Super Crunchers

by Ian Ayres

  Still, Chicago law professor Cass Sunstein worries that there’s a social cost to exploiting the long tail. The more successful these personalized filters are, the more we as a citizenry are deprived of a common experience. Nicholas Negroponte, MIT professor and guru of media technology, sees in these “personalized news” features the emergence of the “Daily Me”—news publications that expose citizens only to information that fits with their narrowly preconceived preferences. Of course, self-filtering of the news has been with us for a long time. Vice President Cheney only watches Fox News. Ralph Nader reads Mother Jones. The difference is that now technology is creating listener censorship that is diabolically more powerful. Websites like Excite.com and Zatso.net started to allow users to produce “the newspaper of me” and “a personalized newscast.” The goal is to create a place “where you decide what’s the news.” Google News allows you to personalize your newsgroups. Email alerts and RSS feeds allow you to select “This Is the News I Want.” If we want, we can now be relieved of the hassle of even glancing at those pesky news articles about social issues that we’d rather ignore.

  All of these collaborative filters are examples of what James Surowiecki called “The Wisdom of Crowds.” In some contexts, collective predictions are more accurate than the best estimate that any member of the group could achieve. For example, imagine that you offer a $100 prize to a college class for the student with the best estimate of the number of pennies in a jar. The wisdom of the group can be found simply by calculating their average estimate. It’s been shown repeatedly that this average estimate is very likely to be closer to the truth than any of the individual estimates. Some people guess too high, and others too low—but collectively the high and low estimates tend to cancel out. Groups can often make better predictions than individuals.

  On the TV show Who Wants to Be a Millionaire, “asking the audience” produces the right answer more than 90 percent of the time (while phoning an individual friend produces the right answer less than two-thirds of the time). Collaborative filtering is a kind of tailored audience polling. People who are like you can make pretty accurate guesses about what types of music or movies you’ll like. Preference databases are powerful ways to improve personal decision making.

  eHarmony Sings a New Tune

  There is a new wave of prediction that utilizes the wisdom of crowds in a way that goes beyond conscious preferences. The rise of eHarmony is the discovery of a new wisdom of crowds through Super Crunching. Unlike traditional dating services that solicit and match people based on their conscious and articulated preferences, eHarmony tries to find out what kind of person you are and then matches you with others who the data say are most compatible. eHarmony looks at a large database of information to see what types of personalities actually are happy together as couples.

  Neil Clark Warren, eHarmony’s founder and driving force, studied more than 5,000 married people in the late 1990s. Warren patented a predictive statistical model of compatibility based on twenty-nine different variables related to a person’s emotional temperament, social style, cognitive mode, and relationship skills.

  eHarmony’s approach relies on the mother of Super Crunching techniques—the regression. A regression is a statistical procedure that takes raw historical data and estimates how various causal factors influence a single variable of interest. In eHarmony’s case the variable of interest is how compatible a couple is likely to be. And the causal factors are twenty-nine emotional, social, and cognitive attributes of each person in the couple.

  The regression technique was developed more than 100 years ago by Francis Galton, a cousin of Charles Darwin. Galton estimated the first regression line way back in 1877. Remember Orley Ashenfelter’s simple equation to predict the quality of wine? That equation came from a regression. Galton’s very first regression was also agricultural. He estimated a formula to predict the size of sweet pea seeds based on the size of their parent seeds. Galton found that the offspring of large seeds tended to be larger than the offspring of average or small seeds, but they weren’t quite as large as their large parents.

  Galton calculated a different regression equation and found a similar tendency for the heights of sons and fathers. The sons of tall fathers were taller than average but not quite as tall as their fathers. In terms of the regression equation, this means that the formula predicting a son’s height will multiply the father’s height by some factor less than one. In fact, Galton estimated that every additional inch that a father was above average only contributed two-thirds of an inch to the son’s predicted height.

  He found the pattern again when he calculated the regression equation estimating the relationship between the IQ of parents and children. The children of smart parents were smarter than the average person but not as smart as their folks. The very term “regression” doesn’t have anything to do with the technique itself. Dalton just called the technique a regression because the first things that he happened to estimate displayed this tendency—what Galton called “regression toward mediocrity”—and what we now call “regression toward the mean.”

  The regression literally produces an equation that best fits the data. Even though the regression equation is estimated using historical data, the equation can be used to predict what will happen in the future. Dalton’s first equation predicted seed and child size as a function of their progenitors’ size. Orley Ashenfelter’s wine equation predicted how temperature and rain would impact wine quality.

  eHarmony produced a formula to predict preference. Unlike the Netflix or Amazon preference engines, the eHarmony regression is trying to match compatible people by using personality and character traits that people may not even know they have or be able to articulate. Indeed, eHarmony might match you with someone who you might never have imagined that you could like. This is the wisdom of crowds that goes beyond the conscious choices of individual members to see what works at unconscious, hidden levels.

  eHarmony is not alone in trying to use data-driven matching. Perfectmatch matches users based on a modified version of the Myers-Briggs personality test. In the 1940s, Isabel Briggs Myers and her mother Katharine Briggs developed a test based on psychiatrist Carl Jung’s theory of personality types. The Myers-Briggs test classifies people into sixteen different basic types. Perfectmatch uses this M-B classification to pair people who have personalities that historically have the highest probability of forming lasting relationships.

