This talk was given March 18, 2016 at the Indie Ed-Tech Data Summit at Davidson College.

“I Love My Label”: Indie versus Industry and the Future of Music Education Technology

The title of this talk could as easily be “indie versus institution” as “indie versus industry.” I have no love lost for either.

I call myself a “serial dropout,” as I’ve successfully failed to complete most stages of education: high school, undergrad, grad school. Nonetheless, I consider myself very fortunate that, even without all the proper credentials, I get to do some of the things I liked most about the “academic” part of my life: that is, I’m always learning, I’m always tackling new research projects.

I call myself a writer, and some days, when I’m feeling serious, I think of myself as a scholar. I’d like to believe that I’m pushing the boundaries, helping shape the future of my field. But what I do now – write about education technology – has nothing to do with what I studied formally as an undergraduate or a graduate student. But if nothing else, there I learned how to be a critical thinker, a thoughtful researcher, and a decent writer – and I’d contend that no matter what major you pursue in school, these are the sorts of skills all students should, ideally, come away with.

I also consider myself fortunate to have peers who believe in the value of “open scholarship” – that is, we share our work online (mostly via our blogs) in ways that bypass the paywalls of the academic publishing industry. As someone who works outside of academia, without an institutional affiliation, I can’t begin to tell you how frustrating it is to be unable to access journal articles. I always get so irked when I hear technology evangelists proclaim “You can learn anything you want on the Internet.” No, you can’t. Huge swaths of knowledge, art, science remain inaccessible; and it’s a loss for scholarship, which need not and does not only happen among those with access to a university research library or with log-in credentials to its online portal. That inaccessibility is a reflection of institutional culture, industry culture, corporate culture, copyright (that is, intellectual property laws), capitalism, and code. That is, when we talk about the future of something like “education technology” or even when we talk about the future of research and scholarship or teaching and learning, we must grapple with issues that are technological, sociological, and above all, ideological.

“Indie ed-tech” – what we’re gathered here to talk about over the next few days – is inherently ideological as it seeks to challenge much of how we’ve come to see (and perhaps even acquiesce to) a certain vision for the future of education technology. An industry vision. An institutionalized vision. Indie ed-tech invokes some of the potential that was seen in the earliest Web technologies, before things were carved up into corporate properties and well-known Internet brands: that is, the ability to share information globally, not just among researchers, scientists, and scholars within academic institutions or its disciplines, but among all of us – those working inside and outside of powerful institutions, working across disciplines, working from the margins, recognizing the contributions of those who have not necessarily been certified – by school, by society – as experts. Distributed knowledge networks, rather than centralized information repositories. “Small pieces, loosely joined.”

“Indie ed-tech” offers a model whereby students, faculty, staff, and independent scholars alike can use the “real-world” tools of the Web – not simply those built for and sanctioned by and then siloed off by schools or departments – through initiatives like Davidson Domains, enabling them to be part of online communities of scholars, artists, scientists, citizens.

I’d like to place the emphasis in my talk today on that adjective “indie” rather than on the hyphenated phrase “ed-tech.” (Yes, the Hack Education style guide insists on the hyphen.) I want us to consider: why would we, why should we embrace the “indie” rather than the “corporate,” the institutional. What difference do these adjectives – well, really, these economies, these ideologies – make to the technologies we adopt, to the technologies we are forced to adopt, to the technologies we force our students to adopt, to the direction we take, to the stories we tell about the future of education?

“Indie ed-tech” draws rather explicitly on the spirit of indie music and the DIY (do-it-yourself) ethos of punk rock.

A connection between music and education might make you wince. After all, comparisons between the music industry and universities are fairly commonplace and not always particularly helpful. The Internet, so one popular story goes, will be higher education’s “Napster moment,” hastening the “disruption” of the institution. According to this particular narrative: just as the development of the MP3 file format and of portable digital music players destroyed the music industry, digital content and communication will soon destroy the university. (The timelines vary. Harvard Business School professor Clayton Christensen famously predicted in 2012 that half of US universities would be bankrupt in a decade. So I guess we’ve got about five and a half years left, by his calculations.)

