Monetizing a Meme: YouTube, Content ID, and the Harlem Shake

YouTube was able to attract a vibrant community of amateur video makers and vloggers (video bloggers) in its early years, as has been well documented by scholars like Jean Burgess and Joshua Green (2009) and Michael Wesch (2008). Much of YouTube’s content was created and uploaded by amateur video makers and uploaders, alongside a smaller but growing network of semi-professionally and professionally produced vlogs, programs, and Multi-Channel Networks (MCNs). In part because YouTube was easy to use, free, and designed to facilitate cross-platform sharing, it quickly established its popularity and drew millions of users in its first year.

YouTube, like comparable “free-to-use” social media sites like Facebook, MySpace, Twitter, Pinterest, Yelp, and others, built their platform around a basic bargain: they provide a free and open platform, users provide the content, and the platform owners could monetize the individualized data gathered from the social and creative activities, interests, and communication of users by selling those data to advertisers. Insofar as it is known or understood by users, this “basic bargain” seems acceptable, and users continue to flood social media platforms with activity and content. This immense assemblage of decentralized social energy and creativity not only provided a new approach to personalized and data-driven advertising but also, perhaps more importantly, allowed these social media platforms to gain market dominance—in some cases crowding out competitors to the point of near hegemonic techno-social power.

Mark Andrejevic (2009) and Toby Miller (2009) have criticized this “basic bargain” for social media users as inherently exploitative, suggesting that what appeared to be a free and open cultural space was actually a brilliant way to have user producers, or what Axel Bruns (2008) dubbed “produsers,” create all the content and thus the economic value, all while gathering their data and providing advertisers with an ever more sophisticated mechanism for reaching consumers. This sentiment was expressed early on by Nicholas Carr (2006), who pointedly dubbed this new form of alleged exploitation “digital sharecropping.” Yet, critiquing the tenuous “labor” relationship of users and corporations on social media as “exploitative” often falls flat for users and critics alike. Part of what has helped avoid a sense of exploitation while maintaining the seeming expressive freedom, social value, and even idealistic enthusiasm for YouTube and other dominant social media platforms has been the sheer volume of creativity they seem to enable. Numerous scholars have lauded the “remix culture” (Lessig, 2008) and “mass amateurism” (Shirky, 2008) that social media platforms like YouTube seem to foster. Amid the celebratory popular commentary and scholarship regarding the “democratic” and “participatory” nature of the Internet and social media, the last few years have seen rising political economic critiques which seek to reconnect discussions of the Internet and social media to the economic logics of capitalism and neoliberalism, including from Robert McChesney (2013), Christian Fuchs (2014), and Trebor Scholz (2012).

Limor Shifman (2014) defines an “Internet meme” as “(a) a group of digital items sharing common characteristics of content, form, and/or stance, which (b) were created with awareness of each other, and (c) were circulated, imitated, and/or transformed via the Internet by many users” (p. 41). In this way, a meme is differentiated from something that is “viral,” mainly in that a meme is a collectively (re)produced phenomena while viral refers to a single video, image, or text that is shared widely. While this difference is useful, Shifman notes that we should think about the viral and the memetic as a “dynamic spectrum rather than a binary dichotomy” (p. 56). That is, while a viral video generally refers to a single video that has been shared and viewed extensively, viral videos often provoke a real-time ripple of remakes, remixes, parodies, and other memetic content. For example, Psy’s “Gangnam Style” went viral, with more than 2 billion views and counting, and it also spawned a multitude of memetic content that parodied, referenced, or remixed the original. On YouTube, Shifman (2011) emphasizes that a video meme is often “a popular clip that lures extensive creative user engagement in the form of parody, pastiche, mash-ups, or other derivative work” (p. 190).

This “extensive user engagement” is often overlooked when considering the tallies of likes, views, and shares rather than a consideration of the incredible amount of effort and labor that composes these cultural phenomena. The celebratory nature of memes such as the “Harlem Shake” has the tendency to collapse the history and the capitalization of participatory culture (as well as overlook the determinants shaping these phenomena) into a footnote, letting the energy of the memetic phenomena take center stage. It is hard to envision the creative labor that generates the phenomena. The “Harlem Shake” and other YouTube memes like it are an immense collection of heavily monetized creative work, often with no financial compensation for the individual video producers, as well as an unclear, tentative, and shifting set of rights and control mechanisms over content.

Harry Rodrigues, known as Baauer, a 23-year-old American DJ and producer released a track titled “Harlem Shake” on 22 May 22 2012. Offered as a free download by his label Mad Decent, the song was well received within it is genre, with the influential online music criticism site Pitchfork awarding it “Best New Track,” describing it as having “an irresistible appeal” and “a purely visceral pleasure” (Fitzmaurice, 2012). The song was also featured as an unreleased track on BBC Radio 1’s “Essential Mix,” an internationally popular weekly radio broadcast of contemporary DJs and electronic dance music.

