Have you ever been to a bazaar, where a hawker offers something at a special price, “just for you?” The world of online retail is getting ready to take the practice to a whole new level — and it doesn’t take much speculation to see how things could go terribly wrong.
Let’s start offline. There are tons of rules about what an old-fashioned retail store can do with prices. Although they vary by state, they tend to impose a reasonably sound standard: You can’t change the price depending on the customer. Stores in different places can charge different prices (as Starbucks does), but in a given store the display price applies to everyone — at least if you don’t count coupons, membership deals or bulk orders.
Going online is more like looking at lots of stores at once. You can easily move from store to store and choose where to buy, which sounds great — as if, behind your veil of internet anonymity, you can get the best price by comparing the offerings of multiple retailers.
The reality is more complicated. E-commerce companies are collecting a lot of information about their customers. You often have to be logged in to get the best deal, especially if you include shipping, which means you’re sharing your entire purchase history. Then there are the membership deals, various kinds of online codes and coupons, all of which conspire to keep you loyal to a single company and provide more data.
Increasingly, companies are finding ways to use those data to steer people toward certain types of goods and services, or even to tailor prices to how much they think someone can pay. Travel sites show fancier hotels to Mac users, auto insurance companies charge more to customers who are less likely to comparison shop, payday lenders focus on people whose search queries show signs of desperation.
Some kinds of price discrimination prey upon the poor and desperate, but plenty of it targets the wealthy, particularly if they are careless or strapped for time. By construction, pricing models follow the money. It’s troubling to think about how that might ultimately affect vulnerable populations, such as older folks on the cusp of senility.
If that sounds bad, here’s a cynical 10-year vision of how things could get worse. Suppose, for example, that companies shift toward using selfies instead of passwords for identification (for example, Amazon has a patent to that effect). Photos may be harder to fake — particularly if they’re in real time and ask you, say, to touch your ear with your left hand — but they will also provide very current information on your health and mood. So a retailer might know if you’re drunk, unhappy or
ovulating. Godiva chocolates, anyone?
Now consider in-home personal assistants, such as Amazon’s Alexa or Google Home. In principle, they could be capable of picking up nuanced inflections in your voice that provide clues to your state of mind. They could then offer high-priced plane tickets to a hurried traveler, a Weight Watchers shake to a person on a diet or a shopping spree to someone feeling reckless or desperate. Utter so much as a whisper revealing a fleeting desire, and you could find yourself inundated with offers of goods and services you didn’t even know you wanted, at prices perfectly tuned to the balance in your bank account.
E-commerce companies already pay to gain access to people’s attention online, ranking customers by their perceived value. So it’s easy to imagine a world in which smart devices gather real-time information on our preferences, vulnerabilities and finances, sending it directly to data collection warehouses that then sell it to retailers, which use it to target customers precisely at the moment when they’re likely to pay the highest price. When we’ve all been separated into our own purchasing universes, the quaint price tags of the retail store may be no more than a distant memory.
Cathy O’Neil is a mathematician who has worked as a professor, hedge-fund analyst and data scientist. She founded ORCAA, an algorithmic auditing company, and is the
author of “Weapons of Math Destruction”