Business Objects Dives Into Predictive Analytics

SAP's Business Objects on Wednesday announced Predictive Workbench, a new module for its business intelligence platform that lets businesses make predictions about such things as customer behavior and business performance. The module, based on technology from SPSS Inc., is the latest example of a traditional BI company looking to move beyond its expertise in reporting and analysis tools.

Predictive Workbench is the result of an OEM deal Business Objects struck last December to offer SPSS's technology. IBM's Cognos struck a similar deal with SPSS in March, and also plans to integrate IBM-developed predictive analytics within Cognos. Both efforts take a direct competitive shot at SAS Institute, by far the market leader in predictive analytics.

The traditional BI companies see a lot of opportunity in predictive analytics, which companies in such areas as consumer goods, finance, and retail use to study patterns and trends in historical data, play out what-if scenarios, and make educated assumptions, all in the quest to maximize profits. Retailers, for example, use predictive analytics to identify what customers are most likely to respond to marketing campaigns, and banks use it to identify financial fraud. IDC predicts the $1.4 billion market for advanced analytics, of which predictive analytics is a subset, will grow 10% annually through 2011.

Since the December OEM deal, Business Objects has worked to integrate the SPSS technology, called Clementine, into Business Objects X1 3.0 so that it's "totally transparent and seamless for customers," said Franz Aman, VP of BI product marketing. The result is Predictive Workbench, from which a user can launch the Business Objects Universe metadata layer to run predictive analytics against various databases and data warehouses. Business Objects will disclose the price of the module only to interested customers, Aman says.

It's not likely cheap, since predictive analytics is traditionally a pricey discipline left to statisticians and scientists with a deep knowledge of data. Those types will typically be the main users of Predictive Workbench, Aman acknowledges, but adds that Business Objects' long-term goal is to tweak the technology so that it's more accessible by common BI users, such as financial and marketing analysts.

The goal also is to convince customers that they need only one BI platform--Business Objects--companywide, rather than go to Business Objects for BI tools and SAS or another vendor for advanced analytics. "A lot of work goes into prepping data for analysis," Aman said. "Our approach is to make it possible for BI folks to work constructively with statisticians and PhDs to tap into data already available."

SAS, however, also offers traditional BI tools. SAS has 31% of the advanced analytics market while SPSS, its closest competitor by far, has 14%, according to IDC. The remaining players have 3% or less of the market.