Of course, the music industry hasn’t been destroyed – that’s just one problem with the whole “Napster moment” cautionary tale. Certainly it has changed.

“Indie ed-tech” posits a different narrative than one where venture-funded MOOCs are, as Clay Shirky put it, the “lightning strike on a rotten tree.” Indie ed-tech offers a narrative that draws on a history of how new technologies shaped indie music in particular. This history might suggest something about what new technologies can and do offer teaching, learning, and research – one that isn’t about doom and disruption (unless we’re talking doom and disruption for those corporate ed-tech executives who want to control the learning playlist).

I want, in part, to build today on that story as told by my colleagues Reclaim Hosting’s Jim Groom and the University of Oklahoma’s Adam Croom. They gave a joint presentation at a conference at Stanford last fall; I wasn’t there, but they both posted their presentations to their blogs, making it easy for me to build on their work. I’ll post this talk to my blog in turn…

So let me briefly recap their argument, which contends that the Napster narrative doesn’t give a full or particularly accurate picture of how digital technologies have changed the music industry: Adam Croom cites UIC communications professor Steve Jones, who says that “The real revolution in popular music in regard to the Internet is to be found in the availability of news, information, and discussion about music and musicians facilitated by Internet media.” The barriers to entry – set high for musicians by radio and television in turn and for music commentators by magazines – have been lowered (somewhat). “In this sense,” Croom adds, “technology is not the industry disruptor, but rather the opportunity to facilitate the building of a community inconceivable prior to its existence.”

New technologies have also changed the production of music, again lowering the barriers to entry (again, somewhat). No longer is it necessary to employ huge teams of people in a cumbersome and expensive process that – ostensibly at least – is best controlled by a record company, as Lars Ullrich famously argued when he testified in front of Congress as to why Napster and the potential to share digital music would destroy music itself. Nowadays – again, ostensibly – one can make music, record that music, distribute that music more easily, without a major label. Hell, you don’t even need a real drummer, sorry Lars. And indeed, according to some figures at least, the number of people who identify as independent artists and musicians has grown substantially since the advent of the Internet.

(Our challenge remains – in indie music and in indie scholarship and in indie ed-tech – to make this path one that is economically and emotionally sustainable.)

For his part, in that Stanford talk, Jim Groom pointed to 80s indie punk as a source of inspiration for indie ed-tech. “Why 1980s indie punk?” Groom explains,

First and foremost because I dig it. But secondly it provides an interesting parallel for what we might consider Indie Edtech. Indie punk represents a staunchly independent, iconoclastic, and DIY approach to music which encompasses many of the principles we aspired to when creating open, accessible networks for teaching and learning at [the University of Mary Washington]. Make it open source, cheap, and true alternatives [sic] to the pre-packaged learning management systems that had hijacked innovation.

I want to latch onto two parts of that last sentence.

One, “hijacked” – this is how institutions and industries often operate when they’re confronted with subversion and innovation. They hijack and disarm it. (That’s how capitalism works. That’s how ideology works.) For its part, the indie punk scene in the US had seen what had happened to the first generation of British punks – how they’d been signed and defanged by major record labels – and they vowed to retain control of their own music and community. For ed-tech’s part, Seymour Papert, one of ed-tech’s most hopeful and subversive radicals, observed this same tendency when he wrote in 1993 that

Little by little the subversive features of the computer were eroded away: Instead of cutting across and so challenging the very idea of subject boundaries, the computer now defined a new subject; instead of changing the emphasis from impersonal curriculum to excited live exploration by students, the computer was now used to reinforce School’s ways. What had started as a subversive instrument of change was neutralized by the system and converted into an instrument of consolidation.

And two: what Groom calls “the pre-packaged learning management system.”

Over the course of the past twenty years, the learning management system has become a cornerstone of education technology – how it’s engineered, how it’s purchased, how it’s implemented – of how schools and their staff administer the course schedule, the grading, the class discussions, the “content delivery,” and so on. It has, perhaps most damagingly I’d contend, become the cornerstone of our imagination – shaping our expectations of what education technology “looks like,” how it functions, to what end, and to whose benefit. The learning management system has become a behemoth, an industry unto itself, part of a larger behemoth of an increasingly technologized university.

