Who are they, you ask?
Banks, insurance companies, employers, retailers, marketing agencies, political candidates, and more are using predictive analytics to gain valuable, yet potentially worrisome insights into people’s lives, such as when we’ll get pregnant, whether or not we’ll take our medicine, how we’ll vote, and even where we’ll be 24 hours from now. Driven by advanced logarithms and a seemingly endless supply of data provided by electronic financial transactions, Internet activity, and cell phone habits, researchers are able to garner a pretty good understanding of what makes people tick. This then allows them to better their target marketing efforts and maximize both efficiency and profits.
From a consumer perspective, the level of access that Big Brother has into our lives is primarily a cause for concern not because of how it’s currently being wielded, but rather because of its potential applications in the future (most of us read 1984 in high school). Nevertheless, since you’re likely being rated on much more than your ability to manage credit these days, it’s a good idea to know what’s going on. Listed below are some examples of the different ways researchers are rating us and/or predicting our behavior.
Credit Risk: Undoubtedly the most well-known consumer “score,” credit scores gauge how dependable we are in paying back money lent to us by a third party. Credit scores take into account payment history on loans, lines of credit, and certain investment accounts as well as the types of accounts you have open, how long you’ve been using credit, your available credit, and your credit utilization, among other things. They’re used by decision makers ranging from lenders at banks to landlords and employers. Some of the most commonly used credit scores are the FICO Score from the Fair Isaac Corporation and the VantageScore from Experian.
Social Influence: Social influence scores are used to determine how much weight your tweets, Facebook posts, etc., carry among your friends and connections. This is important information for marketers, as targeting advertisements to the most influential folks will likely lead to exponential customer gains. The Klout Score is perhaps the most well-known social influence score, and companies actually use it to determine the urgency with which they need to respond to consumer complaints. They want to keep the most influential customers happy, after all.
Profitability: As you might expect, profitability scores gauge how much money a company can expect to make off of you. They’re used to identify a company’s best customers, determine how to earn more off of currently less profitable customers, as well as better target marketing efforts. For example, the ConsumerView Profitability Score from Experian ranks households based on their estimated ability to pay their debts, how likely they are to respond to ads and/or get approved for a given financial account, and their overall profitability.
Attrition: Major banks and other subscription-based businesses use attrition scores to gauge the likelihood that customers will leave them within a certain period of time. These scores take into account internal data (such as whether your account activity or balance is increasing or decreasing) as well as information from the credit bureaus (such as whether you’ve recently opened a new account with a competitor). Attrition scores help businesses concentrate efforts on retaining customers at high risk of departure or avoid sinking too many resources into lost causes. TransUnion was the first credit bureau to offer an attrition score, but most banks use proprietary attrition scoring models that are more customized to their needs.
Bankruptcy Prediction: Bankruptcy prediction scores consider data points such as income, retained income, equity, and liabilities to determine a consumer’s risk of defaulting on their financial obligations and/or going bankrupt. Some of the most popular bankruptcy prediction scores are Experian’s aptly-named BankruptcyPredict Score, Moody’s KMV Score, and the Z-Score.
Medication Adherence: According to the National Consumers League, the failure to follow doctor’s orders for taking medication accounts for approximately 125,000 patient deaths each year. Medication adherence scores are based on publicly available information like job status and home ownership that have been determined to be predictors or whether patients will ultimately fill and take their meds. These scores, the most notable of which is offered by Fair Isaac (yes, the same company that offers the FICO credit score), help medical professionals determine the need for follow-up phone calls and other reminder measures.
Ability-to-Pay: Banks score current and prospective customers based on their ability to handle new financial obligations in order to comply with federal regulations, determine eligibility for spending limit increases, target offers to specific consumer segments, and improve collections efforts. The Income Insight Score from Experian and the Discretionary Spending Index from Equifax are notable examples of ability-to-pay models.
Advertising Receptiveness: Companies ranging from Facebook to clothing retailers to financial institutions conduct research to determine how different types of consumers will react to ads. Using geographic information, purchasing habits, etc., they can therefore target the right ads to the right people.
Predicting Life Events
Pregnancy Prediction: Retailers such as Target have been known to analyze purchasing habits to determine if a woman is pregnant and roughly when she is due, thereby assigning them pregnancy prediction scores. This allows them to better target deals and advertisements. In addition, retailers have found that by garnering the business of women while they’re pregnant, they can make long-term customers out of them.
Voting Habits: Statistical models based on past election results, Census data, and other publicly available information give politicians a sense of how certain geographic regions and consumer segments are expected to vote as well as which voters are up for grabs.
Divorce: Credit card companies are known to analyze purchasing habits (e.g. hotel reservations, marriage counseling, etc.) to determine a customer’s likelihood of divorce and therefore their chances of experiencing financial distress down the road.
Location: Researchers can use tracking data from your phone and those of your contacts to determine your whereabouts 24 hours from now to an average distance of about 70 feet. While this technology obviously isn’t being applied widely due to the legal concerns of tracking people’s phones, it could be a gold mine for law enforcement as well as retailers looking to introduce their offers to the most likely borrowers possible.
At the end of the day, the widespread use of predictive analytics is simultaneously extremely interesting and worrisome. It’s fascinating to see how disparate data points are being combined to accurately predict future behavior, but the seeds of misuse are clearly evident. Not only could governments use this new technology to unlawfully track citizens, but there are also concerns that it might lead financial companies to discriminate against certain consumer segments and effectively trap them with unattractive loans, credit cards, and other accounts. There have already been rumblings in the regulatory community to that effect, and we should expect to see action taken down the road.