Building Attributes? Start with the Right Data

 In Credit Risk Management

In a previous blog, I talked about how attributes are the cornerstone of decisioning policy and creating a centralized approach to attributes across all lines of business and channels is critical. In this blog, I want to discuss how acquiring and/or creating attributes from data sources in addition to the credit bureaus can offer better credit decisions and ultimately better customer experiences.


Historically, credit bureau data incorporates bank trades, public record data, and inquiries. Bank trades include history on payments for different credit products like credit cards and auto loans, while public record data includes things like delinquencies and bankruptcies. Inquiries are the number of times someone has pulled your credit file. This data is used to calculate attributes, such as how many times a consumer has been at least 30 days late in making a payment. Attributes are building blocks of the analytic models used to determine whether or not credit should be offered. The challenge with attributes is that they are only as good as the data that they are based on.

No Data, No Attributes

We know that there are millions of individuals who don’t have a full credit bureau file. There are thin files, no files, and a large range of folks who fall in between. For this population, your standard attributes based on standard credit bureau data won’t work. So, how can we analyze them? By building attributes based on data other than credit bureau data—everything from utility payment data to professional license and certification information.

For example, ID Analytics collects data on individuals’ cell phone usage. This data shows there are a surprising number of individuals who will not pay their rent or credit card bill, but will pay their cell phone bill every month in order to stay “connected” via texts, tweets, and emails. This is very useful data in understanding that this consumer is potentially risky when it comes to offering a long-term loan, but a product that can support instant gratification will probably be paid off every month. Utilizing attributes based on this data could enable a financial institution (FI) to determine which type of product may best fit this consumer, as well as the terms that are associated with it like APR and credit limit.

Better Data = A Better Rank Ordering of Risk

Even when you have the credit bureau data necessary to calculate your attributes, you might not have enough data to make the best possible credit decision. FICO, an industry leader in calculating credit scores, has observed a reduction in the predictiveness of their classic scoring models since the recession. Those scoring models, based on standard credit bureau data, are less effective at rank ordering the risk of different consumers because the underlying data doesn’t paint a complete picture of consumers’ behavior. Consequently, a number of credit scoring providers including FICO are encouraging their customers to adopt their more contemporary scoring models that include alternative data and do a much better job rank ordering consumers based on their statistical likelihood of default.

Don’t Forget Customer Experience

In addition to better credit risk decisions, alternative data attributes can produce radically better customer experiences. For example, relational data incorporated into attributes, like the type that CBCInnovis offers, track all of the different mailing addresses a consumer may have tied to his name. Since this is not provided in standard credit bureau data, an FI may decline a consumer with several legitimate addresses because of an address mismatch. This poor customer experience becomes even more of a glaring problem when you consider that a large percentage of consumers with multiple addresses are affluent consumers with summer homes and business addresses.

Another source of alternative data for building better attributes is an FI’s own internal customer behavior data which can be used to identify the number of accounts and products consumers have and which channel they prefer to interact with. These attributes have the potential to improve future customer interactions by taking those customers’ preferences and past behavior into account.

The More You Know

There is a wealth of information available today, often called “big data”. But, how that data is used is critical in making credit policy stronger. The ability to pull in or create attributes from a variety of data sources including alternative data providers and internal customer data enables FI’s to make better risk decisions and provide better customer experiences.

Recommended Posts

Leave a Comment

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

Contact Us

We're not around right now. But you can send us an email and we'll get back to you, asap.

Not readable? Change text. captcha txt

Start typing and press Enter to search