Indeed, rather than being largely unaffected by the Internet – an argument that you’ll hear Silicon Valley types frequently make about education, forgetting I suppose, that universities helped invent the Internet – the learning management system has prepackaged the university for the Internet, circumscribing scholarship, pedagogy, communication.

The LMS is our major record label. Prepackaged software. A prepackaged sound.

Like all sorts of industries, music and education alike are now supposed to bow to the new insights discoverable thanks to “big data.” Algorithms and analytics will “personalize” our world, we’re told. The problem, of course, is that the algorithms and the analytics also make everything sound the same. That’s the business: a neoliberal mirage around “choice.” Standardization. A familiar and managed image. Profiled. Labeled.

Now, the music business has actually been tracking “data” for a long, long time. Album sales, the singles most frequently played by disc jockeys, the songs most frequently played in jukeboxes – this information has been used to calculate who’s “top of the charts.” It’s been used to track what sorts of sounds people like, and then in turn, to sign more artists that sound just like that.

The music industry likes to highlight the likes of famous “A&R men” who identified and signed artists who found new sounds, who did reshape commercial music, most famously perhaps, John Hammond who “discovered” Billie Holiday and Bob Dylan and Gary Gersh who signed Nirvana. These two choices that seemingly went against the grain of what was being played on the radio, what was selling. But most music executives – now and then – have chosen instead to look for musical acts that replicate what’s already popular.

New music – new commercial music, that is – tends to not surprise us. It has been selected because we are already comfortable hearing it or very much something like it. Once something sells, than we hear it and echoes of it again and again and again and again.

And nowadays it isn’t just “the gut feeling” or “the ear” of the A&R exec that’s poised to perpetuate the bland and standardized sound of popular music. It’s an algorithm.

Indeed, the move to digital music provides new opportunities for data collection and data mining. No one – well, except my parents, I guess – knew how many times I played that 45 of Autograph’s “Turn Up the Radio,” how many times I rewound the cassette to replay Guns & Roses’ “Welcome to the Jungle.” But now the software knows – and remember, whether we stream music or download it, we don’t own this music; we’ve just paid for a license to play it. When we sign the Terms of Use for music services, we likely sign away our ability to keep private what we’re listening to, where we’re listening to it. Sometimes we volunteer even more metadata, labeling our playlists things like “breakup songs 2011” or “super productive writing music” or “best bicycling tunes” or “feeling low.”

This data is incredibly valuable to an industry that wants us to buy and listen to its products.

Using various data sources, researchers at several different universities have boasted that they’ve developed algorithms that can actually predict hit songs – what’s been the “Holy Grail” for a long while in the music industry: identifying the perfect song. One algorithm, for example, uses the popularity of music on bit torrent sites, along with the geography and the query strings of users of those sites. Many algorithms use Echo Nest, a research project itself spun out of the MIT Media Lab that utilizes “machine listening” and “machine learning” to categorize and analyze a song’s acoustic and textual content. Echo Nest, which was acquired by Spotify in 2014, helps the latter’s power music recommendations. Using Echo Nest, other researchers boast they’ve developed an algorithm to predict the probability of whether or not a song would be a top 10 hit – this algorithm did accurately predict the probability of the songs that eventually made the top 10 Billboard Hot Dance/Electronic Songs of 2015. This particular algorithm takes into account things like song length, tempo, key, and “danceability” – whatever the hell that means.

Well, what that means has something to do with successfully predicting that Skrillex, Diplo, and Justin Bieber’s “Where Are Ü Now” would be a hit. Interestingly, when reviewing Skrillex/Diplo album, Pitchfork’s senior editor described the track as “unexpected in all the best ways.”

“Unexpected” and yet algorithmically predictable – the future of music technology.

These sorts of predictive algorithms are, of course, quickly “productized” – to varying success. That’s no surprise as the popular narrative now insists that it’s preferable to avoid decisions once made on instinct and instead take a “Moneyball” approach – that is, it’s best to analyze the statistics and performance data that your competitors aren’t looking for so you can sign talent more cheaply, an approach made famous by the Oakland As’ manager Billy Beane. “Moneyball” trumps “gut,” or so we’re told.