  Not to be outdone, True.com collects data from its clients on ninety-nine relationship factors and feeds the results into a regression formula to calculate the compatibility index score between any two members. In essence, True.com will tell you the likelihood you will get along with anyone else.

  While all three services crunch numbers to make their compatibility predictions, their results are markedly different. eHarmony believes in finding people who are a lot like you. “What our research kept saying,” Warren has observed, “is [to] find somebody whose intelligence is a lot like yours, whose ambition is a lot like yours, whose energy is a lot like yours, whose spirituality is a lot like yours, whose curiosity is a lot like yours. It was a similarity model.”

  Perfectmatch and True.com in contrast look for complementary personalities. “We all know, not just in our heart of hearts, but in our experience, that sometimes we’re attracted [to], indeed get along better with, somebody different from us,” says Pepper Schwartz, the empiricist behind Perfectmatch. “So the nice thing about the Myers-Briggs was it’s not just characteristics, but how they fit together.”

  This disagreement over results isn’t the way data-driven decision making is supposed to work. The data should be able to adjudicate whether similar or complementary people make better matches. It’s hard to tell who’s right, because the industry keeps its analysis and the data on which the analysis is based a tightly held secret. Unlike the data from a bunch of my studies (on taxicab tipping, affirmative action, and concealed handguns) that anyone can freely download fro
m the Internet, the data behind the matching rules at the Internet dating services are proprietary.

  Mark Thompson, who developed Yahoo! Personals, says it’s impractical to apply social science standards to the market. “The peer-review system is not going to apply here,” Thompson says. “We had two months to develop the system for Yahoo! We literally worked around the clock. We did studies on 50,000 people.”

  The matching sites, meanwhile, are starting to compete on validating their claims. True.com emphasizes that it is the only site which had its methodology certified by an independent auditor. True.com’s chief psychologist James Houran is particularly dismissive of eHarmony’s data claims. “I’ve seen no evidence they even conducted any study that forms the basis of their test,” Houran says. “If you’re touting that you’re doing something scientific…you inform the academic community.”

  eHarmony is responding by providing some evidence that their matching system works. It sponsored a Harris poll suggesting that eHarmony is now producing about ninety marriages a day (that’s over 30,000 a year). This is better than nothing, but it’s only a modest success because with more than five million members, these marriages represent about only a 1 percent chance that your $50 fee will produce a walk down the aisle. The competitors are quick to dismiss the marriage number. Yahoo!’s Thompson has said you have a better chance of finding your future spouse if you “go hang out at the Safeway.”

  eHarmony also claims that it has evidence that its married couples are in fact more compatible. Its researchers presented last year to the American Psychological Society their finding that married couples who found each other through eHarmony were significantly happier than couples married for a similar length of time who met by other means. There are some serious weaknesses with this study, but the big news for me is that the major matching sites are not just Super Crunching to develop their algorithms; they’re Super Crunching to prove that their algorithms got it right.

  The matching algorithms of these services aren’t, however, completely data-driven. All the services rely at least partially on the conscious preferences of their clients (regardless of whether these preferences are valid predictors of compatibility). eHarmony allows clients to discriminate on the race of potential mates. Even though it’s only acting on the wishes of its clients, matching services that discriminate by race may violate a statute dating back to the Civil War that prohibits race discrimination in contracting. Think about it. eHarmony is a for-profit company that takes $50 from black clients and refuses to treat them the same (match them with the same people) as some white clients. A restaurant would be in a lot of trouble if it refused to seat Hispanic customers in a section where customers had stated a preference to have “Anglos only.”

  eHarmony has gotten into even more trouble for its refusal to match same-sex couples. The founder’s wife and senior vice president, Marylyn Warren, has claimed that “eHarmony is meant for everybody. We do not discriminate in any way.” This is clearly false. They would refuse to match two men even if, based on their answers to the company’s 436 questions, the computer algorithm picked them to be the most compatible. There’s a sad irony here. eHarmony, unlike its competitors, insists that similar people are the best matches. When it comes to gender, it insists that opposites attract. Out of the top ten matching sites, eHarmony is the only one that doesn’t offer same-sex matching.

  Why is eHarmony so out of step? Its refusal to match gay and lesbian clients, even in Massachusetts where same-sex marriage is legal, seems counter to the company’s professed goal of helping people find lasting and satisfying marriage partners. Warren is a self-described “passionate Christian” who for years worked closely with James Dobson’s Focus on the Family. eHarmony is only willing to facilitate certain types of legal marriages regardless of what the statistical algorithm says. In fact, because the algorithm is not public, it is possible that eHarmony puts a normative finger on the scale to favor certain clients.

