What about the music itself? I keep hearing people express their disinterest in attending any major festivals in the U.S. because “all the lineups are the same.” As a staunch advocate for emerging artists, I’m tempted to agree with this narrative, which aligns well with the ongoing corporatization of festivals and the well-documented shift in big business toward experiential marketing (conglomerates like Live Nation and AEG seem to be scooping up new festivals every month, while brands from PepsiCo and Coca-Cola to Ford and General Motors regularly invest millions in onsite festival activations).
A common hypothesis is that there’s a limited amount of “top talent” (whatever that means) navigating the festival circuit in a given year, usually before or after a new album release cycle, for which a growing number of major festivals must compete. It’s also the job of festival producers to strike a balance between recruiting a high-quality lineup and selling enough tickets to cover their multimillion-dollar overhead—the latter of which relies heavily on the hype that accompanies big-name headliners.
With my numbers cap on, I decided to scour this year’s festival lineups in search of quantifiable answers to these otherwise elusive questions. On the basis of uniqueness alone, which festivals give you the most bang for your buck? Which artists are leaving the largest festival footprint this year, and do they fit our archetype of the top-40 and/or pseudo-hippie cliché? To what extent are the soundtracks behind these sweaty congregations of hundreds of thousands of people really one and the same?
Turns out—as with most other issues in the music industry—the underlying narrative is much more complex than we’ve made it out to be.
How much do we already know?
There’s a small but growing genealogy of online research about, and advocacy against, overlaps in festival lineups. In June 2015, New York Times columnist Stephen Heyman found that 70% of the top 20 summer music festivals in 2015 were headlined by at least one of Florence and the Machine, Muse or Avicii. The Times’ pop-music critics later repackaged this data analysis as an editorial decision, declaring that they would no longer cover the “cross-genre, medium-cool … purchasable bohemia” of the Coachella/Bonnaroo world, in pursuit of more niche festival scenes that allow for more “intelligent” journalism.
Later in April 2016, within the span of two days, both Quartz and Deadspin published their own data-driven analyses of festival overlaps, further perpetuating the homogeneity narrative. Prior to coming across any of this coverage, I posted my own Venn-diagram of lineups on Twitter that confirmed many of Deadspin’s findings.
The rising awareness of lineup overlaps—alongside news about Bonnaroo and Sasquatch reporting their lowest attendance numbers in history—is fueling fears that the world has reached “peak festival.” What does that phrase even mean?
According to onlookers, it’s a three-ingredient cocktail of market over-saturation, corporatization and just plain lackluster music. Eric Danton wrote in Paste Magazine that mainstream outdoor festivals are deteriorating because they’re more interested in “attracting a massive audience [rather] than the right audience.” Speaking at the Hay Festival of Literature & Arts, rock promoter Harvey Goldsmith argued not only that there are too many festivals, but also that “there are not enough big acts to headline them … we are not producing a new generation of these kind of acts — the likes of the Rolling Stones, Muse, even Arctic Monkeys — that can headline.”
Data + methods
To evaluate the validity of these claims in 2017, I studied nine upcoming music festivals of varying sizes and geographic locations, covering a total of 514 artists. Below are the festivals sorted by date, along with their cities, approximate audience sizes as of last year and GA ticket prices—most of which don’t include service fees, shuttle prices or other add-ons, so brace yourselves:
- Coachella (weekends of April 14–23 in Indio, CA; 200,000 attendees; $399)
- Hangout (May 19–21 in Gulf Shores, AL; 100,000 attendees; $309)
- Boston Calling (May 26–28 in Boston, MA; 22,000 attendees; $269)
- Sasquatch (May 26–28 in George, WA; 11,000 attendees; $295) (yes, same exact dates as Boston Calling)
- Governors Ball (June 2–4 in New York, NY; 150,000 attendees; $305)
- Bonnaroo (June 8–11 in Manchester, TN; 46,000 attendees; $349.50)
- Bonanza Campout (June 23–25 in Heber, UT; attendance unknown; $175)
- Lollapalooza Paris (July 22–23 in Paris, FR; inaugural year, so no attendance data; €149 or ~$160)
- Lollapalooza USA (August 3–6 in Chicago, IL; 400,000 attendees; $335)
I mapped out the relationships among these festivals and their artists using Gephi, an open-source network analysis and visualization tool with which I first got acquainted through my sociology class Culture and Networks. The following graphs were generated using the Fruchterman Reingold algorithm, a force-directed algorithm that mimics attraction and repulsion among atomic particles (this will make more sense in a minute). I’ve included my raw data at the bottom of this article, and invite you to play around with it yourself.
