Brandeis Stories

By Julian Cardillo ’14
August 12, 2025

Long before AI became a mainstream force transforming industries like commerce, Brandeis School of Business and Economics professor Benjamin Shiller was exploring the business strategy of personalized pricing and considering how surges in customer data might reshape the relationship between companies and consumers.

AI pricing, or personalized pricing, uses algorithms to analyze data, including purchase history, browsing behavior, location and even device type, to predict what an individual customer is likely willing to pay. The system then dynamically adjusts prices in real time to maximize sales and profits.

Shiller, a pioneer in the economics of AI-driven pricing, discusses the opportunities, challenges and consumer impacts of this rapidly evolving practice.

What sparked your interest in AI-driven personalized pricing?

I’ve been interested in this for a long time. Back in 2014, I wrote arguably the first economics paper on AI pricing. At the time, the textbooks described personalized pricing as a theoretical concept — something firms might aspire to but couldn’t realistically do. I thought, with all the data now available, firms could actually estimate individual consumers’ willingness to pay well enough to raise profits.

In my first paper, I found that if Netflix had personalized prices based on detailed web browsing data, it could have increased profits by about 13%. Just using demographics, the profit gain was only around 0.25%. This showed that the new data sources — like web browsing and other consumer data — are extremely useful for personalization.

There’s been some pushback against personalized pricing from a consumer standpoint. What are the main concerns, and how is this different from traditional concerns about data privacy and price discrimination?

The key issue is transparency. I suspect much of AI pricing is happening “underwater,” unseen because firms disguise it well.

The explosion of data collection isn’t new — it’s been growing for 15 years or more — but AI’s computing power and better algorithms now allow firms to use that data more efficiently. Firms are under pressure to integrate AI, and personalized pricing is an obvious way to show they’re doing so.

Personalized pricing isn’t inherently unfair — consider student or senior discounts — but the concern is the scale and intensity. The price differences can be much larger, and it’s unclear who ends up paying more. You might think it targets the wealthy, but it could just as easily exploit consumers with less time or ability to shop around. Imagine a poor single mom working two jobs. The algorithm might realize she doesn’t have time to compare prices as often and charge her more.

Are there any effective consumer protections or policy approaches?

I looked at a few. One idea was to restrict the range of prices firms can offer. But firms adapt: They might reduce deep discounts, which actually raises the lowest price many consumers pay.

Another approach is banning personalized pricing entirely, like China did in 2022. But then you lose the benefits for consumers who might receive lower prices through personalization. And it’s hard to enforce if firms disguise the practice well.

Are we looking at this all wrong? Could personalized pricing actually benefit consumers?

Yes. The economic theory shows that in markets with multiple competing firms, personalized pricing can intensify competition and actually benefit consumers by offering better deals. It’s a kind of prisoner’s dilemma: Each firm wants to personalize prices unilaterally, but when done collectively, profits may fall and consumers could win.

Some consumers will be worse off, some better off. If the median consumer benefits, personalized pricing could persist politically and economically.

Is personalized pricing here to stay?

I expect some policy will eventually limit it, mostly because of consumer perceptions of fairness, even if actual harm might be less than feared. But firms likely are already adopting AI-based pricing widely, sometimes unintentionally, just by feeding all their data into pricing algorithms that respond to demand signals and individual consumer attributes.

So yes, personalized pricing is likely here to stay, but how it evolves — and how it will be regulated — remains to be seen.