In late 2016, Billy McFarland was raising money for a festival that did not yet have a location, a stage, or a single act confirmed to perform. He had co-founded Fyre Media the year before, built around an app for booking musicians, and the festival was meant to be its launch moment: a weekend on a private Bahamian island once owned by Pablo Escobar, marketed with a single video featuring supermodels running along a beach in slow motion.
By January 2017, tickets were selling for upward of $12,000. Kendall Jenner, Bella Hadid, and a roster of other influencers posted a single orange square to their accounts with no caption, and within 48 hours the festival had generated more organic reach than most concert promoters see in a year of paid advertising. What almost nobody framed correctly at the time was that every ticket sold was not a sale. It was a promise, and every influencer post was another one stacked on top of it, each an obligation the festival would eventually have to make good on or default against, publicly, all at once.
When the first guests landed on April 27, the island had no venue built. The luxury villas promised in the marketing did not exist. Guests were housed in FEMA disaster relief tents left over from a hurricane response, and the catered gourmet meals became, famously, a slice of bread with two pieces of cheese and a small salad. The festival was cancelled within a day. McFarland was arrested by the FBI in June 2017 and sentenced to six years in federal prison for wire fraud, having raised more than $26 million on the strength of promises that could never have been fulfilled at the scale they were sold.
Fyre did not fail for lack of attention. It failed because it kept selling promises without ever building anything capable of paying them back, and the debt came due on a single weekend, in front of everyone who had bought in.
The Reserve Nobody Checks Until It's Due
Banking makes this mechanism explicit rather than metaphorical. Under fractional reserve banking, a bank holds only a small percentage of its total deposits in actual cash, lending the rest out, and the entire system survives on the assumption that not everyone will ask for their money back on the same day. When that assumption breaks, whether at Northern Rock in the United Kingdom in 2007 or Silicon Valley Bank in California in March 2023, the money didn't vanish overnight. A sufficient number of depositors simply tested, at the same moment, a promise that had gone untested until then.
Brands run a less formal but structurally identical arrangement. Every piece of attention a brand earns arrives with an unstated promise attached: that the product will work as described, that the experience will match what was implied, that the reputation being traded on was actually earned rather than borrowed. Nobody counts the debt until a crisis forces the count.
Every Promise Is a Liability Until It's Paid
On April 9, 2017, a passenger named David Dao was forcibly removed from an overbooked United Airlines flight at Chicago O'Hare after refusing to give up his seat. Footage of airline security dragging Dao, bloodied, down the aisle was viewed hundreds of millions of times within the week. United's stock fell roughly 4 percent over the following two trading days, wiping out an estimated $1.4 billion in market capitalization at one point. The damage wasn't caused by one gate employee's decision. It was millions of people who had never personally flown United suddenly calling in a promise, implicit in decades of the brand's advertising and reputation, that the airline had never explicitly made but had always implied it could keep.
Apple shows what it looks like when the promise is actually funded before it's called in. When the company introduced the Apple Watch in September 2014, it launched into an audience the iPhone had spent the better part of a decade earning, and it carried none of the skepticism that usually greets a first-generation device from a company new to the category. The debt had already been repaid, continuously, for years, before the product ever asked anyone to extend it new credit.
OpenAI's ChatGPT offers a live, ongoing version of the same mechanism. Every time the product experiences an outage, the story becomes front-page news within hours, not because the outage itself is unusual for a piece of software at that scale, but because the company's explosive growth manufactured an enormous number of implicit promises about reliability that most users never consciously agreed to but now expect to be honored anyway. Attention did not just win ChatGPT users. It wrote a very large number of promises the product now has to keep paying down, one uptime report at a time.
What the Balance Sheet Doesn't Show
Every company publishes a financial statement. Almost none could produce an honest account of how many implicit promises they are currently carrying, because most have never tried to count them.
That gap won't last. As the cost of buying attention through advertising keeps rising, the companies that separate themselves will be the ones tracking retention, repeat engagement, and the gap between what they've promised and what they've delivered as deliberately as they track cash flow, rather than leaving it to intuition inside a marketing department.
Engine
Every campaign writes a promise. Every product either fulfills it or doesn't. The gap between the two is the debt a brand is carrying, whether or not anyone inside the company has ever bothered to add it up.
Billy McFarland remains barred from starting new businesses without prior approval as a condition of his supervised release, which ended in 2022. United Airlines, years after the Dao incident, still ranks among the largest carriers in the world, because the underlying reserve, built on decades of flights that mostly departed and mostly arrived, was large enough to absorb one very bad week of video without going insolvent.
Fyre never had that reserve. It sold years of promises in a single season and had nothing built to make good on any of them when the guests actually arrived. The lesson was never that attention is dangerous, or that marketing should be more modest. It's narrower than that: the only question that ever mattered was how many promises a brand could afford to have called in on the same day, and almost nobody finds out the real answer until the day it happens.