Network analysis is a useful lens for evaluating and challenging several of the arguments made by the authors in the previous section. For instance, Goldsmith’s remarks are somewhat puzzling. Not only is there more than enough “headliner-worthy” material going around this year—The Red Hot Chili Peppers, Frank Ocean, Twenty One Pilots, Chance the Rapper, Run the Jewels, The Weeknd, you name it—but headliners alone are an inaccurate representation of how the wider music industry interacts with the festival scene. Network analysis helps prove that smaller, emerging artists, not major headliners, are ultimately the glue that holds festivals together.
A watered-down crash course on network analysis terms.
- Node: a point of intersection or connection within a network. e.g. In the following diagrams, an individual festival or artist will be represented by a circular node.
- Edge: a line or curve representing a connection between two nodes. e.g. In the next diagram, there will be an edge between two festivals that share at least one artist.
- Path length: the number of edges required to travel between two given nodes.
- Degree: the number of edges connected to a particular node.
- Betweenness centrality: a ranking system that assigns nodes in a network a value from 0 to 1, depending on their presence in the shortest paths between other nodes. If removed, nodes with high betweenness centrality can potentially divide a network into two disjoint sections.
- Clustering coefficient: a measure of the degree to which nodes in a graph tend to cluster together. Each node is assigned a value between 0 and 1, with 1 being the most (0 being the least) likely to cluster with a fixed group of other nodes.
Results, Part I: The Festival Network
The above diagram is called a “festival-festival network,” showing relationships between festivals in terms of common lineups. Thicker edges correspond to more overlaps; e.g. from observing this diagram alone, we now know that Lollapalooza USA has many artists in common with Coachella, Bonnaroo and Governors Ball, but relatively fewer connections with Bonanza Campout.
The average degree of a node in this network is 7.1 (±1.2), implying that each festival shares artists with an average of around 7 other festivals in the dataset (in fact, five of these festivals have a degree of 8, implying that they overlap with every other festival). Lollapalooza Paris has the lowest degree (only 5; no overlap with Boston Calling, Bonanza Hangout, Sasquatch or its USA counterpart), which is understandable considering that it’s the only overseas festival in my analysis. Still, 5 is a pretty impressive number.
While this diagram is effective at representing festival relationships, it reveals close to nothing about the lineups themselves. Below is a ranking of the lineups from least to most “unique,” measured by the number of overlapping acts with other festivals in the dataset:
- Governors Ball: 48 overlaps out of 66 total acts (72.7% non-unique).
- Hangout: 43 out of 62 (69.4%).
- Boston Calling: 28 out of 43 (65.1%).
- Bonnaroo: 59 out of 103 (57.3%).
- Lollapalooza USA: 83 out of 168 (49.4%).
- Sasquatch: 27 out of 61 (44.3%).
- Coachella: 58 out of 149 (38.9%).
- Lollapalooza Paris: 17 out of 47 (36.2%).
- Bonanza Campout: 8 out of 37 (21.6%).
These percentages may seem astonishingly high, until one realizes that much of the live music business is controlled by a handful of corporate, quasi-monopolistic entities. Live Nation owns Governors Ball, Bonnaroo, Sasquatch and the Lollapalooza franchise, while competitor AEG owns Coachella and Hangout—there goes two-thirds of our list already (Crash Line Productions organizes Boston Calling, while Live Nite Events promotes Bonanza Campout).
