By Joe Mullich

When business intelligence projects fail, it is often because of poor planning: In a burst of excitement, someone at the company rushes ahead with a BI initiative without considering whether it meets the organization's needs or if anybody wants to use it.

Guy Kawasaki, the renowned venture capitalist and author of the new book about influence and persuasion titled  Enchantment: The Art of Changing Hearts, Minds, and Actions.  It has some unusual suggestions for how to avoid BI system development pitfalls. For one thing, he believes a BI failure should be analyzed before the project has even been launched.

"When a product fails, there's often a post-mortem," he says. "You get in this session, and there might not even be the people from the team left, because they've gone onto a different project or company." Also, he adds, some team members are depressed and angry and blaming each other. "It's difficult to get something out of that."

Instead of a post-mortem, he advocates having a pre-mortem. This is a brain-storming session where team members imagine the project has failed and then discuss the reasons why. Such an approach is more effective than simply asking, "Does anybody see anything wrong?," because most team members will not answer this general question. As Kawasaki says, "If you're a marketing person and you think the software sucks, it takes someone with some courage-if not stupidly-to say, 'the software sucks; this is not going to fly.'"

A pre-mortem takes away the sting by coaxing out potential pitfalls-the product has too many bugs, it has lousy marketing or distribution, or the sales force lacks the sophistication to sell it-in a manner that opens up discussion and discovers ways to head off problems.

If analyzing failure in advance sounds counter-intuitive, many people argue that an economic downturn is the time to question every technique that worked in the boom years. "You can't rely on a peacetime general to fight a war," Dennis Carey, a senior partner at executive recruiting firm Korn Ferry International, told Bloomberg Businessweek in 2009, during the heart of the downturn. "The wartime CEO prepares for the worst, so that his or her company can take market share away from players who haven't."

This approach requires a new attitude of innovation and fleet-footedness. For example, while executive sponsorship is often touted as a key to BI success, Kawasaki says people can run afoul by asking for permission.

If you ask for permission from those at the vice president/CXO level, Kawasaki says, they "will always take the scary case." If you want to start a Twitter account, for instance, and you ask the vice president's permission, he will respond, "'I remember reading about somebody tweeting company secrets and getting into it with customers-so no,'" Kawasaki adds.

"My recommendation is just start tweeting," he says, stressing that asking for forgiveness-rather than permission-is the secret of his success. How does this apply to business intelligence? In building a BI program, managers often do not have the data to justify the program until they try it.

"You have to have the spirit to try; and then you have to have the spirit to fix after you try, because you will never be right" initially, Kawasaki says. "The willingness to try is 90 percent of the battle. And the second 90 percent is the willingness to fix. The math works out to 180 percent, but it's that hard."

This perspective dovetails with Corporate Executive Board (CEB) research published in Bloomberg Businessweek in 2010 that reported companies that focus on risk prioritization and response achieve a 20 percent higher revenue growth and as much as 50 percent higher earnings growth than their peers that focus only on risk identification and prevention.

"CEB found that companies that utilize this risk-management approach can often become paralyzed if they attempt to prepare for every possible reaction to every piece of new information," Bloomberg Businessweek reported.

The alternative approach, "just-start-it-and-fix-it," should not be haphazard. Another tactic, proposed by Kawasaki, is to analyze the BI project through a model he describes by the acronym "DICE":

  • D-deep. This is a great product with a lot of features.
  • I-intelligent. This, in Kawasaki's view, is a demonstration that the product understands the user's needs and pain. He points to Ford Motor Company's MyKey, which allows a person to program a car's upper speed limit into the key. If you have a Ford Shelby Mustang GT, "a bad-ass Mustang that can go to 150 miles per hour," you might hesitate to let your 17-year-old son drive the car, Kawasaki says. MyKey will limit him to a speed limit of 55, which is pure intelligence.
  • C-complete. Instead of focusing just on the analytics or software, you need to understand everything that creates the total experience, such as support, documentation, infrastructure, and value-added resellers (VARs).
  • E-elegant/empowering. The product makes people more creative and productive and better decision-makers. "The E is elegant-that someone cares about the user interface, so a mere mortal could use it," Kawasaki says.

Too often, he says, BI projects go wrong because they focus on the D aspect, enhancing features, without considering the I, C, and E elements. However, projects can fail by overemphasizing any aspect of DICE to the exclusion of the others, such as thinking that a BI system is all about pretty scorecards and dashboards (the E part).

"A data set can change a mindset," Kawasaki says, "if you enchant them with a product or service that is deep, intelligent, complete, and empowering and elegant," meeting all the DICE needs of the people who will use the BI system.

  Join the BI TV LinkedIn Group to stay up-to-date with all the new releases