Although somewhat popular in the world of electronic dance music, the viral spread of Baauer’s song would be fueled not by its inclusion in DJ set lists or streaming services but from its prominent use in a YouTube video posted 8 months after its release. On 30 January 2013, “Filthy Frank,” the YouTube persona of a 19-year-old Communication major in New York, uploaded a new video titled “FILTHY COMPILATION #6—SMELL MY FINGERS.” The video began with a 19-s clip of Filthy Frank and three friends, all dressed in skin tight spandex body suits and doing a silly dance in their dorm room; essentially a combination of pelvic thrusts and flailing arms—all to Baauer’s Harlem Shake. In an interview with Naomi Zeichner (2013) in The Fader, “Filthy Frank” explained the inspiration for the video clip that would spark the Harlem Shake meme:

I was in a room with a few people. One of my friends was just playing the song on the speakers and I asked what [it was], and it just happened to be “Harlem Shake.” As soon as the drop of the song came, we just started going crazy. We thought, well, we could turn this into something good.

Within hours of its upload, YouTube user “Gam3xpert” posted a remix video titled “Baauer—Harlem Shake (Filthy Frank Style)” to YouTube that looped the 19-s clip for nearly 4 min. A few days later, a group of Australian high school students known on YouTube as TheSunnyCoastSkate or TSCS uploaded “The Harlem Shake v1 (TSCS original).” Building off of Filthy Frank’s 30 s crazy dance, TheSunnyCoastSkate’s version featured a lone person wearing a motorcycle helmet dancing a-la-Filthy Frank style (pelvic thrusts) along to the clip of Baauer’s Harlem Shake; about 15 s in, as Baauer’s track drops “do the Harlem Shake,” the video jump cuts³ into a crazy dance party of extreme silliness (and more pelvic thrusts). TheSunnyCoastSkate’s (2013) video became the template for the pending dance phenomenon.

Although merely 2 hr after TheSunnyCoastSkate uploaded their video, Filthy Frank uploaded “DO THE HARLEM SHAKE (ORIGINAL)” (DizastaMusic, 2013); it was “The Harlem Shake v2,” based on the video from TheSunnyCoastSkate (2013) and uploaded on the same day by YouTube user PHLOn NAN (2013), that was picked up and posted by the influential web culture site BuzzFeed. Perhaps with no knowledge of Filthy Frank’s original, PHLOn NAN (2013) noted in the description of their video “ALL CREDIT GOES TO THESUNNYCOASTSKATE.” BuzzFeed’s posting of “The Harlem Shake v2” helped to garner it 300,000 views in 24 hr (Broderick, 2013).

Nearly at the same time, another version, “The Harlem Shake v3 (office edition),” created by YouTuber “hiimrawrn,” who was also an employee of Maker Studios, an influential YouTube MCN, was submitted and upvoted to the front page of Reddit, generating additional momentum.⁴ Redditors, historically a source of many Internet memes, commented widely about its potential meme power (Figure 1).

Over the next week, the Harlem Shake meme was spawning remakes from all over the world, filmed in playgrounds, offices, high school cafeterias, and many, many college dorm rooms. The Norwegian army made a version “Harlem Shake (Original Army Edition),” which became the most watched Harlem Shake video on YouTube, eventually garnering over 100 million views (Håkonsen, 2013). Even BuzzFeed made their own “office version,” as did College Humor—who credited for inspiration Filthy Frank, TheSunnyCoastSkate, and PHLOn NAN.

Popular content aggregator BuzzFeed played an important role in establishing the viral sensation, not only promoting specific videos and contributing their own but also framing “The Harlem Shake” as a viral sensation, encouraging users to join in on the fun by producing their own. In their 7 February post “Have You Done A Harlem Shake Video Yet?,” Buzzfeed laid out the steps for reproducing the Harlem Shake meme (Broderick, 2013):

Don’t know how to do it?

Step 1: First, play this song

Step 2: Have one person in a mask dance solo

Step 3: When it goes “do the Harlem Shake,” have everyone in the room start dancing. It’s super simple!

The Harlem Shake meme reached a fever pitch by 12 February, but within another week quickly petered out, with many pronouncing the meme “dead” after mainstream TV programs and corporate advertisers began co-opting the meme. To date, Filthy Frank’s “DO THE HARLEM SHAKE (ORIGINAL)” has received more than 50 million views on YouTube, TheSunnyCoastSkate “The Harlem Shake v1 (TSCS original)” more than 23 million, and PHLOn NAN’s “The Harlem Shake v2” more than 12 million.^5,6 Compilation videos, such as “The Best of Harlem Shake” and “TOP 10—BEST HARLEM SHAKE VERSIONS” have garnered additional tens of millions of views each. Collectively, the tens of thousands of Harlem Shake videos have generated well over a billion views.