Since Lollapalooza USA has the largest lineup, it accounts for many of the overlaps in other festivals—sharing 34 acts with Coachella, 30 with Bonnaroo, 24 with Governors Ball and 19 with Hangout. Sasquatch + Governors Ball is another common pairing for artists, with over half of the former’s overlap acts performing at the latter.
Another interesting tidbit: while there’s a slight, insignificant negative correlation between lineup size and % overlap (-0.015), a more pronounced negative correlation does exist between audience size and % overlap (-0.259). These two stats aren’t contradictory. On one hand, festivals with growing crowds are recruiting even more artists to keep their concertgoers busy during the day; on the other hand, larger lineups can easily be much less “unique” than smaller ones, depending on their target audience (e.g. Governors Ball is explicitly mass-market and cross-genre, so has a lot more overlap than Bonanza Campout, which focuses more on hip-hop and electronic acts as well as a carefully-controlled camping experience).
Results, Part II: The Artist Network
Now, let’s move on to the “artist-artist network,” which shows relationships among individual artists according to their shared presence at music festivals. The average degree of this graph is 142.1 (±80.4), while the average path length is 1.7. In other words, not only does the average festival artist tend to be associated with multiple lineups, but the mainstream festival world at large is also extremely small — operating on fewer than two degrees of separation, rather than six.
For the sake of cleanliness, I removed the edges from this diagram, in order to spotlight a clear clustering pattern among the artists. There are nine discernible reddish-orange clusters, each representing a group of artists unique to a particular festival (e.g. the top, bottom left and bottom right clusters represent the artists unique to Sasquatch, Lollapalooza Paris and Bonanza Campout, respectively).
Here’s where the attraction/repulsion concept I introduced earlier comes in handy: the further away a cluster is from the center, the less overlap it has with other festivals and the less influence its artists have on the larger network. Hence, it makes sense that Lollapalooza Paris and Bonanza Campout are further away from the center, confirming our “uniqueness” rankings in the previous section.
In the middle swims a messy sea of blue-green nodes that represent artists who are featured on several lineups, exerting the most influence on the overall network. The term “influence,” which I’ll discuss in greater depth later on, is largely correlated with degree: indeed, the top 25 artists in the network have a degree of 300 or higher.
If we isolate and assemble these top 25 artists into their own network, we get a graph that’s much more evenly distributed, without any clusters. The average degree of this graph is 24, while the average path length is 1—i.e. every node is connected to every single other node.
It’s also fascinating to note the relative absence of major-label talent, which accounts for just 36% of this graph (The Head and the Heart, Lorde, DJ Snake, Bishop Briggs, Majid Jordan, Mondo Cozmo, Tove Lo, Kaleo, Gryffin). This is almost an exact inverse of the power relations in global recorded music, in which majors command 62.4% of the market, according to the latest Worldwide Independent Market Report. Therefore, despite criticisms of corporatization on the sponsorship side, festival monopolization from a musical perspective is still an indie game.
Results, Part III: The Artist Influence
Here, I’ll dig deeper into what exactly I mean by “influence,” a complex term that yields different outcomes depending on how you measure it. For instance, it’s one thing to be the most prolific artist in terms of festival count, but it’s another thing entirely to be the “most influential” artist in the wider festival network.
Let’s start with this metric because it’s the easiest to calculate. Three artists—Car Seat Headrest, The Head and the Heart and Chance the Rapper—will each be performing at six festivals on our list. Tied for second place are Mac DeMarco, Mondo Cozmo and Marshmello (ALL the Ms!), who will each be performing at five festivals. An additional 15 artists will perform at four festivals; 26 additional artists at three; and 83 additional artists at two (see raw data for complete info).
This might seem like a lot of overlap, but it’s important to realize that only 130 artists are playing more than one festival — leaving 384 artists, or 75% of our dataset, as single-festival performers. We can connect this directly to the artist-artist network in the previous section, in which only a handful of nodes are dark blue.