Now, the music industry spends billions on A&R marketing and promotion. In 2013 alone, it spent $4.3 billion on this – that’s almost a third of the industry’s revenue. So there’s a huge financial motivation to sign artists that will produce hits, clearly, and to identify “the sound” – a pre-packaged sound – that can be replicated into multiple hits, from one or from multiple artists. Algorithms and analytics purport to offer solutions to make decision-making more better, faster, cheaper, more efficient, more scientific. More “personalized.”

“Data-driven decision making.” Sounds familiar, right? Analytics and predictive modeling have also come to education and education technology, of course, with much of this based on the kinds of data that the learning management system and student information system and related software collects: attendance, number of log-ins, number of interactions with a professor, and so on. Some of these products promise to identify algorithmically students who are “at risk” – that is, those who might not complete a course or graduate on time. Other products promise to organize algorithmically the educational content that has been deemed algorithmically the most appropriate to an individual student’s needs as she or he moves through a particular college course. Other products promise to identify algorithmically the courses themselves that individual students will perform best in.

Degree Compass, a piece of software developed by Austin Peay University and acquired in 2013 by the learning management system provider Desire2Learn, is one example of the latter. It is, and I quote from its website, “a personalized course recommendation tool that helps advisors and students plan the most optimal path to graduation using predictive analytics. Tapping into a volume of historical data, the predictive algorithm guides course selection in a way that improves academic success and drives on-time degree completion.” But just like the predictive modeling in music, this process should prompt us to ask a lot of questions about what feeds that algorithm and what are the results: What sorts of classes get recommended? Are students offered something that sounds familiar, comfortable? What signals to the algorithm what a student might find familiar? What happens in the face of an algorithmic education to intellectual curiosity? To risk-taking, to exploration, experimentation, play? To the major that many of us pursue for a while, “Undecided.” Does the educational system as-is, with or without an algorithm, value these things? And what happens when classes are devised in order to perform well according to this algorithm?

Pre-packaged sound. Pre-packaged courses. Pre-packaged students.

The phrase “I Love My Label” in the title of this talk has multiple meanings. It is, as perhaps some of you recognize, the name of a bitingly tongue-in-cheek song by Nick Lowe, later covered by the band Wilco.

I love my label …

And my label’s got high hopes in me

“I Love My Label” is just one of many, many songs that excoriates a major record label or a record executive. There’s The Sex Pistols’ “E.M.I.” There’s The Clash’s “Complete Control.” The Smiths’ “Paint a Vulgar Picture.” “Mercury Poisoning” by Graham Parker and the Rumour. “One Down” by Ben Folds. “Jimmy Iovine” by Macklemore and Ryan Lewis. “The Under Assistant West Coast Promotion Man” by The Rolling Stones. “Have a Cigar” by Pink Floyd. “Joe” by Tom Petty and the Heartbreakers. “Fools” by Motorhead. After it tried to edit out shots of her bare stomach from a music video, Amanda Palmer pleaded for her label Roadrunner Records to “Please Drop Me.” “9000 dollars, that’s all we could win,” Lynyrd Skynyrd sing in on a song from their second album “Workin’ for M.C.A.” (Their first album contains the classic “Free Bird” – how money do you think M.C.A. has made from that song alone?) “But we smiled at the Yankee Slicker with a big ol’ Southern grin,” the song continues. “They’re gonna take me out to California gonna make me a superstar.” Following almost a decade of not releasing an album and a lengthy legal dispute with his old label, Fantasy Records, John Fogerty released “Zanz Kant Danz” in 1985 – a song about a pig that “can’t dance but he’ll steal your money.” In response, Fantasy Records boss Saul Zaentz sued Fogerty for defamation of character. And as Fogerty had previously sold all the rights to his Creedence Clearwater Revival songs to Fantasy Records, the label also sued Fogerty for plagiarism, claiming that another song on his new album, “Old Man Down the Road,” was merely the Creedence Clearwater Revival song “Run Through the Jungle” – a song that Fogerty himself had written – with a new name.

