When shopping online, could you be charged a personalized price based on an AI analysis of your income and habits?

Photo by rupixen / Unsplash

Price Discrimination in the Digital Age: How Companies Maximize Profits Through Personalized Pricing

How much would you pay for your favorite pizza?  Or your dream car?  Frequently, many consumers are able to get things they love for far below the amount they are willing to pay.  Whenever you are able to purchase a good for a lesser amount than you would willingly pay, you experience consumer surplus.  Colloquially, we would refer to this as getting “a deal” on the item.  

Sellers want to eliminate consumer surplus and get you to pay the maximum amount you would willingly pay.  Traditionally, this was an almost impossible task: few, if any, consumers would voluntarily admit that they would pay more for something.  Much of the time, consumers insinuate that they would only buy a good if the price was lower.  But, thanks to modern technology, sellers may have an increasing ability to recognize when a consumer is likely to pay more.

Personalized Pricing Explained

Retail sellers of high volumes, typically seen in venues ranging from convenience stores to supermarkets to big-box stores, usually set the same market price for all consumers.  The price is posted and a steady influx of customers can choose to either accept or reject it.  Changing the price in between customers is rather infeasible, and may even cause consumer outrage.  While the prices may change over time, even on a daily basis, they usually cannot be altered during the day between individual sales.

Personalized pricing occurs more often with periodic wholesale transactions and services.  As the name implies, the seller can set a different price for each customer.  For business-to-business wholesale transactions, where the quantity may be unique for each buyer, there is not necessarily a listed price.  Similarly, services may also require unique characteristics for each job, resulting in individual quotes and bids rather than listed prices.  When the price is based on actual or perceived characteristics of the buyer, rather than characteristics of the product, personalized pricing is being used.

Examples of Personalized Pricing

Historically, price discrimination required market segregation, or the ability to separate consumers into different groups based on willingness and ability to buy.  This was often physical: if consumers visited a certain venue, they could be expected to pay more for goods.  Movie theaters and concerts could charge higher prices for food because visitors could not easily leave and seek lower prices.  While a ticket-possessing viewer might be allowed to leave and re-access the venue, he or she would miss a good chunk of the performance in the time it took to get the lower-cost soft drink or box of candy.

For tickets of virtually any kind, sellers can steadily raise the price as the event approaches.  As the event draws nearer, there are fewer suitable substitutes that depress demand.  Waiting until the last minute to buy an airline ticket or a concert ticket will definitely result in much higher prices, which sellers can charge because they know the buyer has few options for a similar result.  If you need to get across the country tomorrow, you have little choice but to purchase the airline ticket.  If you want to attend the popular band’s concert that same evening, you likely don’t have many other suitable entertainment options.

For individual services, such as contracting work or yard care at someone’s house, the seller can assume the buyer’s financial circumstances and charge more money if the buyer seems wealthy.  While time may not be running short, as it often does with the purchase of tickets, buyers may accept some degree of higher pricing simply because they do not want to spend the time seeking additional bids.  It can take considerable time to determine if one is getting the best deal on a kitchen remodel or a lawn care service, which represents an opportunity cost of trying to avoid market segregation.  You may get a lower price, but at the cost of your own revenue or productivity.

The Economics Behind Personal Pricing

Raise Prices to Ration Scarce Goods

When people are buying very quickly, sellers can take advantage of surge pricing.  They know that, based on timing, more consumers than usual are demanding the item.  Some fast food restaurants raise the prices of popular items during peak consumption times, such as the “lunch rush.”  Airlines raise the prices of tickets during high-travel periods, such as holidays.  They know consumer demand has increased during these times and can still sell all of their available units while raising prices.

Lower Prices to Fill Capacity

Personalized pricing isn’t always harmful to the consumer; it can benefit those who are flexible and able to “fill in” gaps in a sellers’ schedule.  This type of personalized pricing works well when the marginal cost of offering a good or service is low.  This typically occurs when most of the costs of a business are fixed rather than variable.  Examples include events, lodging, and transportation: most of the costs are incurred regardless of how many customers fill the space.  Until full capacity is reached, the marginal cost of each additional unit (i.e., ticket or seat) sold is very low.  Thus, sellers can lower their price to sell all possible tickets and maximize revenue.  

New Technology:  AI Can Help Sellers Engage in Personalized Pricing

Both to take advantage of surge pricing and to sell excess capacity, firms can utilize new AI programs to process tremendous amounts of consumer data and reveal which potential customers are best to target.  AI can also reveal information about customers’ income and wealth, providing insight on potential ability and willingness to pay.  AI can also be used to track and compile potential customers’ previous purchases of similar substitutes and complements.  Profiles can be created for both current customers and potential customers, allowing firms to maximize their advertising dollars by targeting ads to tailored groups.   

Critics may feel that this heavy use of AI is unethical, as it may violate consumers’ privacy.  This is especially true if any of the information used by AI software does not come from publicly available sources.  With so much data stored on the Internet, it may be difficult to draw the line between what information is considered public and what information is considered private.  Some consumers post freely about their purchases and tastes and preferences, while others do not.  This can make firms’ use of AI to target customers controversial: some consumers may appreciate ads that deliver information on goods and services they are likely to enjoy, while others may feel spied upon and fear that their personal information is being accessed.

Could Loyal Customers be Charged Different Prices Based on Usage?

AI software could be used to compile a “loyalty score” for each current customer, which could then be used to adjust pricing.  This would be a more sophisticated and dynamic form of customer loyalty programs, such as those using membership cards to give customers discounts on purchases.  More frequent purchasers could be granted additional discounts to encourage continued loyalty…but they could also be charged higher prices due to the likelihood that they will not change their purchasing habits.  When purchases are made on a website or through an app, the customer would be unaware in the moment whether he or she was being charged differently than other consumers.

As computing power and AI algorithms improve with time, each individual customer could theoretically be charged a different price when shopping online through their user profile or IP address.  This would result, controversially, in firms being able to charge the maximum possible amount for each sale, potentially eliminating consumer surplus.  On a macroeconomic scale, this might be harmful over time as consumers come to realize that they will never again get a “deal” on a purchase.  If that is the case, the incentive to be more productive to earn a higher income is reduced - consumers will come to realize that any increase in income will quickly be neutralized by being charged higher prices.  Thus, AI and personalized pricing could mean nobody’s real income ever increases; an alarming possibility!