A Data Science Central Community
Companies are getting smarter at predicting your next move.
As it becomes easier to gather information on consumers, businesses are crunching personal data in new ways to forecast a wide variety of behavior. In much the same way that credit scores predict how likely you are to pay your bills, a new generation of scores now rate the likelihood that you'll take your medications or redeem a specific coupon.
In some cases, transactions that were traditionally considered off the books — such as rent payments and payday loans — are being incorporated into the growing body of information used to size up customers.
The new uses of personal data raise a host of concerns for consumer advocates, who question the reliability of the scoring models and the accuracy of the information they rely on. Also troubling is that many consumers are oblivious that they've been tagged with these numbers, notes Chi Chi Wu, an attorney with the National Consumer Law Center. In many cases, consumers have no way to learn what their scores are.
"If this score is about me, I should be entitled to it," Wu said.
With credit scores, for example, lenders are required to disclose a score if it was used to deny a loan or assign a higher interest rate. Those who aren't actively seeking a loan can also pay to learn their credit scores from Fair Isaac Corp., which also goes by FICO, the name of its widely used score.
If you're wondering how else businesses are rating you, here's a look at four recently introduced scores you may not know about:
Anyone who has applied for a mortgage understands the importance of credit scores. The three-digit figures not only help determine whether a bank will approve a loan, but its interest rate as well.
Now a company called CoreLogic is developing a score it says will zero in on predicting a borrower's likelihood of repaying a mortgage. The score will be based on a new breed of credit reports the company released in November.
These reports gather information not typically listed on credit reports, including information from CoreLogic's in-house databases of rental records and payday-loan applications. Also included are public court records, such as property liens, evictions and child-support judgments.
The new score is intended to give lenders a more "complete picture" of mortgage applicants, said Tim Grace, a CoreLogic executive. He said that should lead to better lending decisions and reduced delinquencies for banks.
The exact formula for the score is still being developed with FICO. But once they're available in March, Grace said consumers will be able to purchase their scores for a price yet to be determined. For now, CoreLogic is required by law to provide customers with a free annual copy of the more detailed credit reports. Consumers can request their reports by calling 877-532-8778.
The business of scoring consumers isn't limited to financial matters. A score that was introduced last summer seeks to predict the likelihood that patients will take their medications. An individual's score can even vary depending on the condition; the score is available for hypertension, diabetes, high cholesterol, depression and asthma.
FICO says its Medication Adherence Score is intended to help health-care providers flag patients at risk of ignoring doctor's orders. The idea is improve overall patient outcomes and reduce health-care costs. The score is not available to individuals.
Interestingly, a patient's health and credit data are not used to determine the score. Instead, FICO says it can predict compliance based on demographic information such as household size; those who live alone are more at risk of skipping their medications. Owning a car, by contrast, is a good indicator for health-care providers, as is being neither very young nor very old.
And as it turns out, FICO says men are more likely to take their medications than women. Other information thrown into the formula include the rate of bankruptcies in a patient's region and purchase histories culled from the same databases retailers use to target households for catalogs.
FICO, which notes that the scores can't be used for insurance underwriting purposes, declined to say whether the score is being used by any clients yet. But the company has estimated that 2 million to 3 million Americans would be scored by this year, with that number set to rise to around 10 million by fall.
Asking a person how much he or she earns for a living is off-limits in most circles. But credit-card issuers and other companies can get a good idea of how much you make through an outside source.
Experian, one of the three national credit reporting agencies, last March introduced a product that predicts an individual's annual wages rounded to the nearest thousand dollars. The Income Insight W2 is based on the borrower's credit report.
"The intuitive explanation is that if you can maintain a mortgage or credit-card payment at a certain level each month, you're earning a minimum amount," said Brannan Johnston, vice president of income and assets at Experian.
The W2 is a variation of an income forecaster the company rolled out in 2009, which predicted total household income, including investment income and spousal income. The singling out of the individual's wages was a response to new credit-card regulations that require card issuers to assess card applicants' ability to afford their credit lines.
Although credit-card issuers are the most common users of the Income Insight W2, Johnston notes that many other companies — including debt collectors — also use it to gauge how much individuals are earning.
The items you put in your shopping cart aren't free from scrutiny, either. FICO says it has helped a third of the 100 largest U.S. retailers target their marketing based on customer buying patterns.
FICO declined to detail its roster of retail clients. But the warehouse discount club Sam's Club says it worked with the company to develop its eValues program, introduced about two years ago, that offers premium members personalized discounts.
Sam's Club uses its vast database of member transactions to determine "propensity scores," which gauge the likelihood that a customer would act on a particular discount. The scores even factor in the best time to offer that discount. For example, a customer who just bought three boxes of bulk cereal wouldn't be offered a discount on the same items right away.
So far, the program seems to be working. The company says that premium membership — which costs $100 a year, compared with $40 a year for standard membership — has more than doubled since eValues was launched. Customers who redeem an eValue discount also make more than twice as many trips to the store and tend to buy far more items during each visit, according to the company.
Although the scores aren't available to members, the company notes that shoppers are clearly benefiting from them.
"It's kind of like the eHarmony of couponing — we find the very best offers for the customer," said Catherine Corley, vice president of member program development at the company.