Will Pricing Algorithms Spell the End of the Fair Market Price?

I hate haggling. I spent a lot of my life living and working in countries across the Asia-Pacific, where open negotiation for purchases is common, and I always hated doing it. Whenever I visit places where haggling is the norm, I usually end up buying what I need from higher-end stores—likely spending more than I need to—just so that I can pay a fixed price and avoid the anxiety of haggling. I am not good at it. I feel awkward. I don’t like the uncertainty or the confrontation, and I especially don’t like the feeling that I’m being “ripped off” and paying more than I should or more than the next person would.

Given this, you can imagine how horrified I am at the prospect of personalized pricing algorithms, a type of algorithm not yet widely in use but nevertheless on the radar of organizations like the OECD (an intergovernmental organization that advises on economic policy), who has included it among the “dark commercial patterns” and the anti-competitive practices that they monitor.

This type of pricing algorithm uses data collected about a consumer (such as their location, browsing history, or purchasing behavior) to estimate how much that individual is likely to pay for a product or service, and then it adjusts the price presented to that person accordingly, to extract the highest value from the transaction.

My husband, who negotiates for a living (and doesn’t mind haggling in the slightest), dismisses my agitation about pricing algorithms, telling me they are simply another manifestation of market dynamics— that companies have always set prices according to predicted consumer behavior and market fluctuations. But I disagree. 

Yes, companies have long been in the business of segmenting and researching audiences to understand price thresholds and to optimize prices in response to the market—I worked in advertising for fifteen years and have done this work myself—but these new, highly sophisticated pricing algorithms represent a fundamental shift away from traditional market practice, with potentially profound effects.

The importance of shared knowledge in a free market

In a traditional free market economy, in a commercial negotiation or transaction, there is some symmetry of information. If both sides have roughly comparable knowledge about the factors involved in the transaction (price ranges, alternatives, norms, constraints) then neither side can dominate completely. One party may have more leverage, but it is limited—you know what others are paying, you can find out the market range, you know you can walk away, the seller knows you have alternatives, etc. With personalized pricing algorithms, however, information becomes highly asymmetrical—one party is effectively negotiating blindly so it has very little leverage and is potentially subject to systematic price discrimination.

To demonstrate, let’s look at two different scenarios. In the first, I walk into a market stall whilst on vacation with a view to buying something. The vendor sizes me up. He knows that tourists often pay more so he will start with a high price. The vendor knows more about his costs and profit margins, but I know more about my preferences, tolerances, and finances. And both of us know the cultural “signals” of negotiation—the first price is never accepted, walking out means resistance, handling the object a lot means emotional investment, etc. As the negotiation proceeds, it becomes a guessing game about our respective price elasticity until we either agree on a price or abandon the negotiation.

Now, imagine I walk into that same market stall under very different conditions. Imagine the vendor already knows where I live and what I earn. Imagine he knows what airline I flew with and whether I am staying in a luxury hotel. Imagine he knows how much I have paid for similar items in the past, how quickly I abandon online shopping carts when prices rise, how urgently I need the item, whether I am impulsive or cautious, and whether I am tired or stressed. And imagine he knows that I hate confrontation and will tend to accept paying more to avoid prolonging the discomfort of the transaction. What is more, if I decide to walk away, the next market store has the same information and quotes the same price.

Clearly, this second scenario is one where only one side of the transaction has any real power, and the buyer is ripe for exploitation.

In traditional market dynamics, the predictions that vendors make operate within a shared epistemic environment. There is some information asymmetry, but it is navigable. The buyer always has some informational leverage—benchmarking prices, withholding information, walking away. There may be some asymmetry, but it is limited, observable, and manageable.

With the use of personalized pricing algorithms, on the other hand, predictions that vendors make would operate in a sealed or one-sided epistemic environment. Just as the content on a social media feed is optimized to engage an individual based on their behavioral and emotional profile, the price presented would be optimized for that individual’s pricing elasticity. There would be no shared reference price, no ability to benchmark, and little point in trying your luck with a competitor if your “personal pricing profile” follows you across platforms and websites.

This has crucial implications for the idea of a free market. The core idea of a free market is not that it guarantees fairness but that it relies on conditions that make fairness intelligible and contestable—conditions like shared information, open competition, and freedom of choice. “Fair market prices” are formed within this open environment, shaped by buyers and sellers who can compare, question, and walk away. 

Back to haggling briefly: I hate the process, and I hate the feelings it evokes of being treated unfairly or unequally, but it is by rights a fair market practice. If I had the grit to overcome my discomfort around the process, I would potentially be able to secure a fair market price and not feel taken advantage of. This is not the case with personalized pricing. If this became prevalent, every purchase I made would bring all the discomfort of haggling without the agency of being able to meaningfully negotiate.

How likely are we to see these personalized pricing algorithms?

The good news is that few companies seem to be using them for now (from what can be gathered with the patchy research available). Of those that use pricing algorithms at all, most use them only for price setting within the context of competitive and market fluctuation monitoring.

The bad news, at least as I see it, is that given how successfully personalized ad-targeting has transformed the social media and search industries, the trajectory towards companies applying these same practices in their pricing seems a clear and obvious one. 

Why? Advertising driven platforms generate hundreds of billions of dollars a year by predicting and influencing behavior. It seems unlikely that advertisers themselves won’t be tempted to maximize their revenues using the same approach, should the opportunity present itself. The economic incentives are almost irresistible.

What can be done?

The idea of personalized pricing raises almost limitless questions around numerous philosophical disciplines—equality, exploitation, moral responsibility, privacy, consumer ethics, epistemic injustice, and economic theory, to name just a few. There is a lot we should be exploring, researching, and analyzing, from immediate effects to far-reaching ones. But the first step, as always, is to simply be aware of these technological systems, learn about them, and follow their development. These innovations affect us all.

And although, as is inevitable, regulatory bodies are lagging behind the speed and scale of deployment of these types of technologies, it is heartening to see that these potentially harmful commercial practices are at least within their sights and that policy makers around the world are watching with concern.

I encourage you to read more about personalized pricing algorithms and other “dark commercial practices.” Here are some sources that offer very comprehensive and clear overviews: the OECD Dark Commercial Patterns Paper and the OECD Algorithmic Competition Paper.