Intelligent Enterprise

Data, yes: But no organization can ever have enough intelligence. Analysis and discovery are part of the endless quest to find nuggets of information that, when put in the right hands at the right time, fuel decisions that raise business performance, build customer loyalty and create sustained competitive advantage. Commoditization of business intelligence (BI) tools is making the technology more affordable and spreading it to more users. But the act of BI must never become commoditized. Average effort can produce average results. The urgency to find — and use — more and better information should never cease.

To lead off our sixth annual Business Intelligence special issue, we identify six "megatrends." Our selections draw on research and discussions with user organizations wrestling with how to bind business and IT objectives into a sustained quest for higher intelligence.

Cindi Howson's companion feature delves into the megatrend driving BI software itself: improving ease of use so that "BI for the masses" becomes a reality. Using BI to open up data sources to meet the needs of front-line workers raises the business value of data warehouses and other sources, including, soon, content repositories. As you'll see, our megatrends testify to the potential of BI: and also the obstacles that must be overcome.

Break Down Silo Walls

Not long ago, centralized computers held nearly all the data available for business analysis. Users had to beseech IT on bended knee to provide access, usually in the form of voluminous batch reports. With the spread of distributed systems and personal computing, the spreadsheet — perhaps still the greatest of all killer apps — gave users direct access, although data quality chaos ensued.

Personal spreadsheets and departmental BI data marts and tools reinforce what have become known as "silos": applications and data sources beholden to specific business functions or departments and developed with little regard for integration. A major draw of enterprise BI is to apply technology to reorient business performance and execution around multiple departments and functions.

"Banks are fighting to get out of the silos so they can package mortgage, deposit, investment and other products and services," says Tony LoFrumento, Morgan Stanley executive director of BI and CRM. "In the silo world, the mortgage group might say, 'I'm not giving up anything that might hurt my profitability.' What banks need are segment managers who view their clients holistically and look across products to say, 'Hey, give up a percentage point on the mortgage because we'll make it back in the investment relationship.'"

LoFrumento focuses on "rapid development and deployment of BI wherever needed in the organization, from sales assistants and financial advisers on up to the president." Everyone works off the same, qualified data. LoFrumento sees this as critical to breaking down silos and supporting organizational change. "Instead of defending and fencing over their data, managers can immediately deal with situations," he says.

Without a customer-centric focus that relies on integrated data, CRM initiatives are flying blind. "Clients know exactly what they have at your firm," he says. "But if they have to interact with you silo by silo, you aren't treating them well. Competitors suddenly know your clients better than you do. You lose your top clients and are left with the unprofitable ones."

Using Business Objects tools for BI and OLAP, SAS for data mining and other vendors' tools, Morgan Stanley decision makers see customer relationships holistically — and discover that conventional wisdom can be wrong.

Trust, But Verify: Users Demand Data Quality

Regulatory demands have raised the level of data quality awareness among business users. BI scorecards and dashboards are worse than ineffective if they deliver bad data; they can break the hard-won trust established with business users by BI and data warehousing teams. Thus, BI's success is tied to the integrity of the data. With corporate reputations on the line, business users are now supporting data quality initiatives.

Most organizations rely on multiple data sources to gain a complete understanding of customers, processes, risk factors and performance. "Our biggest challenges stem from the fact that business administration depends on data submitted to us by client companies," says Thomas M. Freitas, Transamerica Reinsurance's senior vice president of technology and operations (see "Debriefing," right). "Because there's a lack of applied industry standards, we receive data in multiple formats. Data fields and terminology differ from one company to the next. So we spend a great deal of time and resources on data quality."

With millions of individual policies on its books, Transamerica performs multiple millions of policy-level transactions, Freitas says. Transamerica's BI infrastructure, which includes software from Business Objects, Microsoft and Oracle, fills Balanced Scorecards that help managers evaluate how well the company processes client data. Flash reports reveal the financial impact of policy and premium changes. The company's business is to assume risk from life insurance companies in exchange for part of the premiums. For BI to put the odds in Transamerica's favor, poor data must be vanquished.

Operational BI Focuses on Problem-Solving

In every enterprise, operations and processes are riddled with bottlenecks, which thwart the best-laid plans and degrade business performance. "Operational" BI, which focuses on supplying information to managers and employees for day-to-day decisions, will make a difference if the technology helps them ease chokepoints. BI software suites must stretch to integrate all activities and tools that users employ to zap bottlenecks, including project management software.

It's all well and good to supply senior-level managers with dashboards that indicate the degree of alignment between strategic or financial goals and execution. What's more valuable to operations are metrics that "zero in on the bottlenecks and help front-line employees understand the value of eliminating them," says Jonathan Rothman, director of data management at Emergency Medical Associates, a physician-owned services organization that manages 18 emergency departments in New Jersey and New York. "Before we created dashboards and metrics, we went to these departments, and to our billing and finance operations, to learn exactly what they do and where they're hitting snags."

When a metric isn't performing to the threshold, you want to do something. "The problem is that you must go to another product, usually Microsoft Project," Rothman says. Given BI tools' inability to wean users off Excel, it's unlikely that operational BI extensions for performance management will displace Project or other popular products. "But they don't have to," Rothman says. "I want an integrated approach so that performance management doesn't end with the BI reporting of the metric."

Rothman likes what Business Objects is developing in an upcoming release to integrate with project management. "I can write a report that says metric A performed at X, and then below that graphic I can see that task A started here and task B picked up there. And I can see that when task B finished, I got a little bump in performance, which is shown by the metric. That's a product I can really use."