Golden ticket economics, part 1: Next restaurant

By Felix Salmon
January 6, 2012
have a problem with Grant Achatz's pricing strategy at Next, where tickets are sold at a fixed price and are then free to be resold at an enormous markup on the secondary market. " data-share-img="" data-share="twitter,facebook,linkedin,reddit,google" data-share-count="true">

Economists Justin Wolfers and Betsey Stevenson have a problem with Grant Achatz’s pricing strategy at Next, where tickets are sold at a fixed price and are then free to be resold at an enormous markup on the secondary market. The restaurant is very clear why it won’t auction off tickets instead:

You should auction the tickets. Do a reverse, double blind, dutch auction and give the surplus profit to charity.

An auction would set pricing too high for our sense of value for the meal. One of the reasons that we have so many people trying to buy tickets is because we are trying to do something new, different, and delicious for a great price. We may institute more dynamic pricing in the future, but for now the system is fair precisely because it is blind to everyone – anyone who clicks to buy can buy.

But the Wharton economists aren’t convinced.

“It’s democratic in theory, but not in practice,” said Wolfers…

If a person can sell a ticket for $3,000, the true cost of going to the restaurant — what an economist would call the opportunity cost — is $3000, because that’s how much money the person is giving up for the meal.

Bloomberg’s Mark Whitehouse concludes that Next should “consider selling tickets to the highest bidder and giving the extra money to charity” — precisely the course of action which has been explicitly considered and rejected in the restaurant’s FAQ.

Is Next making a mistake here? Do Wolfers and Stevenson have a point?

My feeling is that the restaurant is the smart one, while the economists are being naive.

For one thing, real people don’t think in terms of opportunity cost — especially not when they’re the lucky winners of a restaurant-reservations lottery. Dan Ariely did research on this at Duke University: he found that once Duke students won the lottery giving them the opportunity to buy sought-after tickets to the university’s basketball game, they valued those tickets at ten times more than the students who lost the lottery.

What’s really going on here, I think, is that the vast majority of people who get tickets hold on to them, go to the restaurant, and eat a wonderful meal for which they paid a reasonable sum. And then there’s a tiny number of people who get tickets, and either discover they can’t use them for some reason, or decide that they’re going to try to flip them for profit.

Because that number is tiny, the supply of Next tickets in the secondary market is tiny — and because the secondary-market supply of Next tickets is tiny, the price of those tickets can become astronomically high. But I suspect that the high secondary-market prices for Next tickets are doing a very bad job of increasing supply — that there are people who can’t use their tickets, and there flippers who are always going to put their tickets up for auction if they win, but there are very few people indeed, and possibly zero, who put their tickets up for sale just because of how much money they might fetch.

As a result, the very few datapoints that we do have, with respect to the secondary-market price of Next tickets, tell us almost nothing about the amount of money that Next tickets would go for if they were auctioned. If Next decided to auction off all its tickets, the total supply of Next tickets in variable-price markets would skyrocket. Demand would probably rise too — but I very much doubt you’d ever see $3,000 tickets in a Dutch auction.

What you would see, on the other hand, would be a lot of semi-disgruntled diners worrying about whether they were suffering from the winner’s curse, and feeling much less chuffed about their meal than the current diners who are generally elated about having won the lottery.

The most important thing in being a restaurateur of a high-end establishment is exceeding expectations; if you auction off tickets, then the price of tickets will naturally gravitate to and possibly past the point at which you can’t do that any longer. That’s why Next is right to worry about “our sense of value for the meal” — because the chances are that their sense is going to be your sense too. If they think a meal isn’t worth more than say $200, and they start selling tickets to that meal at $400 apiece, then they’re setting their customers up for disappointment; I can’t imagine Achatz would ever want that.

Do the handful of people who currently buy tickets for $500 or $3,000 walk away disappointed? Maybe not: there’s a good chance those people aren’t particularly price-sensitive. But when you move away from those people and use the market to set prices for all your customers, big dangers lurk. As Alan Vanneman says, markets are largely foreign to the human imagination. And since restaurant-goers are human, we don’t want to upset them with market mechanisms if doing so is unnecessary.

