Case Study

Using analytics to develop more accurate credit risk scores and identify new business opportunities.

About the client. Headquartered in Atlanta, Equifax operates or has investments in 18 countries, including the US, Canada and the UK, and is a member of Standard & Poor’s (S&P) 500 Index. Its common stock is traded on the New York Stock Exchange (NYSE) under the symbol EFX.

Smart is… helping clients identify business opportunities by accurately predicting which of their customers will move to a new home in the next six months. Equifax sees significant potential for growth in the Canadian market by complementing its traditional credit scoring offerings with innovative analytics services based on advanced modeling of vast quantities of customer data gathered by clients in the Canadian banking, utilities and tele- communications sectors. A prime example of these new value-added services is Equifax’s “movers’ model”, which identifies the customers who are most likely to move to a new home in the next six months. The analytics team has also developed the Equifax Risk Score (ERS), a more sophisticated credit risk scoring system, which significantly outperforms the previous generation of scoring models.

Real numbers, real results.


More effective at predicting
top 10% of movers


In savings per year for top banks in
Canada using the new ERS model



Growing business in Canada.
Equifax sees Canada as one of its most advanced and profitable markets – a market where clients are keen to adopt added-value analytics services that complement traditional credit scoring offerings. More than 30 percent of the company’s Canadian revenue already comes from analytics-based services, and the Technology and Analytical Services team in Equifax’s international division is constantly working to develop new and innovative products and services.

Santiago Villasis, Director, Technology and Analytical Services, explains: “Our clients in sectors such as banking, telecoms and retail gather enormous quantities of data about their customers’ behaviour, and we have over 25 years of experience in analyzing and modeling these data to assess credit risk. But of course, the data can be used for much more than just credit scoring – and this is a huge opportunity for our business. We’re working with our clients to develop all kinds of exciting new technology and analytical services, and IBM Business Analytics software plays a key role in delivering them.”

IBM SPSS Statistics and SPSS Modeler are an important part of the Technology and Analytical Services team’s toolkit – supporting the development of sophisticated statistical models with capabilities such as leading-edge algorithms for predictive analytics, segmentation, decision trees and data mining. The team is also currently looking to extend its use of IBM Cognos Analytics to help present results to clients in a more graphical and intuitive way.

Our custom model provides an uplift in accuracy of 365 percent compared to a baseline of random selection. It is also approximately 40 percent more effective than the client’s original model.



Developing a movers’ model.
As one example of a recent project where IBM Business Analytics software has helped Equifax develop a new service, Santiago Villasis describes the company’s new “movers’ model”.“One of our clients came to us and asked whether we could use their data to predict when their customers were going to move to a new home,” he explains. “They had already tried to build a predictive modeling solution themselves, but the results weren’t as good as they had hoped, and we were convinced that we could do better.”

Using a decision tree methodology, the Equifax team was able to whittle down a huge list of possible variables within a generic customer dataset to a small number of factors which serve as accurate predictors of whether a customer is likely to move homes within the next six months. This generic mover’s model was then customized for the client’s specific dataset, which delivered even better results.“For identifying the top ten percent of individuals who are most likely to move, our custom model provides an uplift in accuracy of 365 percent compared to a baseline of random selection,” notes Santiago Villasis. “It is also approximately 40 percent more effective than the client’s original model.”It is important to note that Equifax does not make available or disclose any individual information for marketing or prospecting purposes without the person’s consent.


Business benefits and smarter analytics.
New insights from predictive analytics help Equifax’s clients make better decisions in areas such as marketing, loan approvals, portfolio management and collections.The movers’ model is 40 percent more effective at predicting the top 10 percent of movers than the client’s own predictions, enabling better targeting of products and marketing campaigns.By improving credit risk management, the new Equifax Risk Score (ERS) is estimated to save a top-five bank $2.5 million per year, compared to the previous generation scoring model.