The headline screamed, in capital letters:


It was an article posted last month on a breezily designed online publication called Quartzy, which is a spinoff effort from a larger publication called Quartz. Quartzy calls itself “a guide to living well in the new global economy”; Quartz, meanwhile, was founded in 2012 by The Atlantic as “a digitally native news outlet” and self-proclaimed publisher of “bracingly creative and intelligent journalism with a broad worldview.”

But you know how it is online. Publications have to post copy relentlessly, and prefer not to pay writers very much, if at all. So bullshit pieces like this article about metal being the “most-loved genre of music” get published, even (indirectly) by The Atlantic. By now they’ve probably forgotten they even posted it.

I can see what Quartzy was trying to do. Someone saw an article in Billboard earlier in November about a new streaming service that wants to cater exclusively to heavy metal fans and figured they could make a story out of it by combining it with that article’s tangential mention of Spotify data that claimed that heavy metal is the music genre with the most loyal listeners. The Billboard piece was focused on the new streaming service, called Gimme Radio; Quartzy decided to lead with the “most loyal genre” factoid, probably surmising this was more generally interesting to their audience than the fact that someone was launching a heavy metal streaming service.

Quartzy overlooked two important things in the process. First, the Spotify data came from a post on Spotify’s “Insights” blog in April 2015. The Billboard article mentioned that the data came from 2015, because they were using the data as background. Quartzy made it the lead and how do you lead with two-year-old information on a constantly updating web site? By not mentioning that it’s two years old, obviously.

This is amateurish and disingenuous, but not at all the worst aspect of this story. To me, the worst part is the statistical sleight of hand perpetuated by Spotify and worsened by Quartzy when converted into that grabby headline about metal being music’s “most-loved genre.”

Spotify never said this, to begin with. What Spotify claimed was that heavy metal is the genre with the “most loyal listeners.” And this is, in fact, what Quartzy reports if you read the article. If you just read the headline, however, you would miss this distinction. Quartzy‘s faulty transfiguration of “most loyal listeners” into “most loved genre” is an all-too-common presentational sin in the age of online “journalism” (which I leave in quotes for good reason), but I’d say is no worse a bungle than Spotify’s muddling of its data in the first place and Quartzy accepting the Spotify spin without the slightest hint of journalistic inquiry.

What is loyal and why do we care?

What, after all, does it mean that a specific genre has the “most loyal listeners”? This is a two-part question. The first is logistical, as in: how would one go about measuring this rather slippery concept in the first place? Spotify assures us they have a workable methodology. (I beg to disagree, as you’ll see. Quartzy never appeared to wonder.) The second part is existential: what does it mean for a genre to have the most loyal listeners in the first place? Is this even a thing you can be? Does it make any epistemological sense? And if so, is being a loyal listener to a genre by any meaningful measure a good thing to be? And if so, good to or for whom? (Quartzy didn’t wonder about any of this either.)

Let’s start by looking at what Spotify did to ascertain listening loyalty. First, they identified what they called “core artists” in each genre; next, they divided the number of streams each core artist had by their number of listeners. Their findings placed metal at the top, with what their chart identified as a one-to-one correspondence between streams of these core artists and their number of listeners. Quoting from the Spotify post:

“We looked for repeated listens to the core artists from each genre—the ones sitting right at the ‘center’ of the genres, as it were. So one could also reasonably conclude that jazz, EDM, classical, and blues listeners play more fringe artists from those genres.”

(FYI: Jazz, EDM, classical, and blues are all genres that had less than a 0.6 correspondence in the “core artist streams divided by listeners” formulation.)

So much bamboozlement here! To begin with: core artist streams divided by listens equals loyalty? What the what? For starters: who or what determines a “core artist” in a genre? All Spotify tells us is they determined core artists via data from The Echo Nest regarding which artists “are most central to each genre.” (The Echo Nest is a “music intelligence” company, owned by Spotify.) Overlooking the unhelpful tautology—core artists are those that are most central—I question the basic premise that genres are best represented by core artists alone in the first place. This penalizes genres in which fans are by nature curious, who routinely explore all sorts of music within a given genre. Such fans could be very loyal to a genre but elude recognition by Spotify. One could also argue that a genre only truly solidifies as a genre when it expands robustly beyond some central group of artists representing a certain musical sound. To go back and ascertain helpful information about a genre by looking only at its so-called core seems like a random decision, made only for its statistical ease than for its connection to truth.

