Better Decisions

A little bit of BI knowledge can be more hazardous than helpful. Here are five not-so-obvious pitfalls to avoid.

If you've shopped at just about any big box retailer, you've likely experienced the high hopes and stinging frustration that can come with knowing up-to-the-minute inventory information. "The computer says we should have two of those," a clerk informs you. And then: "I just can't tell you where they are. And the system won't let me order more because it shows we already have them in stock."

As the customer, you head straight for the nearest competitor.

The corporate equivalent goes something like this: A customer calls to order $12,000 worth of goods. Checking the business intelligence system, the service rep sees that the order exceeds the customer's authorized credit limit of $10,000. He denies the order, and his company loses a long-standing customer -- who heads straight for the nearest competitor.

What's going on here? Putting more accurate and timely information in the hands of frontline workers is supposed to result in better and faster business decisions. But as it turns out, a little bit of knowledge can, and often does, prove to be more hazardous than helpful, experts say.

In some cases, workers are granted access to data but aren't given any helpful information about how to use it to make good business decisions. In other cases, data is distributed but decision-making authority is not, often forcing frontline workers into being little more than the bearers of bad news. All the while, companies are rapidly embracing BI as a key tool for making more and more day-to-day business decisions. Here are five not-so-obvious pitfalls to avoid.

Lost in Translation

The absence of a single set of definitions for business events throughout the enterprise is the No. 1 cause of BI breakdowns, experts say. Without a common language, confusion is inevitable, says Scott Hicar, CIO at hard disk manufacturer Maxtor Corp., which has been running an enterprise corporate data warehouse since 1998.

Maxtor's SCOTT HICAR says finding agreement on common terms is an ongoing process.
Inventory is a prime example. If a warehouse manager includes scrap and obsolete equipment in the count and the sales department doesn't, there is absolutely no way that inventory data can be accurate.

Financial information is another example. At Milpitas, Calif.-based Maxtor, the finance organization is accountable for financials, but it's the supply chain group that actually transacts most of the financial information. That makes cross-functional communication and agreement on common terms essential. Hicar says it's an ongoing process that requires vigilance, since business is continually evolving and changing. "I don't see a point where we say, 'Oh, good, that's done.'"

Darren Taylor, vice president of information access at Blue Cross and Blue Shield of Kansas City, says one of the best ways to ensure data consistency is to create a single BI team comprised of technical experts from IT and business people from various divisions or departments. Over the past year and a half, Blue Cross has combined patient, claims and membership data from three systems into a single, integrated data warehouse. This provides users with a comprehensive view of all data on an individual patient or claim.

"Before, we had 12 people in IT on the technical side building the data warehouse, and we had 20 business analysts in the business units who pulled the data. Now, we have a division of 31 people who create synergy between IT and business. Creating an organization or division dedicated to BI is essential," he says.

Stranding Front-line Workers

In marketing their BI and analytic tools as intuitive, user-friendly and easy to use, too many software vendors have planted the false impression that there is no learning curve associated with BI, says Lisa Owider, a principal at Knightsbridge Solutions LLC, a Chicago-based services firm specializing in BI implementations.

In many cases, the BI software itself may indeed be easy to navigate, yet users have no idea how it can be tapped to actually change and improve the way they do business, she says. "What we see over and over again are companies that give out all this access to data but don't tell users how they're empowered to use it," Owider says.

An example is an international retail chain whose store managers receive a profit-and-loss statement every month. Owider is training these managers how to dig deeper to resolve problems that show up on the statements. For example, if labor expenses are particularly high one month, a store manager has access to data that tells them who worked and when, how many hours they worked, how much overtime they put in, who called in sick, even what the weather was like. All of this information is in the system. Training users to find it, make comparisons, discover patterns and initiate changes is what's needed, she says.

When BI was first being deployed, the key was enabling senior executives to make better long-term business decisions, Owider says. Today, it's about empowering frontline workers to make decisions and changes around everyday events.

Drowning in Data

Giving users access to lots of data without setting concrete BI goals usually results in a lot of meandering through out-of-context information, wasting both time and money. Ideally, you should know the BI value proposition upfront, says Union Pacific Corp. CIO Jim Bell.

For example, the Omaha-based transportation company has two goals when it tracks and collects cell phone data, including information about carriers' service plans and pricing, the number of cell minutes each Union Pacific employee uses and where and when they call. The first goal is to lower overall cell phone costs. The second is to transform all employees into "intelligent buyers," says Bell.

Before turning to BI, Union Pacific put employees in charge of their own cell phone purchases, requiring them to predict how many cell phone minutes they needed, whether they needed roaming and/or nationwide service and how often they expected to travel. "These are all questions the typical business user does not know the answer to," notes Bell. As a result, a lot of costly purchases were made.

Today, Union Pacific still lets individuals and departments make their own cell phone choices. But it provides them with detailed data along with a directive to use the information to reduce costs, so they focus solely on saving money and don't waste time tracking hundreds of cell phone plans.

Bell says the company is extending the same BI goal to printing costs and vehicle purchases. The BI system tracks cost and usage information on different printers, cars and trucks. Employees make their purchasing decisions based on this information and their individual needs.

"I'm using BI to really understand what value we as a company derive from different assets," Bell says. "At the end of the day, the most expensive asset is people, and if people can be involved in lowering the costs of other, less intelligent assets, it helps preserve their jobs. That's the value proposition."

Paralyzed by Procedures

"BI has little to do with technology and far more to do with your understanding of business processes," says Hap Cluff, director of IT for the city of Norfolk, Va. Cluff learned this when his team developed a Web-based BI system that automates what had been a complicated and time-consuming process for obtaining a city building permit. Users across various departments involved in the review and approval process were given electronic access to all application documentation. The new system cut the decision-making process from 19 days to three.

All was going well until the city council mandated that a design review committee see all building plans before the city issued a permit. After that, the IT team noticed that the length of the permit decision process began to creep back up. Cluff subsequently learned that the new design review committee met only once a week and that clerks were physically driving plans to the downtown office where the board met. This added days to the process. So Cluff and his team took digital photos of building plans and incorporated them into the automated BI system, bringing permit application times back down to three days.

This example illustrates a point that IDC analyst Henry Morris says he makes repeatedly to CIOs: "Just having information is not automation. Automation is around a workflow or a certain number of steps." BI technology has to fit hand in glove with business processes, which must be reviewed and often changed.

Petty Politics

Distributing data more widely has a democratizing effect that's sure to upset the balance of power -- or perceived power -- especially in very hierarchical organizations. "It drives the level of transparency, and that can be disconcerting to some folks. Giving people more information ensures that they'll ask more questions," says Stephen Zander, vice president of enterprise business intelligence service at McKesson Corp., a health care services company in San Francisco.

But getting past internal turf battles can pay off handsomely. "If you're sitting with a customer hoping to do a $5 million deal, you might discover the customer is already in the hole to McKesson for $20 million and 120 days," Zander says. "The flip side is if you are presenting to a division of a large customer and the deal doesn't look like it's terribly much. Then you see that customer does lots of business overall with McKesson. It makes all the difference." The data that McKesson aggregates and distributes includes information about sales across multiple product categories as well as data about customer and supplier relationships.

"The goal is to reach a point with BI where people can make choices with an understanding of the impact their choice has beyond themselves," says Zander. "Typically, they're making a decision that could have consequences for the rest of McKesson, so the more information they have, the better the decision McKesson gets."

Special Report


Smarter BI

Stories in this report:

Copyright © 2005 IDG Communications, Inc.

It’s time to break the ChatGPT habit
Shop Tech Products at Amazon