What would happen if sellers could use AI to determine how much you are willing and able to pay for every possible product?

Photo by Anna Dziubinska / Unsplash

The Economics of Hyper-Personalization: When Every Consumer Gets a Different Price

It’s past dinnertime and you’re starving!  How much are you willing to pay for a hot, fresh pizza from your favorite pizzeria?  Probably a lot.  On another day, however, you just got done snacking during a work party and feel bloated.  When your friend suggests getting a pizza from the same pizzeria, are you willing to pay the same price as when you were starving?  Our demand often fluctuates based on many factors, ranging from emotions to physical comfort level to the amount of money in our accounts.  

Traditional Market Economics and Equilibrium Price

Although our individual demand can fluctuate, we have traditionally been able to purchase goods and services for the same price as everyone else in our same circumstance.  When we go to the supermarket, we get the same price as all other consumers there at the same time.  This equilibrium price, also known as the market price or market-clearing price, is determined by the aggregate level of demand of all consumers and the supply curve of the seller.  At this price, the goods and services are sold and, ideally, the firm makes economic profit.

Price Discrimination

But what if firms could charge higher prices to customers who are likely to pay more?  Outside of auctions, price discrimination is difficult.  Firms that engage in price discrimination must be able to segment the market and prevent high-paying customers from leaving their segment to enjoy lower prices.  This is difficult due to the ease of accessing information, especially thanks to widespread Internet access.  Consumers with smartphones can quickly search for the lowest cost deal around.

However, if something is meant to be consumed immediately and cannot be easily transported, customers may have to accept some degree of price discrimination.  Utilities, for example, tend to be geographic monopolies and customers cannot feasibly bring in lower-cost utilities from other markets.  Consumers who are geographically isolated, such as at a sporting event or concert, typically pay much higher rates for refreshments due to the infeasibility of bringing in lower-cost food or leaving the venue and returning during the entertainment.

AI and Price Discrimination

Current price discrimination treats all similarly-situated customers the same, typically based on location and/or time.  The seller does not know how each individual differs in terms of willingness to pay.  Even if the seller could reasonably assume how much each individual would pay, this would likely take too much time to put to fruition.  However, the use of AI could allow sellers to change prices quickly enough to engage in individualized price discrimination by analyzing customers’ information in real time.

If an online shopper creates an account with a seller, his or her information could be rapidly analyzed by AI to estimate income, level of education, demographics, and political beliefs based on publicly-available information or public social media posts.  The more information a customer has willingly released online, the more accurately he or she would be profiled by AI to determine the likely level of demand for each product.  For example, a high-income white male conservative between age 30 and 45 would be profiled as having elevated demand for certain products, thus allowing the seller to charge more than the equilibrium price.

Economics of Perfect Price Discrimination

Efficiency Maintained

Firms are highly motivated to engage in price discrimination if they can, as charging each customer the highest price they are willing to pay leads to greater profit.  Because each customer is being charged their maximum price, there is no leftward shift in the supply curve and no decrease in quantity sold.  All consumer surplus disappears, but there is no deadweight loss - the economy is still operating efficiently.

Productivity Incentives Reduced

Although all mutually beneficial transactions still occur, with no loss in output/consumption, the fact that all consumers are paying the most they are willing to pay reduces incentives to increase productivity.  Because the firm can engage in perfect price discrimination, it has little incentive to improve its productivity and be able to reduce its costs of production.  Why work harder to be able to offer goods at lower prices when you are making economic profit through price discrimination?

Consumers also have little incentive to be more productive and increase their income.  If they know that firms are successfully engaging in perfect price discrimination, they know that any increases in real income will be neutralized by having to pay higher prices.  Therefore, it makes sense for wage-earners to opt for more leisure instead of more labor.  Being profiled as “low income” no longer denies one access to most goods; they can still get them at lower prices thanks to firms’ ability to charge higher prices to wealthy customers and make up any losses from poor consumers.

Controversies of Perfect Price Discrimination

Violation of Privacy

Attempting to profile individuals for price discrimination, as opposed to dynamic pricing based on charging all customers differently based on location or time, involves analyzing their data.  This results in concerns about violations of privacy, especially if sellers use images of prospective customers as part of their algorithms.  As sellers compile the data of prospective customers, what happens to this data after the sale?  In theory, sellers could profit by selling customers’ consumption profiles to other firms, sharing data that customers would rather keep private.

Use of Harmful Stereotypes

When setting up algorithms to analyze the data and images of prospective customers, what assumptions will be included in the coding?  Many firms may perpetuate harmful stereotypes by pre-loading algorithms with consumption assumptions based on race, gender, age, clothing choices, and body type.  This could spread these harmful stereotypes throughout society by formalizing them in a financial and economic sense.  Some sellers may also intentionally adjust their algorithms to charge certain groups of customers more money due to bias or prejudice, hoping to dissuade certain types of customers.

Disincentives to be Fiscally Responsible

Perfect price discrimination charges each customer the highest amount he or she is willing to pay, thus reducing incentives to try to earn higher wages or engage in responsible budgeting.  In theory, no financial maneuvering would net a customer the ability to purchase more goods - all increases in purchasing power would be neutralized by perfect price discrimination.  In the long run, this could result in societal stagnation as nobody felt the need to work harder, innovate, or engage in responsible budgeting.