With algorithmic pricing, the cost of goods and services can respond to immediate market conditions and situations. Doctoral student Liying Qiu and researcher Nikhil Malik discuss what this means for consumers.
Dynamic pricing occurs when the price of a product or service fluctuates over time, sometimes even minute by minute. Instead of a fixed sticker price, the price fluctuates with demand, supply, competition, and timing.
Dynamic pricing is not one idea. It is a bundle of goals. Different industries use it for different reasons, even if it all looks like “the price changed again.”
Sometimes prices rise for a simple, familiar reason: there isn’t enough supply to go around. Think of strawberries in winter. If the stock is unusually low, the grocer sets a higher price—hoping no one shopper buys all.
The same happens with electricity and time-of-use pricing. Electricity usage peaks in late afternoon and early evening. There is simply not enough supply. Dynamic prices are meant to push flexible activities, like charging an electric vehicle or running a dryer, into off-peak hours so the grid does not get overloaded.
However, dynamic pricing for rationing is not always appropriate. For luxury goods, many people accept it: a last-minute concert ticket or a limited-edition collector’s item is fair game. For necessities, it can feel wrong. During the early COVID-19 pandemic, for example, masks and other PPE became expensive because demand surged while supply and manufacturing capacity lagged. Dynamic pricing is simply not a good tool for rationing here. Buyers have no practical substitute. The seller simply needs to find more creative ways of rationing rather than dynamic pricing.
In some cases, dynamic pricing can act as a real-time distress call for suppliers to ramp up. One example is the ride-sharing surge model. When it starts raining, a concert ends, or rush hour hits, requests jump. The platform raises prices, so that more drivers decide it’s worth heading out. Drivers who were not working head out and look for fares, and some who were about to stop driving stayed on longer. So the higher price helps get more cars on the road fast.
Dynamic pricing can simply be an artefact of poor guesswork. For airlines and hotels, some people book weeks ahead, others book the same day. An empty airplane seat and an unsold hotel room expire forever once the moment passes. The company needs to predict the number of possible last-minute buyers weeks or months ahead and adjust prices as that prediction changes. This guesswork is difficult; they often get it wrong and have to keep shifting.
This doesn’t always mean that prices rise. When a store has an abundance of ripe fruit, it may mark prices down late in the evening to clear inventory before it spoils. To a consumer, this can feel like a trick: “Why was it higher in the morning?” The grocer may have just made a bad guess in the morning about how many people are in the mood for that fruit.
Prices can change because rivals changed theirs—and the firm doesn’t want to lose shoppers. These show up most in easy-to-compare goods.
The most common example is e-commerce. Large retailers can adjust prices constantly based on competitor moves and inventory. If a competitor drops a price, algorithms can match or slightly undercut it quickly. The goal is often defensive. Avoid losing traffic and market share.
All of the above seem reasonable. However, there are more controversial intentions that cause consumer frustration and distrust.
Dynamic pricing can sort consumers by desperation! Especially, where the product isn’t meaningfully different, but the seller can detect your urgency.
Sellers will never admit to detecting your urgency, so examples are hard to pin down—but many consumers suspect this: if your phone battery is at 5 percent, does the ride-sharing app know that you are “captive” and you don’t have the luxury to wait for options? If you’re trying to book a hotel for tonight, the context alone—late at night, after several failed searches—can make you feel like the system knows you have fewer options.
People with more patience, information, and flexibility pay less. People who are desperate pay more.
Dynamic numbers can simply be used as a tool to create interest for potential buyers. When a platform shows you a number that moves—an estimate, a score, a forecast—you start asking “what is it now?” That uncertainty pulls you into a loop of checking and re-checking, even when you’re not planning to buy. This is the same as getting hooked on stock prices or the weather.
Zestimate-style home valuations provided by Zillow are a good example. Homeowners may have no intention of buying or selling, but a changing estimate turns into a kind of scoreboard: Did my home go up this month? Did it drop? Why? The number feels personally relevant, so people come back to look—spending time on the site, clicking into nearby comparables, and exploring “what I could get,” even if they never list the home. Zillow gets 50+ million monthly website visitors, way more than buyers and sellers in the market. In this sense, a moving estimate isn’t just information; it can quietly convert curiosity into repeated attention.
You will probably not be able to tell which of these six goals the firm is trying to achieve. Even when dynamic pricing occasionally saves you money, it changes how buying feels. It creates decision anxiety. Instead of asking “Is this worth it?” you end up asking “Is this the right moment?” which invites regret either way. Buy now, and you worry the price will drop tomorrow. Wait, and you fear the price will jump. It erodes trust. People are willing to pay different prices when there is a clear reason, such as first-class versus economy. They react differently when the same service seems to cost more simply because the algorithm caught them needing it. And it makes shopping work. In a fixed-price world, you check once and decide. In a dynamic world, you are nudged to monitor, compare, and time your purchases. Price discovery becomes a part-time job, and those without time or tools pay for that disadvantage.
Dynamic pricing will not disappear. It is too profitable, too measurable, and too easy to automate. The real question is what kind of market we want it to create: one where prices help systems run smoothly, or one where every urgent moment becomes an opportunity for the desperation tax.
What should users consider when using generative AI to create or manipulate images of other people? What policies should companies adopt to ensure that their products don't cause harm? Ethicist Derek Leben joins the Tepperspectives podcast to discuss.
Tepper School doctoral student Pim Assavabhokhin speaks with Professor Emily DeJeu about how AI can enhance short-term decision accuracy and transactive memory systems to improve team speed, but over-reliance on technology can create a dependency.
In the latest Tepperspectives Video, Tepper School researchers Taya Cohen and Sofía Rodríguez Chaves look at how ethics and moral character influence responsible AI use in the absence of policy or regulations.