Why Does it Cost $11.5M to Attend the FIFA World Cup Final?

Dynamic pricing expert and Columbia Engineering professor Adam Elmachtoub has a tip for fans looking to buy World Cup tickets. “Buying very early or very last minute are often the best ways to get a good deal,” he said.

Prices, however, seem to have run amok ahead of the much-anticipated final on Sunday. The Associated Press reported that tickets now range from several thousand dollars to a colossal $11,499,998.85.

In this Q&A, Associate Professor Elmachtoub explains the concept of dynamic pricing and whether there is room to make it fair.

Image

Help us understand dynamic pricing and why it’s used.

Dynamic pricing is a concept where the price of a product or service changes over time. There are many reasons companies do dynamic pricing:

• Learning: Companies change prices to gather data and learn what price point is best
• Inventory: As inventory becomes scarce, prices go up. On the flip side, if there is too much inventory, prices might go down
• Competition: If a competitor changes their price, then the company might update their prices accordingly.
• Seasonality/Time-of-day: Willingness-to-pay and demand for products/services changes over the course of a day, week or season.

World Cup ticket prices have drawn an enormous amount of public backlash. Can you help us understand why the tickets are so expensive? Tickets to the final match have now reached $11.5 million.

This is most likely due to the market price. In essence, there is a giant auction happening for all tickets where the people willing to pay the most will get the ticket. If you are a ticket owner, you have the right to sell your ticket for as much as possible just as you would if you wanted to sell your car or property. FIFA implemented some form of dynamic pricing that adjusts tickets up and down based on demand. But it’s hard to know the details, however.

How can we make World Cup tickets fairer? What would be your fair algorithm for such an important sports tournament?

First and foremost, I would make the algorithm completely transparent and open-source. Most people are okay with dynamic pricing, as long as they understand the rules.

Second, countries are allocated some fraction of the tickets. If it’s important that “anyone” can come, then they should just give them away at a very low price through some sort of lottery. There are various forms of this that have happened.

Why do you think dynamic pricing is often misunderstood and often draws public scrutiny?

Dynamic pricing is often just thought of as real-time price gouging where companies change the prices in real-time to extract as much revenue as possible. This is not to be confused with personalized pricing, where prices are customized to individual user characteristics (i.e., a mortgage loan or an Uber ride that depends on your origin and destination). 

Dynamic pricing is largely confused also with high prices. For instance, World Cup and Taylor Swift concert tickets have skyrocketed. However, much of that is due to individuals reselling their tickets at market prices, which tend to be extremely high.

There is also a confusion with something called variable pricing, which is where prices are not real-time but do change over time (like the off-peak pricing) and what FIFA claims to be doing. This, in my opinion, is a special case of dynamic pricing. Real-time dynamic pricing is often opaque and hard to optimize for as a consumer.

Any tips you have for fans who are trying to get reasonably priced tickets?

Buying very early or very last minute are often the best ways to get a good deal on tickets for entertainment such as the World Cup, but there is never a guarantee. Sometimes buying early can be bad for an event that is not sold out, while sometimes buying late might be the worst-case scenario if you have a large group or the event is extremely popular.

You’ve proposed an approach that would make dynamic pricing fairer. How do you engineer fairness into a dynamic pricing algorithm?

One stream of research has focused on how to make personalized pricing fairer by incorporating fairness constraints. In another line of research, we consider how static pricing might compare to dynamic pricing. We prove that in many settings, the value of dynamic pricing might actually be small enough where sellers should reconsider its value. 

In order to ensure dynamic pricing is fair, one first needs to understand what fairness even means. For instance, if everyone is charged an extremely high price, but always the same, is that fair? Or if the prices are changing dynamically, but according to a published, pre-determined timetable, is that fair?

When prices are changed dynamically without any transparency, then one should check that the average price paid across different demographics is similar. Since companies cannot use protected attributes like race or gender to do pricing, they can still price differently according to time-of-day where perhaps different demographics do their shopping.

Policymakers by and large don’t seem to support dynamic pricing given public sentiment. Some attorneys general are even investigating these practices. Is an outright ban on dynamic pricing the right thing to do? What could be the unintended consequences?

Banning dynamic pricing would destroy the travel industry, where profits are very slim and airlines are always going out of business. They rely on dynamic pricing to collect more revenue from business travelers who tend to book airfare closer to travel than leisure passengers, for example. They also use dynamic pricing to adapt to market conditions and world events. 

However, I do believe more transparency and explanation is needed. People, rightfully so, are more concerned that AI and algorithms are exploiting them. In reality, they should theoretically bring a net benefit to society. My research aims to bridge this gap where we allow sellers to use AI and dynamic pricing, but still ensure the overall policies are fair.

—

Adam Elmachtoub is an associate professor of industrial engineering and operations research at Columbia Engineering. He is also a member of Columbia’s Data Science Institute and the DAPLab. His research spans two major themes: (i) designing machine learning and personalization methods to make informed decisions in industries such as retail, logistics, and travel (ii) new models and algorithms for revenue and supply chain management in modern e-commerce and service systems.