Business Objects or HANA? A guide to SAP analytics environments
A comprehensive collection of articles, videos and more, hand-picked by our editors
SAP shops sitting on the sidelines of big data and predictive analytics now have good reason to consider taking...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
the plunge, according to experts who say SAP's entry into predictive analytics software coupled with its HANA in-memory database offers an integrated solution that has benefits over "best-of-breed" applications.
SAP officially became part of the predictive analytics landscape with the general release of SAP BusinessObjects Predictive Analytics last September. While SAP has previously offered advanced analytical tools through partnerships with SPSS and other vendors, this marks the company's first organically grown offering in what's fast becoming a hot technology area.
With its new offering, SAP is competing with veteran players in this space, including IBM's SPSS arm, SAS Institute, Information Builders and Oracle. It's also competing with specialty "best-of-breed" vendors like Revolution Analytics, Alpine Data Labs, and Opera Solutions, among others.
"The need [for predictive analytics] has always been there; now technology has enabled us to step in," said Charles Gadalla, SAP's director of advanced analytics and solution management. Gadalla added that in-memory database technology and recent advances in parallel processing and CPU technology have made the new predictive analytics offering possible.
Unlike traditional predictive analytics tools, which are designed for specialists and Ph.D. statisticians, BusinessObjects Predictive Analytics targets business users, according to SAP. The company cites its visualization capabilities and support for the R open source statistical programming language as two of the primary capabilities that make advanced analytics more accessible to a general audience. SAP also touts Business Objects Predictive Analytics' tight integration with HANA, SAP's in-memory database technology, which allows companies to run complex predictive models at high speeds.
"Ease of use is a core capability -- we want people using the tool to get up and running quickly," Gadalla said. "We haven't built a tool for the Ph.D. or statistician to use; we built something for the business user. They have to understand the business and be aware of models, but a Ph.D. in [statistics] is not a requirement to use this tool."
Built on SAP HANA
Business Objects Predictive Analytics' promised ease of use and enhanced performance are a powerful combination that may convince some SAP customers to invest in an all-SAP system that includes predictive analytics, according to several industry experts.
They say companies may want to standardize on SAP because doing so can lead to several benefits, including having a single vendor contact and a common data model across all applications. Standardizing on one vendor, these experts say, is one way to streamline processes and simplify architectures.
For more on analytical tools:
Learn about SAP and Hadoop's increasingly close ties
Read how retailers can use SAP BusinessObjects 4.0 and video analytics
Learn where SAP Visual Intelligence hits, misses
But perhaps the biggest selling point for SAP BusinessObjects Predictive Analytics is its tight integration with the HANA in-memory database, although the platform can be used in standalone mode with other databases, such as Sybase and Oracle systems, according to the experts. HANA processes data directly in the in-memory database, where data is also stored in columns, not rows -- both factors that allow it to churn through vast amounts of structured and unstructured data far more quickly than traditional data warehouse technology.
"The combination with HANA is where it all comes together," said John Myers, senior analyst for business intelligence and data warehousing at Enterprise Management Associates, an industry analysis and consultancy based in Boulder, CO. "One of the strong suits of SAP Predictive Analytics and HANA is that they allow people to go through iterative cycles much faster while tuning predictive models."
SAP or "best-of-breed"?
Companies that have specialized data modeling requirements may need to consider a competing vendor, according to experts.
For example, some of the more established best-of-breed predictive analytics tools may have more specialized classification capabilities or more robust financial modeling features, according to Myers.
Analytics experts also advise companies to evaluate tools based on what algorithms are supported, ease of model development and where the models will be run. Companies that have yet to invest in HANA may find that "best-of-breed" vendors offer the best option, Myers added. Organizations should also take some time to think about the skills they have in-house and the skills they want to invest in when trying to decide between SAP and the competition, according to Stein Kretsinger, director of project development at Elder Research, a consulting company specializing in predictive analytics, data mining and text mining based in Charlottesville, Va.
For example, if a company has a stable of statisticians already trained in SPSS or another tool, they are unlikely to want to retrain workers or invest in new skill sets. If they have the business expertise, but lack trained statisticians, then there's more of an opening for the SAP offering, which is geared toward business users.
"Let the talent lead you to the tool," Kretsinger said.
SAP BusinessObjects Predictive Analytics also offers data visualization tools, which allow users to explore information in a highly visual, user-friendly fashion. And that emphasis on visualization may give the product an edge over the competition, according to Cindi Howson, the founder of business intelligence consulting firm BI Scorecard and the author of SAP BusinessObjects BI 4.0: The Complete Reference.
Howson said visualization can substantially improve model development and make it easier for less experienced business users to determine which variables are most important.
What it really comes down to, Howson said, is where customers are in their use of predictive analytics.
"If they are already using SPSS or SAS, I really don't see customers changing because they are too invested in technology," she said. "If they are new [to predictive analytics], SAP will naturally be the first solution they evaluate."