In-Depth

Analysis: New BI Strategies Entail Collaboration, Pitfalls

Lean times may drive business and IT together for new speed and new risk.

• By Ted Cuzzillo
• 09/17/2008

Talk is cheap, and it's a good way to squeeze more BI out of what you have. Several BI leaders I talked to recently recommend it, though they also see risks.

"One of the cheapest things available to BI people is talking to business people," says one data architect at a large IT vendor.

Just sit down and see if you can come up with something valuable that's outside of conventional BI wisdom, he suggests. If you've already done the data integration, now's the time to put data to use in ways you might not have had time for before. "Maybe you can connect the data in some new ways. You can have fun with it now," he says.

Slow periods are a chance for IT and business to make heroes out of each other. Pull out ideas you've been saving, look around for data -- either easily acquired from outside or from unused heaps inside -- and see what you can do.

"Maybe there are some nuggets lying around in someone's heap of unused data," he says. You may find good data or what would be good if you could solve data quality problems. "Either way, you're ahead."

A slightly riskier method that also calls for collaboration is on the rise: agile BI. Dave Wells, now an independent consultant after five and a half years as TDWI education director, says that though it didn't develop out of a need for lower costs, lean times might have hastened its use.

"It's about prototyping your way to a solution," he says. It starts with someone from IT and someone from business teaming up. "Let's just see what we can whip up in an hour or two or three," he says. If the prototype doesn't give the most useful answer, redo it.

Forced collaboration is agile's most refreshing aspect, says Larissa T. Moss, co-author of Business Intelligence Roadmap (2003; Addison-Wesley) and several other books. "There's no better method to bridge [IT and business]. It's built on constant feedback."

Not all business people are so open to collaboration, she says, but those who've tried it wouldn't do it any other way. "They want to be involved." There's a transition in the business community about how much they want to be involved in developing BI solutions. "It's really a business solution," she says. "They should play a role."

Some may wish the business people stayed out. Agile's speed threatens to undermine data warehousing teams as they trudge forward with their slow, unglamorous work. In one large organization Moss is familiar with, impatient business users have gone around them to develop agile-BI applications. That outnumbered and slower-moving team is falling behind and becoming alienated. Thought the company is committed to the data warehouse, Moss worries that the data warehouse project may be cut during future cost-cutting.

Perhaps agile data warehousing will replace the traditional method. Like agile BI, agile DW also requires collaboration -- but entirely on the business side. In an integrated environment, business people have to negotiate with each other over data collision, priorities, standards and other things. Moss says, "It's no longer IT asking, 'What do you want, Charlie?' It's 'Does Charlie agree with Tom, and does Tom agree with Mary?' and so on'" as data sources collide.

Eventually, even agile applications will need data integration and all the well-known disciplines – such as data governance, data stewardship, and master data management, says Moss. It will have to come together at some point, she says. "Master data management might bite them in the butt."

By then, we hope, everybody will have learned to talk.