Managing big data at the speed of risk


“When I started using predictive analytics in 1991, I had a desktop computer with a 600 megabyte hard drive running SAS® 5.0 something,” said Olivia Rud, respected business intelligence thought leader and author of Data Mining Cookbook: Modeling Data for Marketing, Risk and Customer Relationship Management. Technology has vastly improved since then, but Rud’s point was that predictive analytics is still the tool of choice.

According to Rud, she started out building logistic models for a credit card company. “It took 27 hours to run one model,” she said. “So I would spend the whole week getting the models ready – as linear as I could – and then I would start the model on Friday and pray that it didn’t crash once I’d started! Today, I think that same model can run in about 10 seconds.”

Rud was speaking at the Insurance and Finance SAS Users Group (IFSUG) Summit in Cary Monday about the challenges of big data and the value predictive analytics provides to banks to meet those challenges.

“The challenges of today are compounded by a global economy and the growing data volume. During this time of fast growth and complexity, predictive analytics is a very powerful way to manage that data,” she said.

Predictive analytics for risk management

Rud defines risk as the uncertainty of future outcomes. “Risk management is defining the optimal level of risk for your organization. It’s also putting in place tools and processes to manage your risk. So you have to start out by predicting your current risk levels and then using these financial tools, especially predictive analytics, to get your current risk level to your optimal level,” said Rudd.

Examples of predictive analytics use in risk:

  • Predict which customers will default, when it will occur and the amount of loss.
  • Predict the probability of a claim and the expected loss.
  • Develop optimal compensation packages and determine best candidates for high-level positions.
  • Predict customer lifetime value.

Today is the future

Rud’s presentation was one of many at the IFSUG Summit. Her point was well received – the world and technology are moving at a lightning pace. There were some in the room, and many at the summit, who could also remember working from a main frame or a desktop before gigabyte hard drives and Internet connections. Most now wrangle a sea of data and see the tsunami on the horizon.

Several summit presenters talked about SAS High-Performance Analytics (HPA), including Dave Macdonald, Keith Collins, Brent Lever and others. Their message was uniformly a challenge to rethink the constraints imposed by storage and access limits. It’s not just about having all of your data in one place or having good, clean data; you have to be able to use the data.

Keith Collins talked about increase in production that organizations will realize with HPA. He also wrote about it in a recent blog post, “It’s not the speed so much as the ability to ask – and answer – 20 times more questions, and then change the business as a result. So now the percentage change you can make in the business – the lift in sales, the identification of the fraud, and so on – goes up. So now you are really having a bottom-line impact on the company.”

To get a good understanding of HPA, I’d recommend you read Alison Bolen’s HPA once a day blog series. Get started with these posts:

Continue reading other posts collected from the IFSUG Summit:


About Author

Waynette Tubbs

Editor, Marketing Editorial

Waynette Tubbs is a seasoned technology journalist specializing in interviewing and writing about how leaders leverage advanced and emerging analytical technologies to transform their B2B and B2C organizations. In her current role, she works closely with global marketing organizations to generate content about artificial intelligence (AI), generative AI, intelligent automation, cybersecurity, data management, and marketing automation. Waynette has a master’s degree in journalism and mass communications from UNC Chapel Hill.

Leave A Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to Top