Exploitative algorithms are using tricks as old as haggling at the bazaar A few years ago I received a text message from my mobile phone network informing me of the good news that I was already on the cheapest possible tariff. They should have let sleeping dogs lie: I called their bluff, and within minutes they offered a cheaper one. Another time, I quit only to receive phone calls pleading for forgiveness and offering me an iPad if only I’d come back. It was like being in an emotionally abusive relationship with Santa Claus. Nobody wants to feel that they are being taken for a fool. It is hardly surprising, then, that the UK business secretary Greg Clark has made some noise about his plans to scrutinise how firms may use big data or other digital tools to produce “abusive
Tim Harford considers the following as important: Undercover Economist
This could be interesting, too:
Tim Harford writes Technology can be the friend of creativity
Tim Harford writes Stop sniping at central banks and set clear targets
Tim Harford writes Why good forecasters become better people
Tim Harford writes The economist’s guide to the perfect Christmas
Exploitative algorithms are using tricks as old as haggling at the bazaar
A few years ago I received a text message from my mobile phone network informing me of the good news that I was already on the cheapest possible tariff. They should have let sleeping dogs lie: I called their bluff, and within minutes they offered a cheaper one. Another time, I quit only to receive phone calls pleading for forgiveness and offering me an iPad if only I’d come back. It was like being in an emotionally abusive relationship with Santa Claus.
Nobody wants to feel that they are being taken for a fool. It is hardly surprising, then, that the UK business secretary Greg Clark has made some noise about his plans to scrutinise how firms may use big data or other digital tools to produce “abusive outcomes” such as exploiting loyal customers.
Another review is under way courtesy of the UK’s Civil Aviation Authority into how budget airlines “use algorithms” to seat families separately if they don’t pay extra for assigned seats. You and I don’t want to be on the sharp end of an exploitative algorithm, do we?
There seems no harm in having a good hard think about how well competition works in a data-rich age. Customers have new tools at their disposal to find the best deals; companies, in response, can pick off lucrative customers like stray wildebeest, offering hidden discounts to some, even targeting adverts and offers by sex or race.
Yet the striking thing about the concerns of consumer champions is that for all the digital window-dressing, this struggle is as old as haggling at the bazaar.
When companies take time and trouble to make their wares less attractive, we call this “product sabotage”. Printers may come in a high-cost professional version and a lower-cost home version with a chip to slow it down. “Value” supermarket pasta or rice is packaged to look like famine relief. And airlines may split families who do not pay extra — a practice that hardly requires a mysterious “algorithm”.
In each case, the company is seeking a premium from premium customers while grasping for volume by offering low prices to the masses. In order to achieve both goals, it may need to damage the mass-market offering. If the cheap product is insufficiently dreadful, the risk is that even wealthy customers may buy it.
The economist Jules-Emile Dupuit spotted an example in 19th-century France. “It is not because of the few thousand francs which would have to be spent to put a roof over the third-class carriage . . . that some company or other has open carriages,” he wrote of the railways. “What the company is trying to do is prevent the passengers who can pay the second-class fee from travelling third class; it hits the poor, not because it wants to hurt them, but to frighten the rich.”
The problem, then, is more than 150 years old. And it is not clear that the situation would be improved by insisting on equal treatment for all passengers. The railway company (or airline) might then offer only the first-class service at the first-class prices, perhaps even higher.
This is aggravating, no doubt. But the root of the problem is that the company has some market power, which allows it to squeeze customers and raise prices. The product sabotage is the symptom — and not necessarily a harmful one.
What of the idea that loyal customers are exploited rather than rewarded? That was my experience with the phone companies, but the infuriating practice is, again, not new. Every time I shave I can praise King Camp Gillette for inventing the disposable razor blade, and curse him for embracing the pricing model of cheap razor, expensive blades. What is that, if not a penalty for loyalty?
In truth the word “loyalty” leads us astray here. Any profit-seeking company will want to exploit customers who never walk away, so considerable effort is devoted both to identifying those customers and to inducing them not to look elsewhere.
“Loyalty cards”, whether an airline gold card or a rubber-stamped bit of cardboard from your local espresso bar, are designed to persuade high-volume, high-value customers both to identify themselves and to stick around. The result is a less competitive market in which everyone pays a higher price.
It is possible that in the initial scramble to sign up new customers, companies reward them so lavishly as to compensate them in advance for years or decades of locked-in high prices. But it’s not likely.
Who loses out from such behaviour? Understandably, we worry about “vulnerable” consumers. But for the companies, their target is clear: they will try to price-gouge the customers most likely to pay. Often, those customers will be rich and busy, while the ones who enjoy the bargains will be poorer and have more time to shop around. That is no calamity. When the victim is a lonely octogenarian in the early stages of dementia, the cat-and-mouse game between producer and consumer takes on a cruel and tragic edge.
Regulators are right to be vigilant. Still — the very fact that such tricks are as old as commerce itself suggest that we will not succeed in stamping them out. Buyer, beware.
Written for and first published in the Financial Times on 5 October 2018.