As I alluded to before, the coveted group of five- and six-festival artists are mostly indie, and mostly not headliners (with the exception of Chance). In fact, many of the major-label artists in our dataset, such as the Red Hot Chili Peppers, D.R.A.M., The Weeknd and Schoolboy Q, are actually performing at only two festivals—suggesting that established members of the major-label elite don’t necessarily prioritize festivals over other, more easily-controllable live revenue streams, such as solo tours.
Going one level deeper than festival count, a more useful influence metric is degree, i.e. the number of artists with which one shares a lineup. The Head and the Heart tops the list with a whopping degree of 437, followed by Car Seat Headrest at 432, plus new entrants Joseph at 380 alongside Glass Animals and Kaleo tied at 379:
Note that while Chance the Rapper is tied with The Head and the Heart and Car Seat Headrest for festival count, he ranks lower than these acts on the basis of degree, because the festivals that he’s headlining have smaller lineups altogether.
Another common influence measure is betweenness centrality, which, as previously defined, ranks nodes according to their presence in a path between other nodes. The top-10 list for this metric yields a larger presence of hip-hop and electronic artists, suggesting that those specific genres serve as the interlocking energy between festivals—in contrast to our previous lists for festival count and degree, which seem to favor a more pop-rock sound.
Jai Wolf is the surprising victor of this list with a betweenness centrality of 0.04, while producer DJ Snake and rapper Kaiydo also join the ranks:
Below is a chart of the 10 artists least likely to cluster with other artists, due to their presence across multiple festivals. The list seems to be a hybrid of those for degree and betweenness centrality, featuring Canadian electronic duo Bob Moses as a new entrant:
The Head and the Heart, Car Seat Headrest, Chance the Rapper, Kaleo and Glass Animals are the only acts included in the top 10 across all three influence metrics. What’s more, every single one of these artists has released a new full-length album within the last year, and are clearly maximizing the festival circuit as a cross-promotional tool for their recordings.
We tend to oversimplify the story of music festival lineups in America.
In the modern streaming era, both legacy and emerging artists view live music—and festival presence in particular—not only as their primary revenue stream, but also as a key barometer for success. Legacy artists are eager to cash in on “pseudo-hippie” nostalgia, either staging comebacks or performing music composed years or even decades ago that the puppetry of terrestrial radio still deems relevant (Red Hot Chili Peppers, Rolling Stones, MGMT, and so on). Emerging artists see festival placements as stamps of legitimacy and gateways to new music fans—even if they’re just being paid in “exposure” and the cards are ruthlessly stacked against their long-term success.
Legacy acts may rule the media, but the constantly-evolving if unstable independent music scene is the unsung hero in the festival world. I hope my analysis reinforces this fact that any perceived homogeneity across music festivals is simply the result of advertising. As Sasquatch founder Adam Zacks told A.V. Club, when you make your purchase decisions or aesthetic critiques only based on the top headliners, “you get the snapshot view and so much gets missed.”
From an economic standpoint, it also doesn’t make sense for music festivals to pursue lineup popularity alone. According to insider trading analyst Jake Mann, lineup popularity correlates neither with size nor with profits, and a festival’s overall atmosphere and reputation are equally instrumental to its success. Even if Governors Ball or Hangout might be among the least musically unique festivals in the world, their locations are certainly distinct (one an iconic metropolis, the other a cozy beach town), and each offers its own adjacent, carefully-curated culinary and visual palettes in addition to sonic ones.
I would love to conduct further analyses that incorporate either a more international (Glastonbury, Primavera Sound, Sónar) or a more genre-specific (Ultra, Electric Zoo, Electric Daisy Carnival) outlook. The significant “influence” of indie artists on the festival circuit also begs the question of to what extent festivals actually propel and sustain these artists’ careers, considering the minuscule odds of moving up the ladder towards larger venues and more prestigious headlining gigs.
In the meantime, is the global live music industry’s $28-billion-and-counting valuation really banking on its lineups becoming more homogenous? I hope not.
If you want to explore the data some more, here’s a spreadsheet of my raw node-edge data and centrality statistics.
If you have any other interesting insights to offer, or if I blatantly left something out in my analysis, please let me know in the comments below! Let’s start a productive, exciting conversation around live music.