Dynamically pricing the Warriors

Back in the early 90s, I was in graduate school in the Bay Area and playing in a regular faculty-PhD student basketball game. At the time, the Golden State Warriors held a high draft pick and while waiting for everyone to show up for our game, there was an animated discussion about what the Warriors would. Our conclusion was that they would screw it up. After all, they had recently traded away Mitch Richmond and had previously had traded away Robert Parish so they could get their hands on Joe Barry Carroll. Twenty years on, the Warriors basketball  performance hasn’t gotten a lot better. Sure, they are currently sixth in the West and would make the playoffs if the season ended today. But that would only be their second playoff appearance since I left the Bay Area many years ago.

That long history of mediocrity makes the team’s business performance all the more remarkable. In each of the last eight seasons, they their average home attendance has been over 18,000 — giving them one of the highest average attendance in the NBA. How have they managed to do this? They used lots of data in order to make sure that seats don’t go unsold while also making sure that they don’t give out unnecessary discounts (Warriors Go On Offense to Fill Seats, Wall Street Journal, Feb 6).

The team looks at data generated by Ticketmaster as well as resale-market sites like StubHub, and pores over weather forecasts and ticket sales from competing entertainment in the Bay Area.

The Warriors examine, for example, how much ticket buyers historically paid for a Tuesday night contest against the Houston Rockets versus a Friday night game against the Rockets. This helps the team predict how much it can charge for tickets without curbing demand. In some scenarios, price-points can change hourly, part of a growing sports-business practice adapted from airline bookings known as “dynamic pricing.”

The Warriors also use data when planning to move the last unsold seats for a game with an online offer. The team will take its 200,000-person email list and break it into chunks, testing different times, different subject lines and different links in the body of the message to gauge what brings the quickest response.

“We’ll send 100,000 messages 10 minutes after a victory, and 100,000 the next day,” says Mr. Schneider, 33 years old, now in his 11th season with the Warriors. Data on response time, tickets sold and the price point for each ticket tier will determine how the team tailors future pitches.

Pricing sports tickets dynamically is not exactly news anymore as many teams have started to do this. (See, for example, here.) What’s interesting here is the amount and kind of data that the Warriors are using. For example, the article reports that they adjust prices based on whether the star player of an opponent is injured or not. Intuitively, it is an obvious thing to consider, but it is also a difficult effect to measure since, for example, when a player gets hurt is going to matter. If the star gets hurt the night before, it presumably matters less than when a star gets hurt two weeks before. In the former case, most tickets are presumably already sold while two weeks out gives the market a chance to adjust.

The article also reports that scheduling is another consideration. That is, the Warriors try to build a schedule that is easy to sell. That might mean avoiding going head-to-head with football teams early in the season while also keeping weeknight games to a minimum. There are obviously limits on this. Almost every team in the league will have the same preferences and it can’t be the case that everyone plays just road games during the week.

A final point. The Warriors currently have two-thirds of their capacity committed to season tickets. That means that all the work to fine tune prices is aimed at a relative sliver of their seats. One of the challenges, I think, with dynamic pricing for sports teams is how this interacts with season ticket pricing. This is a component of the market for which there is no good analog in, say, the airline industry. Season ticket holders are in some ways a team’s best customer and it is not clear to me whether dynamically pricing tickets in the season helps or hurts season ticket holders. On the one hand, if the team commits to not undercutting the price season ticket holders paid for a game, then holders benefit.  If a game proves to be unexpectedly popular, then the “official” price rises and a season ticket holder had a better opportunity to sell her tickets if she can’t make the game. The same would not be true if the team were willing to aggressively promote to move unsold tickets. Now the season ticket holder will be undercut by the team if she wants to sell her tickets. Said another way, dynamic pricing probably benefits  season ticket holders if the team is doing well but likely hurts when the team does poorly. At least this year, the Warriors are winning.