"It's about the right to begin again. The right to be free of your past. That's grace. So, yes, you can write lyrics about debt relief."
Vorsprung durch Technik: U2 Fans & Ticketmaster Algorithms
December 06, 2017
Sometimes the air is so anxious
When Live Nation and U2 announced that the 2018 U2 Experience and Innocence Tour was going to feature Ticketmaster’s new Verified Fan system, many fans took a deep breath in anticipation of a bumpy ride.
It is not my goal in this article to revisit the recent presale experience. Rather, I want to fuse my love of U2 with my love of machine learning and see if there are lessons for us: on the role that machines are increasingly taking in the lives of the concert-going U2 fan and the ways in which fans, tour promoters and even scalpers will need to adapt in a technological arms race.
That deep breath fans took may not have been warranted, but it was a natural human response. A combination of anxiety and intuition bracing for a worst-case scenario. It's something we do well as humans: assess risk with minimal information. Don't hope for too much. We know that change can bring surprises and prepare ourselves for disappointment because we know it is better to be positively surprised, especially when it comes to securing "Rock’s hottest ticket."
This screwed up stuff is the stuff of dreams
Computer algorithms have a different kind of intuition. They don't suffer from irrational anxiety. Instead, every decision they make is based on data and rules. Some of these rules are designed to match human intuition (Don't sell 20 tickets to one person; it's unlikely they are planning a family reunion at Madison Square Garden in June) while others pick up on patterns humans would never consider (Don't sell tickets to somebody who only buys tickets to weekend shows in multiple major cities using a credit card they change every three months). This is machine learning, and the future on which Live Nation is pinning its hopes of winning the battle against scalpers.
The arms race of fans vs. scalpers is nothing new – we've had CAPTCHA images for well over a decade to block ticket tout bots. And uberfans – a term I am using to describe the fan who travels to multiple shows away from their home town – know to be careful about ordering too many tickets using the same name, address and credit card.
What is different now is the role that big data is playing in determining who is a fan and who is a scalper. Ticketmaster introduced the ticket-selling computer-algorithm equivalent of the carnival trickster that will guess your age or your weight. And they bet big that it could guess correctly.
You're the best thing about me
Your online activities, purchases, travels, listening habits become a digital representation of who you are and by extension your fandom.
This "digital footprint" becomes one of your biggest assets and potentially your biggest liability.
If your behavior is a little too close to that of a scalper, you may be in for a rocky experience.
It turns out that in a world of millions of casual fans, there isn't a huge difference between a rabid fan who buys 20+ tickets to 20+ shows and a scalper who does the same thing and immediately turns around and sells them.
Telling the difference requires that you pay close attention and, sadly, computer algorithms are not perfect. This is something we intuitively know, and unless we are a labcoat-wearing Jeff Goldblum in a sci-fi summer blockbuster, nobody blindly trusts a computer to make important decisions without some form of human oversight or really good programmers.
It's the little things that give you away
When you use your credit card, a computer algorithm instantaneously decides whether or not it thinks your purchase is fraudulent.
Is this a shop you normally shop at? Was the purchase made in a place you could have traveled to naturally since you last purchased? Was the purchase for something expensive that could easily be resold on the black market?
These are the obvious things an algorithm looks for, but by looking at miniature variations in your behavior, an algorithm can detect anomalies you wouldn't think of. Perhaps you do use that gas station regularly, but you've never spent over $40 on gas. Perhaps you are normally at work in the city during those hours. The algorithm detects something strange and raises a flag.
Red flag day
What *should* happen next is that your charge is put on hold while the bank frantically contacts you to confirm that the purchase is legitimate. For banks this activity happens on a regular basis, so they have a huge human army to handle those flagged transactions. Ticketmaster had a unique challenge: Tours like U2's are somewhat rare and come together quickly. Fan accounts of calls to the Ticketmaster call centers suggest they didn't have much time to get their human verifiers in place, if they even realized that this was a necessary part of a machine learning verification process.
This human part of the process is necessary because humans are incredibly unpredictable. If humans weren't driving or walking on roads, we'd probably all have self-driving cars already.
I can’t see you through the fears
I've spent most of my career teaching computers to predict human decisions or behave like humans, and it’s an unforgiving task. In aggregate, computers can predict certain behavior with eerie accuracy. But ask a computer to predict a single human behavior or preferences over a lifetime and you are continually chasing a moving target. I call it job security.
