Computerworld - John Kopcke has been in the decision-support systems business -- now we call it business intelligence -- for about 26 years. He's seen it all, from the early days of executive information systems and online analytical processing, to today's data analytics. Computerworld talked with Kopcke, chief technology officer at Hyperion Solutions Corp. in Sunnyvale, Calif., about the latest trends in business intelligence.
What's driving business intelligence projects these days: better financial reporting or customer relationship management (CRM)? From our point of view, it's financial reporting. But we'd expand that to a concept called business performance management, which goes beyond just reporting. My experience is that when companies are being hit by external political and economic forces -- and companies aren't making money hand over fist -- the attention tends to be focused on how to run the business better. You haven't seen a lot of new CRM initiatives recently; it had its heyday 12 to 18 months ago.
We hear a lot about real-time business intelligence. How much of that is hype and how much is reality? There's certainly hype associated with it. There's also multiple definitions of the term real time.
Most of the real-time technology work in the last few years has been about handling massive amounts of transaction data and looking for anomalies or things that break particular rules. That's really real time. Applications such as fraud detection, call centers, stock prices. Hyperion has tried to give those transaction alerts some context. It doesn't do the manager a lot of good just to get an alert that some threshold has been broken, if they don't have the supporting information to help them analyze it and decide what to do.
Another definition of real time is up-to-date information, which is as fresh as it needs to be. The focus is on a very small latency between when an event occurs and when the business has responded. If you have a latency that is significantly shorter than your competition, then you'll enjoy significant advantage.
No organization in the world missed the fact that 9/11 occurred. But when you look at how long it took for organizations to react and change their business after 9/11, you see a huge difference. Southwest Airlines took only two days to change its entire business from top to bottom, while others took months or probably are still coming to grips with it.
So we try to reduce that latency between when an event has occurred -- say, a competitor has launched a new product, war has broken out -- to where the business is recast to take advantage of it.
Are there downsides to real-time data? If you've got a completely automated decision system that executes a decision based on an event, like a transfer of funds, you have to pay special attention to the quality of that data.
The next problem is not bad data but partial data. It's better to make a decision based on no information than it is to make a decision based on partial information. Partial information, even if it's correct, might be as misleading as bad data. You have to ask yourself if you've pieced together a complete view of the situation.
Executive dashboards: What's the level of adoption? The first executive dashboards actually went into organizations around 1985. We called them executive information systems at the time. And they had limited success because they were executive systems -- the chairman of Merck would have it on his desk -- but then that was it. What we're seeing today are management dashboards, which have been pushed down through the organization, providing relevant information to a particular manager. At Southwest Airlines, they call them cockpits, and they're specialized, so that the guy in charge of putting peanuts on airplanes gets a different view than the guy who's in charge of purchasing jet fuel. But they all see what planes are flying where.
So I'd say dashboards are leaving the early-adopter phase and becoming more mainstream.
Fancy data visualization screens are pretty, but does anyone really use them? At the low end we've had pie charts, maps and histograms, and that's quite common. At the high end of the spectrum, with advanced visualization, you see much less adoption. That's not because the techniques aren't useful, but it's very difficult for the business user to actually understand what the visual is trying to tell them. Yet the whole purpose of visualization is to simplify, not to become more complex. In some cases you spend more time explaining the visualization to the business user than they actually get out of it.
There will be new advances in visualization -- we keep looking for the next killer app in visualization, and there will be one -- but they come slowly, and they really have to deliver value.
The other thing to keep in mind is that what works in visualization varies from one part of the world to another. Things like spider charts are important in Japan, but you rarely see them in the U.S., for example.
Read more about BI and Analytics in Computerworld's BI and Analytics Topic Center.
Free Insider GuideCopyright © 1994 - 2012 Computerworld Inc. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of Computerworld Inc. is prohibited. Computerworld and Computerworld.com and the respective logos are trademarks of International Data Group Inc.
