tickets

here’s something i wrote and sent to my management when i was a software engineer at live nation in april of 2008, prior to their merger with ticketmaster.

Revolutionizing the Ticketing Industry

How Live Nation could sell out every show, eliminate the economic incentive in ticket scalping, even out spikey site loads, and increase average customer satisfaction, all by utilizing the same economic concepts that drove Google Adwords to legendary success

(See an explanation by Google’s Ivy League economic luminaries here: http://video.google.com/videoplay?docid=126302100433476950)

Game Theory and Economics

William Spencer Vickrey was a Columbia University professor who won the Nobel Prize in economics, three days before his death on October 11, 1996. He authored the seminal 1961 Journal of Finance paper, “Counterspeculation, auctions and competitive sealed tenders.”, which was the first instance of an economist using the tools of game theory to understand auctions. The paper dealt with the underpinnings of auction theory, and suggested a simple yet revolutionary auction system called the “second-price auction” (also called the Vickrey auction, after him), in which the highest bidder wins, but only pays what the second-highest bidder bid. Here is why that is so brilliant.

Imagine we were selling tickets by asking customers to place a bid, knowing that the n available tickets for any given seating tier would go to the n highest bidders. A potential buyer for whom a ticket is worth $50 dollars could artificially bid, say, $40. If he doesn’t get a ticket, he gains nothing, but also spends nothing. His expected value is zero. But if he does get lucky enough to get a ticket, he then has spent $40 for something that is worth $50 to him. This means that in an economic sense, the artificially low bid has a positive expected value, and is strategically advisable. It also means there is an exceptional incentive for e.g. ticket brokers (a.k.a. scalping agencies) to place numerous artificially low bids, knowing that every time they win, they make a profit.

In a traditional “English auction”, this problem is addressed by allowing participants to know what other participants are bidding. Say that Bob thinks a Coldplay ticket is worth $50, and Alice thinks it is worth $60. Bidding begins at some amount, and then Bob and Alice play leap-frog, progressing in whatever increments they have the patience for (theoretically, two cents at a time), until the bid is greater than $50. Alice then gets the ticket, that is really worth $60 dollars to her, for $50.

A second-price “Vickrey” auction accomplishes the exact same result, except without the cost in time and labor of a real auction. Bob and Alice, and the other participants, simply write down their real bids in sealed envelopes (or on a form on a web site), and when bidding stops, Alice wins but only pays what Bob bid. This saved the hassle of the English auction, but also prevented the strategic “gaming the system” that we saw with the first-place-bid system above. For instance, there is no reason for Alice (or a scalper) to under-bid, because she’ll only have to pay the second highest price anyway.

What About the House’s Bottom Line?

Many would look at this scenario as sub-optimal for the auction house. If only they had just put a sticker price of $60 on the ticket, Alice would have bought it, and they would have increased their profit. But there are two reasons this isn’t a problem.

For one, whenever the number of bidders is large, it’s a sure bet that the next-highest bidder wouldn’t be $10 behind, but more like 20 cents. Imagine 1000 people were bidding for the ticket — certainly someone would have bet just shy of Alice’s $60. Realistically, she’d probably end up paying something like $59.93. Or if you’re really talking about thousands of bidders, $59.99. So the house isn’t loosing much, especially again considering the time-labor savings of avoiding an English auction.

Secondly, we have to consider all the money the house loses if they decide to guess the best selling price on a volume of many items, and as a consequence of trying to be sure to set the price high enough, they lose some sales. In the concert business, we call this “empty seats”. We don’t want rotting bananas in our stands because we asked a dollar more for them than someone out there was willing to pay.

Being First In Line

In a system where winning is all about being first in line, it is a natural unavoidable consequence that buyers will flood the sales floor the moment the doors open. In a brick and mortar establishment this poses staffing problems. In an online environment those huge spikes in traffic pose a unique challenge, and require us to spend 100% of the time equipped with resources that we only need perhaps 1% of the time.