And then there’s the underlying formula itself. I for one can’t wrap my mind around what dividing streams by listeners even does—can you? There’s no coherent meaning here; it’s not like a batting average in baseball, where hits divided by at-bats creates a clear and meaningful statistic. Let’s say all the heavy metal “core” bands had one million streams, and one millions listeners, great—what does one stream per listener mean? Nothing that seems clear. Let’s go further and say the core jazz artists had 600,000 streams and one million listeners. This would create that 0.6 correspondence mentioned above. Given that we don’t know what the one-to-one correspondence means, we don’t, now, know what six-tenths of that circumstance means either, outside of the already clear realization that jazz listeners collectively listen to fewer songs from the genre’s most mainstream artists than do people listening to metal.

What, in turn, does that mean? Not necessarily what Spotify says at all. Listeners who focus on the most mainstream artists in a genre may not be “most loyal” as much as least informed—as in, they only are aware of the most popular bands. Or, perhaps, rather than “most loyal” these core-oriented folks are nearly the opposite: music’s most casual listeners, in that they don’t care to investigate beyond the usual suspects. Is a genre filled with uninformed and/or casual listeners a genre with the most loyal listeners? I instead argue that a genre where listeners listen to all sorts of so-called “fringe” artists (see above excerpt) would be the genre with the most loyal listeners—meaning, in this case, listeners who appreciate a genre’s musical landscape enough to branch out and listen to many different versions of it. This is precisely the opposite of Spotify’s conclusion, and Quartzy, in the queasy tradition of inexpert internet posts, swallowed the self-serving corporate line without chewing.

And then let’s back up and again ask ourselves what good is it to identify a genre with the most loyal listeners in the first place? Is being loyal to a genre of social significance? Whose purposes are being served by figuring this assignation?

A cautionary tale

Clearly the only thing going on here is a sales pitch. It was first a sales pitch by Spotify, which continues to present itself as a storehouse of quality by wrapping itself in heaps of quantity, and whose musical warehouse is, for better or worse (worse, mostly, I’d say), compartmentalized by a category concept that sounds more definitive than it mostly is (i.e., “genre”). Beyond that, it was indirectly a sales pitch for anyone (hello, advertisers!) seeking to identify target-able groups of consumers by economically meaningful new ways. This is obviously something in which Facebook has been specializing, often to the detriment of civil society, and it’s something that Spotify wants us to realize it can do too. Implicit in the misbegotten message that metal has the “most loyal listeners” is the idea that these listeners can be aggregated and sold to. That was the point of the original Billboard article, after all: here’s a new streaming service for metal fans, and here’s why it’s a brilliant idea.

As for Quartzy, they have no excuse at all for conflating loyalty with “most loved” except as blatant click bait. In this case, their sales pitch is for their own web site. Because we all know that web sites that fool you into clicking through to their articles are in fact the most loved sites on the web.

Look, I know this was just a throw-away article in the throw-away world of constantly updated web sites offering ongoing posts for an audience ever-ready to click away to something more interesting. And I know no active harm was intended here, unlike what’s out there from purveyors of misinformation and lunatic conspiracies. The stakes seem very low in an article misrepresenting music listenership.

But if this is the kind of piece it’s easy enough to click past, ignore, and move on from, it’s also exactly the kind of thing that illustrates the godawful limits of a digital world ruled by algorithm and monetized eyeballs. It’s death by a thousand poorly written and reported cuts. Maybe it helps to look down and say, “Hm. Maybe I’m bleeding a little.”

And then maybe it helps to begin to see each cut as its own little cautionary tale until the long-awaited day arrives when we may collectively break free of our digital trance and re-imagine our relationship with the world at large, and with each other. A new year approaches. One can always hope.