In her 1933 book, “The Economics of Imperfect Competition,” Joan Robinson introduced the economic theory of pricing discrimination. She essentially structured a theory around a practice that had been practiced before modern currency versions were introduced. We experience this daily when buying consumer goods and household items. It is not always bad, though. For example, when transportation companies charge more during peak seasons in some countries, they also offer subsidized tickets during slow seasons.

Traditionally, price discrimination has been powered by analytics, where demand and supply dynamics have been known better (obviously, again, using data and analytics). The yield management pricing approach of airline tickets is an example. Even for consumer goods, we sometimes see price gouging (like toys between Thanksgiving and Christmas). While pricing methodologies have become more data-driven, the number of industries and markets where price discrimination has been leveraged has remained relatively the same over decades.

Price discrimination can be leveraged across many more industries than currently practiced. But as you can imagine, the underlying foundational knowledge of price discrimination is understanding demand and supply dynamics, combining them with consumer behavior data and historical pricing strategy performances. A good mix of quantitative and qualitative analytics is needed, and these two need to be merged as well. 

With the current approaches of piecemeal manual analytics we perform, the ability to expand price discrimination, leveraging the approaches that were used to establish price discrimination in industries where it is practiced, will take ages. This is because the current price discrimination in many industries results from observation and analysis over decades to fine-tune the process to an optimal point. Not that there are no opportunities to enhance the process significantly, but the process works well enough in these industries’ current state.

And this is where AI can help accelerate that “learning.” What took decades to fine-tune in other industries can be established relatively much faster in different industries. If you think deeply, you can list at least five sectors where price discrimination is not leveraged extensively, but there is an opportunity to leverage it. Give me any of these industries, and I can provide a roadmap of how AI can help learn what was learned over decades using traditional approaches. 

While leveraging AI to enhance current sub-optimal approaches to analytics is necessary, exploring expanding the current constraints of capabilities leveraging AI is also essential.