“I Love My Label.” These songs – and so many others – all comment on the music industry’s practices of exploitation and control: who owns the artist’s lyrics, who owns the artist’s image, who owns the artist’s sound? What position are unsigned artists in to review, rewrite, rebuke the legal contracts offered to them by major record labels?

What position are we in to review, rewrite, rebuke the legal contracts now offered to us by major (education) technology companies? That is, of course, the Terms of Service that we all click on without reading. Who owns the student or scholar’s work? Who owns the student or scholar’s data? Who’s been granted a license to use it, display it, data mine it, feed it into an algorithm? And an algorithm to what end?

What position are we in – artists, musicians, scholars, students, citizens – to challenge algorithmic decision-making, something that is even more opaque and “black-boxed.”

“I Love My Label.” I chose that song as the title of this talk because of some of the other meanings of the word “label” too – it isn’t just a company that produces records; it’s a word used to describe or categorize someone. Labeling. It’s the process that takes place in order to build the profiles against which we’re sold advertising, recommended movies on Netflix or songs on Spotify, and, if you buy the software, told which content model to take next or which course to take next in order to graduate quickly with the highest possible grade point average. “Personalization” might sound like it’s designed especially for us; but “personalization” is an algorithm based on a profile, on a category, on a label.

I’ve argued elsewhere, drawing on a phrase by cyborg anthropologist Amber Case, that many of the industry-provided educational technologies we use create and reinforce a “templated self,” restricting the ways in which we present ourselves and perform our identities through their very technical architecture. The learning management system is a fine example of this, particularly with its “permissions” that shape who gets to participate and how, who gets to create, review, assess data and content. Algorithmic profiling now will be layered on top of these templated selves in ed-tech – the results, again: the pre-packaged student.

Indie ed-tech, much like the indie music from which it takes its inspiration, seeks to offer an alternative to the algorithms, the labels, the templates, the profiling, the extraction, the exploitation, the control. It’s a big task – an idealistic one, no doubt. But as the book Our Band Could Be Your Life, which chronicles the American indie music scene of the 1980s (and upon which Jim Groom drew for his talk on indie-ed tech last fall), notes, “Black Flag was among the first bands to suggest that if you didn’t like ‘the system,’ you should simply create one of your own.” If we don’t like ‘the system’ of ed-tech, we should create one of our own.

It’s actually not beyond our reach to do so.

We’re already working in pockets doing just that, with various projects to claim and reclaim and wire and rewire the Web so that it’s more just, more open, less exploitative, and counterintuitively perhaps less “personalized.” “The internet is shit today,” Pirate Bay founder Peter Sunde said last year. “It’s broken. It was probably always broken, but it’s worse than ever.” We can certainly say the same for education technology, with its long history of control, measurement, standardization.

We aren’t going to make it better by becoming corporate rockstars. This fundamental brokenness means we can’t really trust those who call for a “Napster moment” for education or those who hail the coming Internet/industrial revolution for schools. Indie means we don’t need millions of dollars, but it does mean we need community. We need a space to be unpredictable, for knowledge to be emergent not algorithmically fed to us. We need intellectual curiosity and serendipity – we need it from scholars and from students. We don’t need intellectual discovery to be trademarked, to a tab that we click on to be fed the latest industry updates, what the powerful, well-funded people think we should know or think we should become.

Liner Notes: The album covers in this presentation come from Rage Against the Machine, Pink Floyd, Minor Threat, Black Mountain, Metallica, Sonic Youth (well, Giulia Forsythe), The Antlers, The Sex Pistols, The Flaming Lips, Gorillaz, Autograph, Yo La Tengo, TV on the Radio, The Clash, Radiohead, Michael Jackson, Black Flag, and Daft Punk (with apologies). The hair metal “gods” include Bon Jovi, Warrant, Poison, Mötley Crüe, RATT, Whitesnake, Cinderella, and Stryper. The ‘zine art first appeared in Sideburns (not Sniffin’ Glue as it’s often credited). Fair Use FTW. The full slide deck is available via SpeakerDeck. You can listen to the “I Love My Label” playlist on Spotify, but you should support artists by buying their music. Unless it's Metallica. Then share freely.