In-Depth

Escaping BI's Groundhog Day

Escaping BI's Groundhog Day may take new vision and new interpersonal strategy.

You may have heard this conversation before. Two eminent BI leaders ran into each other at the Sunday night reception at the TDWI World Conference in San Diego.

"Business Objects and Cognos just don't get it," one said, by which he meant all the BI vendors that became part of bigger companies last year. As these companies get dragged further into the new parent, he said, they'll "get it" less and less.

The other added, "They don't get collaboration, and they don't get visualization. The big vendors just can't innovate fast enough for the market, and the next generation of BI will come from a new generation of vendors."

Over the next few days, I heard clues for how BI might find its way out of its post-consolidation malaise.

One vision came from one of the industry's most refreshing thinkers. At the Executive Summit, Mark Madsen of Third Nature presented "Clues to the Future of Business Intelligence." According to Mark, "Expertise blinds us." We lose sight of opportunities, perceive the unfamiliar as risky, and assume that the future is linear. When we finally see the future "barreling down on us," it has actually been on our doorstep a long time -- creeping forward like a snail.

In fact, he said, BI's "snail" has been in front of us for about five years. Although business has been shifting to more decentralized processes and decision making, he said, business intelligence practices still rely on old assumptions of hierarchy and control.

Good BI, said Madsen, needs three pieces: access, delivery, and social features. "We're zero for three." In fact, the iPhone, the Mac, Google, and Web 2.0 sites such as YouTube, LinkedIn, and others have shown users that they're much better at all three of these pieces than BI vendors themselves. YouTube's user-supplied content, for example, is sifted with a system of comments and ratings. IBM's Many Eyes, an experiment in visualized data, is "dead simple" and can be learned in a few minutes, unlike some training for BI tools. "Users expect a superior experience," he said.

Some developers fear that user-friendly interfaces cost more to develop. Madsen disproved it, at least in one example. He compared a hard-to-use map of Oakland crime -- developed by the city for more than $300,000 -- with an intuitive, almost Google-like map created in a developer's spare time. (The city reacted by shutting off his data access.)

A similar plea for simplicity came from Bill Schmarzo, Yahoo! vice president of advertiser analytics. The iPod is "relevance perfected," he said in his Thursday keynote, and "relevance is everything."

The iPod's success derives from its simplicity -- giving users just what they need and no more. Its designers worked hard to keep the irrelevant off the interface -- unlike dashboard makers whose confusing and ineffective interfaces boast "infinitely flexible and infinitely powerful." That's "infinitely confusing," he said, and just lazy design. Instead, he said, start with recommendations instead of forcing users to "crawl through charts."

Sid Adelman, co-author of Data Warehouse Project Management (2000, Addison-Wesley) and longtime TDWI instructor said, "The tools are fine. … What we've done is hand a Stradivarius to a kid and then wonder why it doesn't sound so good."

A player violin might help, but for BI the problem is not the instruments but the orchestra. "The organizational and cultural issues will screw us every time," Adelman warned.

Sometimes there's a solution. Adelman recalls a bank manager with a staff of five managers who refused to cooperate with each other until their bonuses were linked. No one got a bonus unless performance for all five was good. Suddenly, those five and their staff found harmony.

The bank manager's solution worked for a different reason, however, than most people think. "People think they can't integrate because they don't like each other. Actually, it's the opposite: they don't like each other because they haven't integrated," explains Lorna Rickard. She and Maureen Clarry teach "Power, Politics, and Partnership in Business Intelligence Projects" at TDWI conferences.

In the course, role-playing students discover that roles determine behavior. That is, executives, middle-managers, entry-level workers, and customers all react in predictable ways to common problems -- most often ineffectively.

Clarry and Rickard urge students to try a new approach. For the "middles," with the most complex role, the best strategy is allying with peers. They're urged to ask themselves, "Who can help from my level?"

People in any role, say Clarry and Rickard, can learn to avoid their own version of the movie "Groundhog Day," in which Bill Murray's character is caught in a forever-recurring day.

I'm tired of the groundhog; I believe Mark Madsen and Bill Schmarzo are, too. They have earned release from the BI industry's forever-recurring conversation. By spreading their own visions, they invite alliance and they might even stimulate vendors.

Meanwhile, enterprises are certainly able to make their own adjustments. It really may be possible with more effective interpersonal responses to break out and get what they need for themselves and their organizations.

Must Read Articles