Why is BI's penetration rate still no more than 25% of its potential, especially in numbers of users?
The biggest thing holding BI back is fragmentation, which makes consolidation look like a good idea. But if you do that just to save money, you're going to miss the boat. The trick is to standardize, consolidate and then deliver things that are meaningful to everyone in the organization.
That's why I've been talking about establishing competency centers outside of IT. Organizations must get people who understand what "best practices" are to become gatekeepers — and give them hooks into purchasing. The best BI initiatives have key users involved up their eyeballs.
IT management often inhibits greater BI tool use to keep license costs down and protect performance from runaway queries.
Yes, but organizations should want people asking questions and getting answers. The alternative — keeping them uniformed — is much worse. Besides, the software industry is changing. Standard procedure used to be paying up front for more functionality than you could possibly consume. Now the trend is toward role-based functionality.
The key is committing to an information-centric culture that breaks through fears of database meltdowns and ideas that BI is too hard or costly.
Why do you call performance management "BI with a purpose"?
Performance management is the new BI. Before Gartner, I was with Digital Equipment Corp. I had a counterpart on the consulting side. We were peers, but we had different systems and numbers. The night before our meetings with management, we'd get together to reconcile stuff because the managers would have shot us if we showed up with different views of the same business.
You want to get as many eyes on good, consistent information as you can so that when people show up at those types of meetings, everybody has the same perspective. Actionable insight is necessary for aligning BI processes with business purposes.
What do you see as a technology game-changer in BI and performance management?
If you throw 10,000 simultaneous BI users at a relational database, it will melt down. One of the ways you solve that problem is by delivering insight. I know: We went through this whole "push" technology debate a million years ago. What bothered people then was the mindlessness of it. You could only alert yourself to things you were expecting to hear about. The applications were silly and superfluous.
But push is going to change with the implementation of intelligent agents. Adaptive agents can examine what's going on. They can evaluate event data against a pattern base or set of characteristics that describe what the data means and who cares about it. You'll get a nugget of information that you weren't expecting but find really useful. Then, your BI activity will be focused instead of diffused.
If I were looking at 10,000 or a million simultaneous users taking the diffuse path to BI, I'd be very afraid. Performance management, planning, forecasting and dashboard metrics will be augmented in the future with smart, adaptive alerting agents. This is the only way organizations can scale up to extraordinarily large numbers of constituents.
Why do you live in New Hampshire? I can't tell you because then everyone will want to move here. Most influential book you've read recently? Equations of Eternity by David Darling. What do you do to relax? When I'm not traveling in the back of a 757, I fly my own Cessna.