in analyzing the information to aid in making decisions. Business intelligence focuses primarily on reporting. Core business intelligence, for instance, according to research firm IDC, looks at how often something happens.

Predictive business intelligence, however, is used to determine an event’s probable outcome, notes Henry Morris, IDC’s VP and general manager, integration, development and application strategies for the firm’s office in Framingham, Mass.

EDM is the next logical step to predictive business intelligence. But it doesn’t stop with analytics. It goes a step further, treating the decision-making step as a separate process. EDM capitalizes on business rules management systems to automate and improve decisions.

“The idea of ’enterprise decision management’ has been a lofty, and yet unfulfilled dream of BI vendors for years,” acknowledges Keith Gile, principal analyst at Forrester Research, Milford, Conn. “Much like performance management requires significant coordination between applications, data, metrics and business roles within an organization, so too does EDM require coordination.”

According to Gile, EDM falls into the category of operational BI applications, which are “highly dependent on process definitions, real-time data updates, highly granular data and the transaction level and have a short time window in which to make a decision.”

EDM is most common in automated customer-interaction decisions, such as marketing, product recommendation, up-sell/cross-sell offers, pricing, credit decisions, fraud detection and insurance underwriting. It is also used to track and analyze results once a decision has been made.

Take the example of a call center, where a call-center representative intercepts a client wanting to increase his or her line of credit. In a matter of seconds, the rep must determine the validity of granting this request based on predefined business rules and customer scoring. Rather than the rep making the decision, however, the software application makes a recommendation based on myriad variables and either grants the decision one way or the other or refers the case to a higher level of management.

EDM on the Rise

Still in its infancy, EDM is slowly rising in popularity, although research firms have not actively started tracking it yet. However, IDC reported that the core analytics business intelligence market was $7 billion in 2005 versus the predictive analytics business intelligence market of $2.3 billion. Again, predictive analytics is the precursor to EDM. So according to these figures, the predictive market is already one-third of the analytics market.

And EDM shows promise. Rules-based systems offer many advantages for automated decision making. For one thing, rules are stored separately from application and database logic, allowing the system’s decision-making logic to be changed without recompiling the application every time rules are changed or added. In addition, rules work well for simplifying complex business logic.

Another benefit of rules-based systems is that they are nontechnical and thus easy for various employees to update and maintain.

Rules-based EDM packages lend themselves to real-time decisions and offer the ability to automate 85% or more of the day-to-day business decisions typically dealt with manually. That’s according to James Taylor, VP of product marketing for enterprise decision management at analytics software vendor Fair Isaac, Minneapolis.

Real Results

Berkshire, UK, telecommunications firm O2, for example, has realized an 85% improvement over a monthly period on sales calls that have an offer made on some products. That’s thanks to EDM, notes Aly Richards, head of O2’s CRM strategy and architecture.

With more than 15 million customers and annual revenues in excess of 4£ billion, the organization faced the challenge of creating a relationship between one customer and a large corporation.

Over a two-year period, O2 adopted Chordiant Software’s Decision Management package after searching for an end-to-end solution from mining through to the action taken with a customer. “Chordiant was the only solution that provided a critical part in the middle of this that allows for multiple models and rules to be combined for immediate action with real-time scoring,” Richards points out.

The organization also considered products from SAS and E.piphany but found that at the time neither company had a viable solution for the strategy management piece. “This meant that neither option could provide an end-to-end architecture that connected the predictive models and intelligence to the real-time action without stored scores,” Richards adds.

The most challenging aspect of EDM adoption was overcoming customer turnover. With a mature wireless market in Europe (nearly 80% of adults have mobile phones, according to Tom Verna at Chordiant, speaking on behalf of O2), the only means to grow a business is to take customers away from competitors, increase the amount your customers spend with you or to buy the competition. O2 started with basic product penetration and then worked on customer satisfaction through relevancy.

The organization also learned how to integrate with incumbent architectures, warehouse and building platforms and how to get the most out of existing technology investments. In addition, O2 overcame a cultural challenge of changing the way “front line” behaves while remaining “on brand,” Richards says.

What really helped O2 in the entire EDM adoption process was defining business needs early and building an end-to-end technology architecture that allowed it to turn customer insight into front-line action in real time, Richards points out. “And we were always open to the fact that this would change the way we run our business.”

In addition to the increased sales calls-to-offers rate, O2 realized a daily conversion rate of 64.1%. “The ability to monitor trading and make immediate changes with no IT intervention has meant that we have been adjusting the rules to get to this rate,” Richards comments. Based on early results, O2 has a projected ROI of 2,500%, she adds.

Auto Club Group (ACG), a large Michigan auto insurance underwriter, has also realized the benefits of using EDM, including lowered expenses, improved processes and increased overall business results. When challenged to integrate three rate-quoting systems built across two processing systems, ACG opted to assemble a front-end Web-based quoting system that would compete for independent agents’ business.

ACG wanted to use a rules management architecture because separating the business logic from the operational decision system would allow for easy updates to the strategic decision framework, policies and exceptions. ACG enlisted Fair Isaac’s Blaze Advisor software to make this a reality.