You think you know what makes people tick, then their collective behavior changes. Investors who ignored early economic warning signs start freaking out at the first sign of trouble. Shoppers who would buy records on Record Store Day decided to stream on Spotify. Longtime U2 fans find themselves reselling their concert tickets because of a family commitment. Is it the beginning of a recession? The end of retail? A scalper in the making? What you thought was certain is now a question, and the validity of your predictions is now a matter of judgment.
In a paper I wrote earlier this year in the Journal Of Investment Management, we used machine learning to mimic human preferences for selecting articles to read and were able to mimic them 78 percent of the time. Did we fail the other 22 percent? No. Turns out the human preferences differed 17 percent of the time for no discernable reason. There is a limit on how well any computer can discern who we are because we often don't know who we are ourselves.
Get out of your own way
What does this mean for U2 fans trying to manage their online identities and stay in the running for concert tickets when they are humans stuck in the middle of a war between AI-driven tour promoters and AI-driven scalpers? I'll offer a couple of suggestions, all of which should be taken with a pinch of salt (I don’t work for Ticketmaster or begin to understand how their machines actually work).
1) Blessed are the superstars: Let Ticketmaster see evidence of your fandom
Make use of your online identity judiciously. It’s ironic that fans of the band who predicted and lamented a big-brother style tracking of music preferences would need to intentionally flash their u2 fandom to the "eye in the sky," but anything you can do to help Ticketmaster recognize you as a true fan is going to help. They really want you to link all your social accounts and allow their app to access your location when at the venue.
Live Nation currently is monetizing that social media information. According to Samantha Sichel, Live Nation vice president, digital product & business development, “Live Nation is laser focused on curating an authentic, two-way conversation before, during and after a live entertainment event. Over the last two years we have assembled a team to focus specifically on finding new ways to allow brands to capitalize on social API’s and develop innovative digital tools to improve communication. We have seen tremendous success so far and we have only scratched the surface."
Your online identity is valuable information for Live Nation and is probably worth more than the cost of your tickets. Freely letting them profit from your identity seems like a bad idea. The risk that Ticketmaster could use its knowledge about your fandom to coerce you buy into more lucrative concert experiences probably makes your blood boil. But giving them something that can authenticate your fandom may be worth it if it lets you breathe when tickets go on sale for the next tour leg. As Michael Kosinski notes in "The End of Privacy," it only takes 200 data points to know certain preferences better than your spouse does. (It’s reasonable to assume that the data scientists maximizing revenues are better incentivized than the ones running a scalper filter.)
Don't go overboard. They don't need to know any more of your deepest secrets. But a tweet from a spare Twitter account linked to the TM app stating, "Can't wait for #u2eiTour to start @U2 @atu2 @LiveNation #LittleMoreBetter" can’t hurt. Just be glad you aren't Taylor Swift fans who get tweeting duty as part of their fan homework assignment.
2) Blessed are the filthy rich: Let your wallet demonstrate your fandom
If money is no object, you can sell your soul to the corporate beast and do all your Christmas shopping through Fanfire. Buy the stuff no self-respecting scalper would buy.
Now that your u2.com account is linked to Ticketmaster, your purchases should count.
3) Blessed are the liars for the truth is awkward: Don't be a scalper-fan!
This should go without saying, but don't do anything that would make you look like a scalper. Don't buy extra tickets to scalp to fund your tour habit. Try to minimize the chance that you'll have to sell you ticket for legitimate reasons. Sometimes it's unavoidable, but if you have to choose between a friend’s wedding and a U2 show, remember where your priorities lie!
4) Blessed are the bullies: Call Ticketmaster and give them a piece of your mind
When all else fails, use your voice. As painful as many fans' experiences were when calling Ticketmaster to get help, it creates data that says, "This seems to be a U2 fan and they are really angry."
Scalpers are going to call too, but you are forcing Ticketmaster to have a human-in-the loop. Once this human figures out that you are an actual fan, the computer is more likely to trust that person’s judgement rather than its own intuition.
Ironically, the recent fan backlash likely created a new data cluster in the Ticketmaster computer: pissed-off U2 fan. Hopefully for everyone’s sake that profile is one that is hard for scalpers to mimic.
© @U2/Steve Lawrence, 2017