Friend, brilliant reporter, true believer and now NY Times best-selling author Julia Angwin has a thrilling book out explaining the real-life consequences of privacy erosion, and her own personal journey in escaping Eyes in the Sky. She’s been good enough to provide me with a free excerpt, which I’ve broken into a couple of pieces. Below is the first I’ll share, about how companies toy with your ability to figure out what price you should pay for items. Remember, they have Big Data, and you don’t. Here’s why that hurts:
As companies gather more digital data about potential customers, they have the ability to use that information to charge different prices to different users or steer different users to different offers.
Ryan Calo, a law professor at the University of Washington, calls this the “mass production of bias,” in which companies use personal data to exploit people’s vulnerability. For example, companies can chip away at consumers’ willpower until they finally give in to making a purchase. Or a computer algorithm can set prices for each individual at exactly the price that is the most he or she is willing to pay for a given product or service.
The credit card companies have started using some of these techniques. In 2010, my colleagues at the Wall Street Journal and I discovered that Capital One was showing different credit cards (with different rates) to different website visitors, based on its guesses about their income and geographic location. The result was that when Thomas Burney, a Colorado building contractor, visited Capital One’s website, he was greeted with offers for a card for people with excellent credit, the “Capital One Platinum Prestige.” By comparison, when Carrie Isaac, a young mother from Colorado Springs, visited the website, she was shown a card described as being for people with “average” credit.
The reason was buried in the computer code. Contained in the 3,748 lines of code that passed between Thomas’s computer and Capital One’s website were the credit card company’s guesses about his income level (“upper- mid”), education (“college graduate”), and his town (“avon”). Capital One had assessed Carrie as having only “midscale” income with “some college” education. A Capital One spokeswoman told us, “Like every marketer, online and off – line, we’re making an educated guess about what we think consumers will like and they are free to choose another product of their liking.”
By 2012, when my team again tested for market manipulation, the techniques had become more widespread and increasingly sophisticated.
We found that credit card companies were still offering different cards to different users. Discover was showing a prominent offer for the “it” card to computers connecting from cities including Denver, Kansas City, and Dallas, but not to people connecting from Scranton, Pennsylvania; Kingsport, Tennessee; and Los Angeles.
But we also found that websites were varying prices based on their guesses about where users were located. In our tests, Lowe’s was selling a refrigerator for $449 to users in Chicago, Los Angeles, and Ashburn, Virginia. But it cost $499 in seven other test cities. Similarly, a 250- foot spool of electrical wiring was displayed at six different prices on Home Depot’s website depending on the user’s location: $70.80 in Ashtabula, Ohio; $72.45 in Erie, Pennsylvania; $75.98 in Olean, New York; and $77.87 in Monticello, New York. Both Lowe’s and Home Depot said the variations were an attempt to match online prices to the closest store.
We found the most comprehensive price differences on the website of the office supply giant Staples, which appears to use data about visitors to guess where they live. It then displays different prices to different users based on its estimate of their geographic location. The end result: when Trude Frizzell logged on to Staples.com from her work computer in Bergheim, Texas, she saw a Swingline stapler listed for sale for $14.29. Just a few miles away, in Bourne, Kim Wamble saw the same stapler listed on the same website for $15.79. The difference was not due to shipping costs, which are calculated after purchasing the item. Rather, the prices seem to reflect how far Staples believes the user lives from a competitor’s store.
Staples confirmed that it varies prices by a number of factors but declined to be specific.
It’s not illegal to charge different prices to different users, as long it is not based on race or other sensitive information that could constitute redlining. But offering price variations to different users can result in unfair results that are unintended. Our tests of the Staples website showed that areas with higher average income were more likely to receive discounted prices than lower- income areas.