In a system where winning is all about bidding the most, this abominable engineering challenge is slain. Instead of tickets for a show going on sale June 1, they go on auction May 20, and the auction closes on June 1. No need to rush, except in the edge case that the show will probably sell out, and you assume your bid will be tied with the lowest successful bid. But in that case, a lottery tie-breaker can eliminate even that early bird incentive.

Selling Out Every Show

For most reputable acts playing an appropriately-sized venue, it’s entirely conceivable that there is some low-enough price at which the show would sell out. But as just discussed, if a ticket vendor were to sell tickets at that price, the loss incurred would be greater than the loss incurred by failing to sell all tickets at a higher price. Vickrey auctions provide the best of both worlds, by selling as many tickets as would be sold if the tickets were free, but selling them all at just a little less than what each concert-goer thought they were worth. Let’s contemplate that. Every single ticket-buyer gets a ticket for as much, or less, than he thinks it is worth, and we sell as many tickets as could be sold for free. That point serves as a nice segue into an economic discussion of utilitarianism.

Customer Satisfaction — An Issue of Utility Theory

One of the more esoteric facets of economics is the concept of utility, a measure of real subjective value that transcends strict monetary measures. Utility has been defined by various thinkers (e.g. Jeremy Bentham) as satisfaction, happiness, pleasure, etc. The difference between monetary value and utility can be demonstrated by the fact that a fine has a smaller deterrent effect to someone with greater disposable wealth, whereas jail time or community service has a more equitable nature.

Alternatively, for any two people with congruent disposable wealth, the one who is willing to spend more for some product demonstrably values it more. This has enormous consequences for average customer satisfaction in instances in which demand exceeds supply. For instance, imagine our Coldplay show, in which tickets sell for $50, and Bob won the “click page-refresh furiously” lottery, and Alice did not. Bob, by sheer dumb luck, got to see a show that Alice would have gotten more enjoyment from. If we implement a system in which Alice would have instead gotten this ticket by paying more (because it really mattered that much to her), then Bob will be compensated for missing out on Coldplay when he comes to our site looking to buy tickets for his favorite band in the world. Instead of his success or failure being largely the product of randomness, so that the buyer often gets tickets to shows he “sort of” wants to see, but fails to get tickets to the shows he really wants to see, he will now be most likely to get the tickets he most wants. This explains a certain sense of satisfaction I’ve personally experienced after buying an expensive scalped ticket to a show that sold out in minutes. I got to see a show that I passionately wanted to see, because I was willing to prove my enthusiasm with my wallet. If you multiply this effect by the millions of fans out there, this is an absolutely enormous increase in the total/average satisfaction of our customer base. And our ability to generate the maximum possible revenue will make even the venues we do not own want to come solely to us for their ticketing needs. At least, that’s my inexpert prediction.

What About Artist Backlash?

One of the most predictable results of implementing this kind of policy is that those numerous musicians for whom the art matters more than the business (I count myself among them), will complain that we are “gouging” the fans. They might argue, so what if zillionaire X got to see Coldplay because he genuinely did like them more passionately than zillionaire Y?! What about all those non-zillionaires out there? This is where we would need to give the artist a gentle economics lesson, as well as a concession of sorts.

See, if there are lots of people who wanted to see a show and didn’t get to, that simply means there is more demand than supply. More people want to go see U2 than there are copies of U2. We only get one Bono, unfortunately. Limited supply is just an inescapable fact of economics, and if the artist is really so bothered by it, then he needs to play more shows (or become less popular).

The reality that artists really need to look at is that when demand exceeds supply, and tickets are priced below market value (to ensure that the kids can go see their heroes), the market demands that scalpers fill their niche. And they do. And stopping them as long as there is an economic incentive for their trade is a monstrous engineering challenge of its own. And what’s even more pernicious about scalpers is that they both take money that should belong to the artist and the legitimate ticket broker, and waste money. They waste money because it takes money for them to operate, but they are just re-selling tickets for a higher cost — which we could have done in the first place, and at least given it to the artist or even a charity of the artist’s choosing. This whole racket would make an economist want to tear hi hair out. We have to educate artists into understanding that wasted money doesn’t do anyone any good. It’s hugely, vastly, incomprehensibly in our best interest (and theirs, and the fans’) to do that. We are talking about billions of dollars worth of sales and customer satisfaction that is otherwise being wasted.