In fewer than nine months, the organization created more than 3,000 centralized business rules to drive its decision processes. The application executes decisions based off the centralized regional systems across six states and external data sources. Blaze Advisor executes the policy quotes and publishes them in XML, providing for easy distribution and availability across a variety of quoting systems.

Blaze Advisor appealed to ACG because of its ability to create a series of rule templates.

“The power of the rules system will give us the ability to execute strategic business practices, such as processing comparative insurance rate information, and will automatically provide a personalized, competitive quote to our internal and independent agents to help us win more profitable business,” notes Michael Koscielny, director of regional underwriting operations.

Before employing EDM technology, all applications at ACG were reviewed manually. Now, with the help of Blaze Advisor, 99% are automated with 1%held back by underwriters for manual review. In addition, the organization has realized a 35% increase in applications processed.

“Looking back now, our old process seems so archaic and constricting to our business success,” Koscielny says. “Blaze Advisor propels the ACG system that precisely calculates and assigns pricing tiers based off of external data inputs, such as driving records. It then automatically executes an action, such as making a pricing offer for desirable candidates or knocking out ineligible drivers that might have multiple violations.”

Facing Challenges

These two examples represent the minority of organizations investing in EDM. Because EDM is still a relatively new concept, others are hesitant to jump on the bandwagon for a variety of reasons. For one thing, decisions are risky. So it would make sense that automating them might also be risky.

According to IDC’s Morris, some technical challenges keeping organizations from EDM are dealing with integrating all the various different pieces that go into an EDM application. “A rule has to be effective in various different applications,” he explains. “How do you translate that in a way an application can absorb? How do you build these predictive models?”

Forrester’s Gile believes any EDM solution must include rules engines, business process management solutions, BI reporting and analysis solutions, predictive solutions and even knowledge management. That’s a pretty complex application. “There may be solutions providers that claim to be able to build this type of integrated application. However, there are very few BI vendors that can address all these issues in a single product.”

In the view of Curt Hall, senior consultant at Cutter Consortium, EDM apps must combine rules-based systems with enterprise application integration, data warehousing, BI, CRM, and messaging and middleware. Hall refers to EDM as “personalization on steroids” in his Jan. 25, 2005, report entitled “Enterprise Decision Management.” And for good reason. EDM provides the ability to analyze customer interactions and determine their significance for the whole customer relationship. No wonder so many pieces of technology are necessary.

In addition to technical challenges, organizations must overcome organizational challenges in order to embrace EDM. These, IDC’s Morris contends, are more important than the technical challenges. Organizations may lack their employees’ trust and confidence in these systems. People may look at these as taking away their personal, human core expertise. “An automated rule system without any intervention is difficult for some to accept,” Morris elaborates.

The adoption process can be eased by first focusing on a decision that will have some kind of measurable impact, meaning “some significant consequences one way or another on how you make the decision,” Morris continues. Organizations need to see what kind of impact a decision will have on business, measure the results and move forward.

The highest return investments, Morris notes, are on projects having to do with a company’s core operations. “If you can do that better, more efficiently and give greater customer satisfaction, that’s an overall win,” he adds.

O2’s Richards encourages organizations considering EDM to use prototypes or proof of concepts to generate company buy-in and learn how to develop conversations. “But make sure you set goals and determine what you are going to measure,” she warns. “If you have no measurements, you don’t prove anything.”

Looking Ahead

Cutter Consortium’s Hall notes that EDM is more talked about these days than it is actually practiced. “This is because EDM requires more than just deploying a rules-based system,” he explains. “It requires an infrastructure that allows the technology to be applied across multiple channels. And this means integrating different business processes as well as integrating data from various operational systems and analytic applications.”

According to Fair Isaac’s Taylor in an April 2005 report entitled “Beyond BI: Building Intelligence into your Operational Decisions,” companies that have already invested in BI are in the best position to take advantage of EDM technology. “Because BI and EDM are complementary, automating decisions builds on your investment and realizes more of its latent value,” he explains.

”BI was built on the promise of information democracy and instant insight,“ he continues. ”Every decision maker could, from their desktop, drill down and discover the answers to any question on performance and trends.” Taylor says the same thing needs to be done with operational decisions: They need to be removed from a staff’s queue. “This improves the experience for customers who get approved more quickly, enables businesses to scale as volume grows and gives staff more time to deal with exceptions that require expert review.“

EDM can lend itself to quickening and improving even processes that cannot be fully automated, Taylor adds, “Through the instant application of relevant analytics that can guide decision makers…most businesses that begin with an EDM system gradually increase the percentage of decisions they can automate.”

BI vendors may be in the best position to take advantage of EDM technology, but it is still an elusive concept to the vast majority. “As this is still in its infancy, few vendors can even discuss the topic. It remains a services engagement and is many years away from being a ‘product,’” notes Forrester’s Gile.

Wright is a technology writer based in Phoenix, Arizona. She can be reached via e-mail at rwright@softwaremag.com.