≡ Menu

Predictive Analytics spotlight from IBM

Share

Deepak Advani kicked off day 2. SPSS was acquired by IBM in 2009 and, like all acquisitions, has to report on how well it has met the expectations set at acquisition time. SPSS has beaten every target set for it at acquisition and reviews have apparently been a “bit of a love fest”. SPSS has always had a strong direct to modeler business but the IBM acquisition has really boosted their ability to sell to the enterprise and that’s part of the reason for its success. In addition, SPSS’s ability to support the core Smarter Planet message has been a critical advantage – being able to put the “smart” in Smarter Planet. Finally SPSS has taken the use cases that resonate – like reducing churn, hospital treatment survival rates or proactive maintenance of equipment – and promoted them to show the practical value of predictive analytics to both customers and internal IBM people.

Deepak sees three broad categories or pillars of analytics within which they are “hardening” industry solutions where the use case has been proven out and is widely supported and understood:

  • Predictive Customer Analytics
    Help companies acquire, grow and retain customers. Industry solutions include up-sell/cross-sell, market basket analysis, churn prevention, customer segmentation.
  • Predictive Operational Analytics
    Manage, maintain and maximize value of systems and components. Industry solutions include predictive maintenance, assortment planning, condition monitoring and reverse logistics
  • Predictive Threat and Risk Analytics
    Monitor, detect and control risk and fraud. Industry solutions include claims fraud, insider threats, credit card fraud.

Examples from customers include reducing direct marketing cost by 18%, 600% ROI, reduction of repeat repairs by 25%.

These scenarios are supported by the four elements of SPSS products – data collection, statistics, modeler and decision management (predictive analytics deployed with business rules to turn predictions into actions).

SPSS is also bringing social media analytics to bear on these problems. The real value of social media analytics to SPSS (and I would agree) comes when it is integrated with other analytics. For instance, bring what people are saying on social networks in and use to identify advocates. Use predictive analytics to predict which actions have the highest probability of increasing advocacy. Apply these models in marketing campaigns to drive appropriate actions.

SPSS Decision Management is a key product in the SPSS stack. By combining business rules  with predictive analytics SPSS Decision Management allows for rapid model deployment and the turning of these models into action. SPSS Decision Management offers a simpler interface for real time model deployment and integration with business rules. Pre-packaged solutions and easy to use simulation make the product appealing to business owners.

Integration with other IBM products is proceeding apace. Integration with Cognos so that Cognos data sets can drive models. Integration with Netezza for in-database analytics. Taking advantage of IBM research algorithms in model development and using IBM’s industry frameworks as the basis for new SPSS Decision Management solutions.

The world says Deepak is changing and creating new opportunities for predictive analytics. Micro targeting is growing with 1:1 personalization but opt-out is moving to opt-in, increasing complexity. Using time and spatial dimensions in conjunction with propensity to buy analytics as well as operational factors to make a compelling offer for both customers and companies. Big Data, unstructured data, social media data are all making it both more compelling and more complex to build predictive models.

The core message from Deepak – Predictive Analytics is hot and opportunities are widespread.

Share

Comments on this entry are closed.