Auction-Lottery Combination

Still, many artists will feel at least some seats should be available to devoted fans who don’t happen to be rich. This is an unfortunate kind of naïveté, because what they don’t realize is that if little Timmy lucks his way into a ticket that costs him $40 dollars, but is actually worth $200 to him (even if he doesn’t have $200), then he will have every reason to scalp that ticket (e.g. on Craigslist) if he can get more than $200 for it. And he will be able to do that if he was allowed to buy a ticket for less than someone else would have paid for it. You just can’t beat the market. My friend Milton Friedman taught me that much.

Admittedly there will be artists who simply won’t be able to grasp this, or just won’t want to because it violates their altruistic leanings. And to a certain degree, it’s essential to our artist relationships and company image that our system assuage their fears that we might be another Ticketmaster, trying to bleed them for every last penny (literally). The “solution” is a hybrid system in which a certain number, S, of seats are set aside at a set price, P, (e.g. $30) for a lottery among people who bid at least P. All lottery winners are limited to buying a maximum of, say, 4 tickets, and the person whose name is on the credit card must show up at will call, and actually show the card and an ID to get the tickets, and must either go directly from there into the venue, or re-present his ID when he does. [This actually happened at a Coldplay show I tried to see at the Fillmore a few years back. This is an additional step to help prevent scalping, although the worst that could happen with this system, even if scalping were unchecked, is what would happen if we implemented the system identically, except auctioned 100% of the tickets with no lottery at all.]

These lottery contestants could be ordinary ticket-buyers, or even members of the artist’s fan club, like the massive one that Pearl Jam maintains. The people who end up winning these tickets may have bid $30 or $30.01… or $400. That may violate the spirit of wanting to give these tickets to less wealthy people, but if we try to make a cutoff point, so that the lottery only applies to bidders within a relatively low range, then we simply give wealthier people an incentive to place artificial low bids. Strategic/insincere low bidding is bad, so we have to just make the lottery blind to bid size (so long as the bid at least meets the minimum, P). The idea is that the probability dictates something like half of the lottery tickets will go to non-rich fans, and so we simply reserve around twice as many seats as the artist (and our P.R. advisors) want to make available for them, knowing that sometimes more will go to the non-rich fans, and other times more will go to the rich fans — this is more of that troublesomely inescapable economics stuff.

Some implementation notes are:

• Lottery winners’ bids must be removed from the bid pool. (I.e. if bids exist for $20, $25, and $30, and the $25 bidder wins the lottery, then if the $30 bidder wins a normal auction ticket, he only pays $20 for it, not $25 — this is mathematically necessary for the system to work.)
• Seats should never go empty, so customers should be free to bid e.g. 10 cents if they want (or whatever minimum is necessary for us to break even on their bandwidth usage from our site, and the show cleanup/usher costs they incur), and if there are seats for them, they get to go — courtesy of Live Nation, the music company they now worship, even though we are just giving them a product that we were going to “throw out” anyway.
• That scheme could be utilized by giving concert-goers “frequent rocker miles”, which they could use to bid on shows, so that potentially wasted empty seats are at least used to build fan loyalty.
• This entire scheme requires re-training the minds of artists and fans, but it seems to me to be a hugely economically advantageous undertaking, that could, with some skill, be spun into a narrative of our revolutionary departure from the old school “Ticketmaster” approach

Hopefully you have found this diatribe stimulating. It was great getting a chance to talk with you about your college days and your family this past December. I look forward to hearing your thoughts should any of the ideas I’ve presented here strike you as remotely viable.

Best regards,

Clay