Big data and artificial intelligence have made it easier than ever for corporations to relentlessly fine-tune the cost of individual products. However, a new economic study warns that deploying algorithms to create highly granular, heavily differentiated pricing is actively destroying profit margins.
According to new research published in the journal Marketing Science, the conventional corporate wisdom — that sharper data segmentation naturally yields better profits — is fundamentally flawed because it ignores a powerful quirk of human psychology.
The study reveals that rather than relying on complex AI price optimisation across a product line, companies are far more profitable when they offer fewer, strategically chosen price points.
The power of class pricing
The research, authored by Zuhui Xiao from the University of Wisconsin-Milwaukee and published by the Institute for Operations Research and the Management Sciences (INFORMS), investigates a widespread market phenomenon known as “class pricing”.
Class pricing is the practice of assigning a very small number of distinct price points to a massive assortment of related products. Consumers encounter this daily: a pub might offer dozens of draught beers but charge only three distinct prices across the entire menu. Similarly, a supermarket aisle might feature hundreds of unique products, but group them into just a dozen shelf prices.
While this was historically done for operational simplicity, Xiao’s research proves that the strategy’s true power lies in manipulating consumer expectations.
The trap of loss aversion
Consumers do not evaluate a price tag in isolation. Instead, they scan the products in front of them, form a baseline price expectation, and then compare what they are being asked to pay against nearby alternatives.
This comparative behaviour triggers a well-documented psychological phenomenon known as “loss aversion”. Humans are inherently more sensitive to perceived losses than to equivalent gains. In a retail environment, paying more than expected is deeply felt as a psychological loss, while bagging a cheaper item is only mildly registered as a gain.
When artificial intelligence algorithms assign highly granular, individualised prices to a line of products, it forces the consumer to directly compare every single item.
“When firms introduce more granular pricing, it triggers consumers’ direct comparison of prices,” Xiao explained. “Consumers perceive higher-priced items as losses relative to cheaper alternatives and tend to resent higher prices more than they reward lower ones. As a result, the price disadvantage of higher-priced items is psychologically amplified, making them look worse than the underlying price difference alone would suggest.”
A fatal profit asymmetry
Because AI pricing models amplify this psychological disadvantage, companies suddenly find that they cannot convince consumers to pay top dollar for premium items — even if those items genuinely boast better taste, higher quality, or greater prestige.
Simultaneously, the algorithm forces the firm to keep its lower-tier products incredibly cheap just to maintain baseline demand. This creates a fatal asymmetry: the business is sacrificing significant profit on budget items, but the consumer’s psychological resentment means the firm cannot recoup those losses on premium items.
“This asymmetry can reduce consumers’ total willingness to pay across the assortment and outweigh the benefits of differentiating prices based on cost or value,” Xiao said. “That is why adding more price points can actually backfire.”
The findings serve as a stark warning to retail giants and tech companies rushing to implement AI-driven dynamic pricing models without considering human behaviour.
“Even with advanced technologies, firms should be cautious,” Xiao said. “More pricing flexibility does not necessarily translate into higher profits. In many cases, simpler pricing structures are more effective.”
