Want a crystal ball for business? Some software comes close

So why bother? BI software lets companies find out what sorts of customers buy their products, analyze how long they have to wait on hold and how angry they are when they call customer support, calculate the cost and benefit of hiring new customer service representatives, and track the results.

It can present the information in pre-packaged reports or let people dig into it to try to uncover bottlenecks. And as with many complicated enterprise software packages, it's now becoming available online as a service that means companies don't have to install it themselves.

At least that's the promise. Earlier, large-scale software trends such as enterprise resource planning (ERP) and customer relationship management (CRM) didn't always pan out. But BI grafts more neatly into an existing business without requiring big changes, said Clint Johnson, vice president of the company's data warehouse and business intelligence work at Zions Bancorp, which has just begun using its own GreenPlum BI system to aid its eight-state network of banks.

"BI is much more strategically applied," Johnson said. "I think people are trying to solve some strategic problems but not the world's problems."

Hit, but not as hard
Gartner, which monitors spending on information technology, saw strong growth from 2007 to 2008 for business intelligence software, an increase of 22 percent to $8.8 billion. However, "we see weakened buying and hence expect growth to be in the single digits in both 2009 and 2010," said analyst Dan Sommer.

One of the big reasons for the growth is that mainstream software companies jumped aboard. In 2007, Oracle bought Hyperion, IBM bought Cognos, and SAP bought BusinessObjects.

"BI had never had a really big enterprise sales force before," said Marge Breya, general manager of intelligence platform group at SAP BusinessObjects. "Now we've got three gargantuan sales forces with massive customer bases."

Gartner's Kurt Schlegel agrees. He now believes he overestimated the degree to which BI was hit by the downturn, in part because of that sales force but also because of inquiries at Gartner and interest during its BI conferences. "The people who came had budget," Schlegel said.

Microsoft, meanwhile, has a sales pitch tuned for the economy: you probably already bought much of the software you'll need. That's because its PerformancePoint software works in conjunction with its SQL Server database and Excel spreadsheet software. "Those are products a customer typically already owns and have deployed. We look at how a combination of those technologies can create a BI solution," said Kristina Kerr, lead product manager of Microsoft business intelligence.

New directions
The economic turmoil crystallized new directions for BI. "Transparency and tighter alignment between strategy and execution are issues on top of the agenda," Sommer said, making corporate performance management a "major growth driver" within BI.

Indeed, the economy shifted directions for Zions Bancorp's BI work, Johnson said.

"A lot of our earlier priorities on BI were marketing, but not risk management," he said. "Under TARP (the federal government's Troubled Assets Relief Program), the 19 largest banks had to do stress testing. We did it also just because our investors and rating agencies were saying, 'What is the impact going to be? Do you need more capital to support your business?'"

But BI has helped the company focus on its more traditional business requirements, though. With it, the company discovered that customers who use online banking services have annual turnover more than four times less than those who don't. That affects the company's marketing efforts.

"If I'm acquiring customers, I want to invest in customers who are going to stick around for awhile and give them the services that keep them here," Johnson said. And BI should help when things pick up:

Such work pays off: Overall, Zions Bancorp expects to recoup the expense of its BI purchase within two years. "It all gets back to money," he said.

Crystal ball
In times of crisis, everybody wants a bit more security, and BI software companies say they're having some success supplying it through forecasting technology.

"Forecasting is extremely hot right now," said Tammy Kay George, SAS's global business analytics marketing.

Schlegel sees this aspect of BI as nascent. The first step is "reporting," or using BI to find out what's going on in a company. The next step is to set targets and monitor progress or failure in meeting them as time progresses. "That's where a lot of folks are," he said. So far, only "a minority is saying, 'Can we get predictive and forecast?'"

Johnson is a fan, though. "Reporting is all about water under the bridge. The next level we're trying to take the program is being able to know what's going to happen, whether it's how customers are going to behave with a certain set of products, or what the risk is with respect to loan or credit portfolios."

Forecasting is tough. BI software must typically be fed data that has a bearing on where things are headed--the classic leading economic indicators such as new housing starts. A company has to have staff who are good at forecasting--and BI "is mostly run by an IT organization that doesn't' have skill sets for econometric forecasting," Schlegel said. And then there's the issue of finding a forecasting model that's appropriate and that accounts for jarring economic changes.

"Predictive modeling can be hard when everything is changing all the time," Balliet-Antorcha said. "Models like stability."

But predictions need not just be about next quarter's orders. Breya spoke to one customer that's worried about which companies it deals with are going to be in business. "With 85 percent certainty, they have a view on which companies will be solvent and which ones won't," she said. "We're moving from inventory management to solvency management."

But if BI gets brainier, will people rely too much on a computer's conclusions? Computer-based trading algorithms famously ran amok during the 1987 "Black Monday" stock market crash. George thinks not.

"There is a certain level of concern about the machines making the decision," she said. "But they're not making the decisions solo. The machines are making recommendations, the people take actions."