Hyperion's predictive analytics purchase demystified |
 |
By Hannah Smalltree, News Writer
15 Feb 2007 | SearchDataManagement.com |
|
|
In a business world filled with uncertainty and risk, wouldn't it be nice to be able to predict the future?
Hyperion Solutions Corp. hopes to get its customers a little closer to that impossible dream with its latest acquisition. The Santa Clara, Calif.-based business intelligence (BI) and corporate performance management (CPM) vendor recently agreed to acquire Denver-based Decisioneering Inc., which makes predictive analytics software that can help forecast the future -- even without historical data, according to Ivo Bauermann, director of product marketing with Hyperion.
Instead, Decisioneering's Crystal Ball software uses advanced modeling and simulation techniques that enable companies to better assess risk in uncertain situations -- such as launching a new product or starting a new project. Hyperion's existing software platform has some predictive analytics features now, Bauermann said, but those are based on current or "what is" scenario analysis and historical regression. The Crystal Ball software has more sophisticated "what if" analysis capabilities, enabling companies to set ranges of possible outcomes and assumptions, compare results, and model potential risks of business decisions.
"This closes the gap between guesswork and more statistical decision-making capabilities. It helps companies decide which scenarios are most likely to occur and which will have the highest impact," Bauermann said.
Hyperion plans to run Decisioneering as a standalone business unit and has no timeline for potential integration into its products, he said, but he believes that the next major release of Hyperion's platform will likely have some integration with the Crystal Ball software. Hyperion says it will continue to support Decisioneering's 4,600 customers. It also noted in its press release that the Crystal Ball software is used by 85% of Fortune 500 companies -- such as Becton, Dickinson and Company, a medical technology company based in Franklin Lakes, N.J.
At Becton Dickinson, both engineers and businesspeople use Decisioneering's Crystal Ball software, according to Marty Stout, principal engineer. The software helps Stout make engineering decisions about its medical products.
"It's impossible for us to test all the things that could occur when we make hundreds of millions of products. The only way we can really understand the way those products will perform is through statistical simulation," Stout said.
Crystal Ball also helps business users make decisions about whether to take on new projects, he said. The ability to combine some basic data with assumptions and statistically model potential outcomes is a useful function that cuts across departments. Of course, it's still "garbage in, garbage out," Stout noted. The tool supports human decision-making, and it's only as good as the data and assumptions that users plug into it. But the software has enabled Becton Dickinson to produce more "optimal systems," Stout said.
Supporting high-level decisions
Predictive analytics offers big benefits to high-level business strategists, according to John Hagerty, vice president of research with Boston-based AMR Research. Business and finance people frequently do a lot of modeling, he explained -- analyzing different scenarios to determine the possible outcome of business decisions. But they often end up with multiple potential outcomes -- and no way to determine whether one situation is more likely to occur than another. High-level statistical analysis aids decision makers by helping them assign probabilities to the likelihood that certain events will occur.
"The Decisioneering tool can give you a lot more comfort in the suppositions that you're making," Hagerty said, "and what the statistics say about the probability of the events happening."
But Hyperion isn't the only vendor touting advanced analysis capabilities. Other BI and CPM vendors also have predictive analytics, Hagerty said, such as Stamford, Conn.-based OutlookSoft Corp., Chicago-based SPSS Inc. and Cary, N.C.-based SAS Institute Inc. But while the latter two have had "lab-coat" or highly advanced tools with these features, he said, Crystal Ball is designed to work on the desktop, as an extension to the familiar Microsoft Excel interface.
And, Hagerty said, more BI vendors are likely to be promoting predictive analytics in the future, because these are the kinds of features that end users are asking for in a BI/CPM platform.
"The ideal situation is to say, 'Based on everything I know and what I expect to happen, this is the likelihood of the possible range of outcomes that I can get' " Hagerty said. "Predictive analytics can give you a lot more perspective on the potential results of a possible business strategy."
|
|
|
|