  But the big idea behind these new matching services—the insight they all share—is that data-based decision making doesn’t need to be limited to the conscious preferences of the masses. Instead, it is possible to study the results of decisions and tease out from inside the data the factors that lead to success. This chapter is about how simple regressions are changing decisions by improving predictions. By sifting through aggregations of data, the regression technique can uncover the levers of causation that are hidden to casual and even expert observation. And even when experts feel that a particular factor is an important determinant of some outcome, the regression technique literally can price it out.

  Just for fun, Garth Sundem, in his book Geek Logik, used a regression to create a formula to predict how long celebrity marriages will last. (It turns out that having more Google hits reduces a marriage’s chances—especially if the top Google hits include sexually suggestive photos!) eHarmony, Perfectmatch, and True.com are doing the same kind of thing, but they’re doing it for profit. These services are engaged in a new kind of Super Crunching competition. The game’s afoot and it’s a very different kind of game.

  Harrah’s Feels Your Pain

  The same kind of statistical matchmaking is also happening inside companies like Lowe’s and Circuit City, which are using Super Crunching to select job applicants. Employers want to predict which job applicants are going to make a commitment to their job. Unlike traditional aptitude tests that try to suss out an applicant’s IQ, the modern tests are much closer to eHarmony’s questionnaire in trying to evaluate three underlying personality traits of the applicants: their conscientiousness, agreeableness, and extroversion. Data mining shows that these personality traits are better predictors of worker productivity (especially turnover) than more traditional ability testing. Barbara Ehrenreich was appalled when she took an employment test at a Minneapolis Wal-Mart and was told that she had given the wrong answer when she agreed with the proposition “there is room in every corporation for a non-conformist.” Yet regressions suggest that people who think Wal-Mart is for non-conformists aren’t a good fit and are more likely to turn over. It’s one thing to argue that Wal-Mart and other employers should reorganize their mind-numbing jobs to make them less boring. But in a world where mind-numbing jobs are legal, it’s hard for me to see what’s wrong with a statistically validated test that helps match employees that are most compatible with those jobs.

  Mining for non-obvious predictors is not just about hiring good applicants. It’s also helping businesses keep their costs down, especially the costs of stagnant inventory. Businesses that can do a better job of predicting demand can do a better job of predicting when they are about to run out of something. And it can be just as important for businesses to know when they’re not about to run out of something. Instead of bearing the costs of large inventories lying around, Super Crunching allows firms to move to just-in-time purchasing. Stores like Wal-Mart and Target try to get as close as possible to having no excess inventory on hand at all. “What they have on the shelf is what they’ve got,” said Scott Gnau, general manager of the data-mining company Teradata. “If I buy six cans of yellow corn off the shelf, and there are now three cans left, somebody knows that happened immediately so they can make sure that the truck coming my way gets some more corn loaded on it. It’s gotten to the point that as you’re putting stuff in your trunk, the store is loading the truck at the distribution center.” These prediction strategies can be based on highly specific details about likely demand. Before Hurricane Ivan hit Florida in 2004, Wal-Mart already had started rushing strawberry Pop-Tarts to stores in the hurricane’s path. Analyzing sales of other stores in areas hit by hurricanes, Wal-Mart was able to predict that people would be yearning for the gooey comfort of Pop-Tarts, finger food that doesn’t require cooking or refrigeration. Firms are engaging in “analytic competition” in an explicit attempt to out-data-mine the other guy, struggling to first uncover and then exploit the hidden determinants of profitability.

  Some of this
Super Crunching is done in-house, but truly large datasets are warehoused and analyzed by specialist firms like Teradata, which manages literally terabytes of data. Sixty-five percent of the top worldwide retailers (including Wal-Mart and JCPenney) use Teradata. More than 70 percent of airlines and 40 percent of banks are its customers.

  Crunching terabytes helps predict which customers are likely to defect to rivals. For its most profitable customers, Continental Airlines keeps track of every negative experience that may increase the chance of defection. The next time a customer who experienced a bad flight takes to the air, a data-mining program automatically kicks in and gives the crew a heads-up. Kelly Cook, Continental’s onetime director of customer relationship management, explains, “Recently, a flight attendant walked up to a customer flying from Dallas to Houston and said, ‘What would you like to drink? And, oh, by the way, I am so sorry we lost your bag yesterday coming from Chicago.’ The customer flipped.”

  UPS uses a more sophisticated algorithm to predict when a customer is likely to switch to another shipping company. The same kind of regression formula that we saw at play with wines and matchmaking is used to predict when a customer’s loyalty is at risk, and UPS kicks into action even before the customer has thought about switching. A salesperson proactively calls the customer to reestablish the relationship and resolve potential problems, dramatically reducing the loss of accounts.

  Harrah’s casinos are particularly sophisticated at predicting how much money they can extract from clients and still retain their business. Harrah’s “Total Rewards” customers use a swipeable electronic card that lets Harrah’s capture information on every game played at every Harrah’s casino they’ve visited. Harrah’s knows in real time on a hand-by-hand (or slot-by-slot) basis how much each player is winning or losing. It combines these gambling data together with other information such as the customer’s age and the average income in the area where he or she lives, all in a data warehouse.