In the past, I’ve advocated auctioning off restaurant meals in certain contexts, but never as the only way of selling tickets. Restaurant-reservation auctions should be rare things, applying to a minority of your total diners. Most of the time, prices should be fixed, and it’s always nice when demand outstrips supply. That’s how successful restaurants have always worked, and it’s hubristic to imagine that there’s an obviously better way.

Update: Next owner Nick Kokonas responds in the comments, happily demolishing a key part of the auction-happy crowd’s argument: it’s untrue that a $100 ticket ever sold for $3,000. In reality, he says,

a TABLE has sold for $ 3,000. That table was a kitchen table for 6 people that with wine pairings, service, and tax was nearly $ 2,500 face value. This was a case where one blogger got it wrong and EVERY news source since has reported it as if a $ 100 ticket sold for $ 3,000. Big difference.

What’s more, Kokonas confirms my suspicion that the overwhelming majority of tickets are not in fact resold: 99% of them are used by the people who manage to buy them, or their family and friends.

But he does add that there will be a Dutch auction for one two-top per night, with all proceeds going to the University of Chicago Cancer Center. I hope it raises lots of money!


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see

$3000 opportunity cost?

Yoram Bauman shot that one a while ago with two Snickers bars :-)

Posted by Frwip | Report as abusive

If markets are conversations, a lot of people wish the other interlocutors would just STFU.

Posted by Curmudgeonly | Report as abusive

How about selling 90% of the tickets at a fixed price, with allocation by lottery, then auction the remaining 10%?

Posted by ABritAbroad | Report as abusive


Your point about endowment effects and expectations/framing/anchoring are well taken. However, over time, isn’t the equilibrium solution a small number of dedicated arbitrageurs (“scalpers” in the vernacular) who buy in the primary and sell in the secondary market? Their facebook page notes that entire days’ allocations are selling out at 3 second clips. An enterprising coder could just set up a script that buys em all and sells at a profit.

Doesn’t it make more sense for the restaurant to capture this surplus rather than a parasitic middleman?

Here we have a couple of reasons why the dynamic is favorable to a Dutch auction: an extremely limited supply (relative to, say, a sports event or concert), and a convenient, low transaction cost method of running the auction (the web, electronic payment systems, edocs, etc.)

Posted by Sunset_Shazz | Report as abusive

1. Enterprising coders did set up scripts — but those have been effectively blocked by both passive (captcha) and active (ip filtering) means.

2. I don’t believe a ‘ticket’ has ever sold for $ 3,000 — a TABLE has sold for $ 3,000. That table was a kitchen table for 6 people that with wine pairings, service, and tax was nearly $ 2,500 face value. This was a case where one blogger got it wrong and EVERY news source since has reported it as if a $ 100 ticket sold for $ 3,000. Big difference.

3. The number of scalped tickets is very low… we can track transfers and to an extent which are scalped and it is less than 1 %. The majority of transfers that are not between family and friends are processed through our site — a few per day only.

4. Most economists I have spoken with on this (and yes, they do call / email me) fail to consider the fact that we have to tightly control the flow of people into the restaurant and that we are not making widgets… it is not a scalable operation. Those that realize this immediately think — raise prices and find price stability through an auction. I completely understand that from the point of view of maximizing utility (in the economic sense). But it would be a PR / customer service DISASTER. Right now we are offering an experience that is perceived as a great value and that ensures that we have a full house every night. As soon as we take that to ‘parity’ we run the risk of living a bright but short existence. We are planning for the long term — and that includes tangential businesses (our iBook series with Apple for example) that are not factored into their plans.

5. For our El Bulli menu we will be running a Dutch Auction for 1 table of 2 per night… with 100% of the proceeds benefiting the University of Chicago Cancer Center where chef Achatz was treated. We will indeed see where El Bulli pricing settles, albeit for charity and for a very limited supply.

6. I am in the process of building a software suite that will allow restaurants, galleries, barbers, theaters, etc. charge variable pricing in both direction… and indeed use systems to easily sell fixed, variable, and auction pricing in a mix based on both supply and demand — while linking through social media seamlessly… and without having to use third-party deal sites to interact with opt-in customers. Next is in the first iteration of that experiment — there is plenty more (and more interesting) models to come.

– nick

Posted by nickkokonas | Report as abusive