Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

June 30, 2003

Power Thinking

On storage, analytic architectures, and Erik Thomsen's new adventure

by David Stodder

I would hardly be in the minority declaring that we're living through momentous times that fill us with uncertainty as we try to make the leap from one age to another. Public and private organizations draw from this context a clear sense that they must change how they view risk, security, and unfolding economic threats and opportunities. In an information age, data is the critical strategic resource for interpreting change and analyzing what decisions must be made for success — if not survival — in the future.

The problem is, as Richard Scannell points out in our cover story ("It's What's Inside That Counts"), IT's desperate hunt for cost efficiency — and the response from key IT providers in the storage management community — has focused on resource utilization and capacity, and not enough on the valuable contents that the infrastructure supports. Frankly, this isn't surprising: projecting out mushrooming data needs, many organizations overbought in the 1990s. IT managers are determined to make full use of current storage investments.

Ever since we launched this publication, we've been tracking the emergence of storage management. The sector has evolved from an unglamorous (albeit profitable) adjunct to server hardware into what we view as one of the primary components of the intelligent enterprise infrastructure. It would be a shame, at this propitious time, for the storage community to lose the initiative. We hope Scannell's thoughtful discussion helps IT managers and solution providers refocus on "what's inside" — and on how to make it count when the enterprise demands knowledge to comply with new regulations or to determine the best business strategy.

How will the increasing emphasis on accessing and analyzing data influence the development of enterprise IT architectures, which for so long have focused on the "moment of truth": the closure of business transactions and all that follows? In "A New Analytic Perspective", Kemal A. Delic and Umeshwar Dayal bring us to the proper starting point: the user. The authors discuss how emerging analytic architectures must align with user needs — and build on that basic notion to allow innovative technologies to reach their potential.

Erik in Vegas

Corporate governance is probably not the first thing you'd imagine visitors obsessing about on a trip to Las Vegas. At the risk of understatement, it's fair to say that America's self-styled capital of pleasure and licentiousness is adept at convincing normal folks (possibly even CFOs and accountants) to not think too hard about where their money went. However, the subject dominated "Solutions 2003," Hyperion Solutions' annual North American Conference (May 5-7), held at The Venetian.

There, I sat down with Erik Thomsen, newly installed as distinguished scientist at Hyperion. As an independent researcher and consultant with DSS Labs, Erik has served as a contributing editor to Intelligent Enterprise since our first issue in 1998. Our association began in the pages of one of this magazine's predecessors, Database Programming & Design.

Our commitment to editorial independence prevents us from having contributing editors who are employed by software providers — and so we have had to say farewell to Erik in that capacity. In a concluding Author's Note to his recent article ("BI's Promised Land, Part II"), Erik offered gracious parting comments, which I heartily return. On behalf of our readers, we thank him for a memorable series of columns and articles that have contributed much-needed clarity and intellectual vigor to the conception, implementation, and management of business intelligence, data mining, decision support, online analytic processing, and related applications.

BPM: Tackling Complexity

While swapping impressions about the conference focus on corporate governance and how the heat on this topic might influence the emerging field of business performance management (BPM), Erik noted the irony that we were discussing these topics not only in the land of loose slots, but also in a hotel/casino fantasy of Venice, Italy. Some seven centuries ago, during that real city's heyday as one of the world's capitals of international trade and finance, merchants put the new invention of double-entry bookkeeping to good use as a measure to improve accountability. "People have always wanted transparency," Erik said. "But today, business environments are becoming so complex. Controlling that complexity demands new levels of accountability — and will require different techniques from those that have been successful for the last 20 or 30 years. Managers can't be faced with the old game of telephone when they might have 75 stations between the operational end and themselves. Will the message be the same by the time they get it? This is a real accountability challenge.

"Plus, think of all the stakeholders you've got now, including some who play multiple roles," he continued. "You have not only managers, who are held accountable: You have creditors, regulators, and owners, including employee pension funds that might control 30 or 40 percent of the stock in the corporation. So, there's huge pressure to improve transparency and remove the friction, if you will, in the flow of information. In turbulent times, we can't ignore things the way we used to. Organizations are much more global and interdependent, which leads to increasing volatility. We need to figure out how to measure, estimate, and internalize our understanding of this external volatility. We have to do this at the same time that we're figuring out internal risk — and not just financial, but operational risk."

Many companies are reaching for predictive analytics to understand risk. "But you can't predict until you have some concept of pattern," Erik said. "And you can't think about patterns unless you measure. Measuring is the first step, and is a key part of BPM. You can't manage what you don't measure.

"You also need a model. I've always had a gripe about data warehousing approaches that say, 'Here's the data, now go ahead and analyze it.' You can't really do that; you need some model of what the data is. To develop models, we need to look at processes and historical behavior. How else can you differentiate between random variances and systemic change? That's what you're trying to figure out. It's all part of this beautiful dance between data and the models of data."

Time is Right

"My first real love was the creation of technologies," Erik answered, when I asked him why he decided to leave consulting. "My first company was in software development. But the seas are pretty rough right now, and it can be tough with just a little raft. Hyperion has staying power and direct access to an existing customer base. It's going to be great to develop software for people who are actively using the solutions and thinking about this set of problems we've been discussing. Hyperion is balanced between a core multidimensional engine, on the one hand, and a set of domain-specific, managemen-oriented applications on the other. The company is totally focused on performance management. Both sides are in this together — we have to be."

Rate This Article

Comments:

Optional e-mail address:

For BPM and BI providers, user expectations remain a major challenge. "We're moving from an age where a small number of people have been doing all the analysis, to one in which everyone is doing some." Thomsen concluded, "You might only need access to analysis once or twice a week for 15 minutes, but that might be an incredibly important 15 minutes. If you have to read six manuals and make 17 calls to coworkers and software companies, the effort is useless. Everyone has a right to expect that the system be as simple to use as the information you're trying to get. Business complexity — not software complexity — is what you're trying to understand."