Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

HANA injects SAP's predictive analytics lineup with speed, flexibility

As SAP clarifies its predictive analytics strategy, companies must take special care to match each product's complexity and target user to their business needs.

When SAP put forth the initial use cases for adopting its HANA in-memory database several years ago, the touted benefits were significant accelerations in executing operational analytics and reporting. As HANA has evolved into both a full-fledged application platform and the core infrastructure that powers all SAP applications, the strong initial focus on HANA-based analytics has broadened to include planning, textual, spatial and predictive analytics.

Loosely defined, predictive analytics is a multi-stage process that helps companies to arrive at some previously unclear or unknown outcome. Companies tend to use predictive analytics to discover new revenue opportunities or to shine a spotlight on previously unforeseen risks to their businesses. Predictive analytics begins in the world of statistical modeling, machine learning and data mining to identify key points of interest or specific patterns in a firm's past and present data. Those data points form the basis of models that capture relationships between data points, models which, once built, can then be run and scored against new data sets to determine matches or deviations from the patterns and thus arrive at predictions.

For instance, a cable provider can deploy predictive analytics to help its call center employees pitch the service package that has been proved to appeal most to a particular kind of customer as defined by income range and geographic location. The result is an improvement in customer service and retention leading to increased revenue. In another example, a firm can use predictive analytics to help identify patterns suggesting that one of its suppliers may be in financial difficulty. In this case, the outcome is being able to anticipate and manage the risk associated with a potential bankrupt supplier and to set in motion alternative sourcing plans if necessary.

Breaking down SAP predictive analytics options

SAP has three main predictive analytics offerings -- SAP Predictive Analysis, InfiniteInsight and Lumira -- all of which run on HANA and target different types of users.

SAP Predictive Analysis is on-premises data discovery, visualization and predictive analytics software aimed at the traditional predictive analytics user, a data scientist who feels at home using such complex technologies as neural networks and is adept at creating custom analytics. The software has a graphical user interface into the open source R statistical language. SAP debuted Predictive Analysis after IBM acquired its longtime predictive analytics partner, SPSS. The market is a mix of large software players like IBM, Oracle and SAS (with whom SAP has a HANA-focused partnership), as well as open source players and startups.

SAP InfiniteInsight is the renamed KXEN cloud-based technology, which SAP acquired in 2013, and is aimed at business analysts who have some familiarity with predictive modeling and algorithms, but don't have the deep analytical and customization skills of data scientists. InfiniteInsight also has a stronger focus on usability and automating model creation. Retailer Macy's is using InfiniteInsight to improve its targeting of customers through its email and website marketing campaigns.

Lumira, formerly Visual Intelligence, a cloud-based business intelligence (BI) client for predictive analytics, is intended for business people, and its primary focus is data visualization. SAP has infused some of the acquired KXEN technology into both the Predictive Analysis product and Lumira. The vendor is reflecting a general trend in predictive analytics to open up some of the processes and automate others so a broader group of people can use more elements of the software. Many businesses have yet to adopt predictive analytics and one key barrier is the lack of sufficiently skilled analytics personnel.

What HANA brings to predictive analytics

Alongside its trio of predictive analytics products, SAP is also coming out with a growing number of industry-specific and business applications, such as fraud management, that contain embedded predictive analytics drawing on the HANA platform's Predictive Analysis Library. There are plans to include predictive analytics in more applications, with a particular focus on SAP's emerging cloud-based financials software.

More on SAP analytics and HANA

Read an e-book on SAP BI best practices

Understand SAP's business intelligence strategy

Decide between third-party and SAP analytics software

Running on an in-memory platform like HANA has two main benefits for predictive analytics users: speed of execution of predictions and the ability to build models based on a mix of structured and unstructured data. The types and volumes of unstructured data such as text, video, audio and sensor-based inputs continue to increase at an immense pace and, again, companies are looking to have more of their staff trawl these data sets in combination with the data generated by their business applications to identify actionable patterns.

SAP's moves in predictive analytics mirror its companywide strategic goals of increasing HANA adoption, improving usability, and embedding more functionality in its business applications. It's currently unclear how it will differentiate standalone predictive analytics products from embedded predictive analytics, and whether, as seems likely, over time the latter will predominate.

For SAP customers, the choice is how much control in the form of customization they require from their predictive analytics and the cost equation of buying specialized software over having some, more limited capabilities embedded in their business applications. Quantifying the benefits of predictive analytics can be clear cut in the case of unearthing new revenue opportunities, but less so in dodging risk, since once you act on that knowledge, it changes the future.

About the author:
China Martens is an independent business applications analyst and freelance writer. Email her at chinamartens@gmail.com or follow her on Twitter at @chinamartens.

Dig Deeper on SAP business intelligence

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.