How predictive analytics can deliver strategic benefits

The technology can help companies quickly identify and respond to new opportunities

PHOENIX -- Enterprises can gain significant long-term benefits by applying predictive analytics to their operational and historical data, said analysts and IT managers at Computerworld's BI & Analytics Perspectives conference being held here this week.

Unlike traditional business intelligence practices, which are more backward-looking in nature, predictive analytics approaches are focused on helping companies glean actionable intelligence based on historical data.

If applied correctly, predictive analytics can enable companies to identify and respond to new opportunities more quickly, they said.

In a keynote address, James Taylor, CEO of Decision Management Solutions, said predictive analytics is especially useful in situations where companies need to make quick decisions with large volumes of data.

Predictive analytics practices can help companies in three key areas: minimizing risk, identifying fraud and pursuing new revenue opportunities, Taylor said.

For instance, predictive analytics can help companies fine-tune their ability to identify risk in areas such as loan and credit origination or fraud in areas such as insurance claims, he said.

Importantly, by embedding predictive analytics into operational data, companies can put themselves in a better position to identify new revenue opportunities, Taylor said. For example, by looking at a customer's historical purchase patterns, companies can make reasonable predictions about the kinds of promotional offers and coupons that are likely to resonate with that customer.

Blue Cross and Blue Shield System (BCBS) is one organization that is already deriving considerable benefits from predictive analytics. As an organization that provides healthcare insurance to nearly one in three Americans, BCBS has amassed a huge amount of claims-related data over the years.

A few years ago, the BCBS Association, the entity that holds the Blue brands, created a single database called Blue Health Intelligence (BHI) to consolidate all the claims information maintained by each of the 39 companies that are part of BCBS. The database is one of the largest repositories of de-identified healthcare data anywhere and contains claims-related information about more than 100 million people.

BHI operates as an independent unit and provides a range of business intelligence services that is enabling better healthcare services for members while also transforming the manner in which BCBS manages its costs.

The impetus for the effort came from the need for BCBS, like other health insurers, to control spiraling costs, said Swati Abbot, president and CEO of BHI, during a presentation.

A disproportionate share of healthcare costs goes toward the care of people with chronic illnesses, Abbot said. In fact, the top 5% of healthcare users account for more than 55% of healthcare costs, she said.

By applying predictive analytics technologies to its vast trove of claims data, BCBS has been getting better at not only identifying the risk factors that lead to several chronic diseases, but also identifying individuals who are at heightened risk of getting such diseases, she said.

"For every member enrolled in a health plan, we have a health score," which represents the likelihood of that individual needing lifelong treatment for a chronic illness, Abbot said. BHI has even developed disease-specific modules, such as one for diabetes, which predicts an individual's risk of getting diabetes based on previous data, she said.

The goal is to be able to use the data to get doctors to provide better, more targeted care for high-risk patients, reducing their need for expensive, long-term treatment, she said. The predictive modeling is enabling BCBS to move toward a more incentive-based healthcare model in which doctors get incented for performance, Abbot said.

Online dating site is another company that relies heavily on predictive analytics to run its service. The company collects and maintains a lot of information, some collected from subscribers and some collected by monitoring subscribers' interactions on

The company's challenge is to find a way to improve revenue per subscriber by delivering the best possible matches based on each subscriber's preferences, said Jim Talbott, director of consumer insights at

It is a task that is complicated by the fact that subscribers might indicate a specific set of requirements for a potential partner but then interact with people who fall outside of their specified range of preferences, he said.

To meet the challenge, has developed a predictive model that matches people based on not just their stated preferences, but also their site behavior and interactions.

Companies interested in predictive modeling need to have a clear idea of their objectives before they start, Taylor said. They need to know what sort of decisions will be powered by their predictive analytics and work backward from there, he said.

To develop a good predictive model, enterprises need to focus on defining a clear set of business rules for each decision and then focus their analytics on driving the best decisions, he said.

Jaikumar Vijayan covers data security and privacy issues, financial services security and e-voting for Computerworld. Follow Jaikumar on Twitter at @jaivijayan or subscribe to Jaikumar's RSS feed . His e-mail address is

Copyright © 2011 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend
Shop Tech Products at Amazon