The Harlem Shake meme was collectively pieced together by a network of relatively unknown YouTube users, each adding to the reproducibility of the meme by adding new ideas and forms. Its quick growth seems to have been sparked by Redditors upvoting one of the early versions, then fueled by Buzzfeed.

Notably missing from meme’s conception was Baauer, whose song seemed to almost incidentally provide both the musical backbone and the name. This odd sequence of events was noted by Filthy Frank, who, after the initial fervor of the meme died down (and having amassing tens of millions of views to his original video), tweeted at Baauer with some irony (Figure 2).

Baauer did not respond.

Yet according to Content ID, YouTube’s automated system to detect copyrighted material, the audio was the only piece of verifiable intellectual property that might warrant an ownership claim and monetization. Baauer and his label Mad Decent, although missing from the generation and spread of the meme, would not remain silent for long. The massive global popularity of the Harlem Shake meme gave his label powerful leverage, and even control, over the distribution and monetization of the meme.

While the speed with which the Harlem Shake meme was born, spread, and died was remarkable, dance crazes are nothing new. Looking at dance crazes well before the Internet reveals the complex tensions between grassroots generativity and commercial exploitation. The Tango, The Twist, The Mashed Potato, The Hustle, The Macarena—these are just a few of the dance crazes that swept through the dance halls, living rooms, and discotheques of the United States, and often the world, throughout the 20th century. They were memes in the pre-Internet, Richard Dawkins’ sense of the term—pieces of culture that successfully spread, replicated, and transformed into popular phenomena. The term “meme” was coined by Dawkins (1989) in the book The Selfish Gene, as a way to conceptualize a “unit of cultural transmission” (p. 192), that acts like a “gene” in the method of genetic transmission—simply stated, a meme as a cultural idea or phenomena that spreads.

Dance crazes, or perhaps we could say “dance memes,” emerge organically from a certain subculture and ultimately morph into the mainstream, not unlike video memes like the Harlem Shake. Take, for example, The Hustle. The Hustle’s origins go back to a Puerto Rican dance style that emerged in the South Bronx in the early 1970s. By 1974, it became known as “Spanish Hustle” (also known as the “Latin Hustle”), and in 1975, the Fatback Band made a song with that name. Building on this grassroots dance style, Van McCoy and the Soul City Symphony’s chart topping hit song “The Hustle” took the emerging dance and propelled it into a mainstream dance craze that swept the disco club scene in 1975. Along with the dance craze, the song rocketed to the top of the Billboard Pop Singles chart. The Hustle was enshrined as a dance standard when it was featured in the movie Saturday Night Fever in 1977.

The original Harlem Shake was a 1990s dance style from which Baauer’s song Harlem Shake and later the Harlem Shake meme inadvertently take their names. The dance style emerged collectively from the unique hip-hop culture of Harlem and was utilized and transformed by hip-hop artists who brought the dance into the mainstream—from Al B’s “drunkin mummy” dance in the 1980s to hip-hop music videos of the early 2000s. Nobody considered the Harlem Shake dance something someone “owned” or that any one person had created. It was an existing cultural form and successful dance meme that could be freely used, shared, and transformed. Each time someone performed the dance, or used it in a music video, their actions contributed to the dance form, as well as the larger hip-hop culture. Certainly, hip-hop artists sold records and gained fame based in part on songs featuring the Harlem Shake, but the dance itself was a public good of sorts. The Harlem Shake would help spawn other similar dances in that time period, such the “Chicken Noodle Soup,” which combined the dance styles of the Harlem Shake and the Toe Wop.⁷

In 2006, within the first year of YouTube’s popular existence, two teenage hip-hop artists, DJ Webstar and Young B, both from Harlem, recorded a song called “The Chicken Noodle Soup,” which became popular among teenage hip-hop fans, who began uploading videos of their dances, creating YouTube’s first Internet dance craze. The dance craze spread, thanks to hip-hop enthusiasts emulating the dance on YouTube (with millions of views) and linking to their videos on blogs and Internet forums.⁸ Writing in The Phoenix, Carly Carioli (2006) explains how The Chicken Noodle Soup dance spread online:

In the old days, dance crazes migrated slowly—block by block, dance floor by dance floor, eventually city by city, like a disease, until a song or a movie came along to spread the instructions to the masses. Most of the good dances like the Fila, the Whop, and the Pee-Wee Herman were dead before music videos began speeding up the cycle—thanks to which we got the Macarena, the electric slide, and the cabbage patch. But now we’ve got YouTube, thanks to which you don’t need to wait for someone to discover “Chicken Noodle Soup,” press the song onto vinyl, sell it to a major label, and have Hype Williams produce the video. Now you can see 14-year-olds doing the Chicken Noodle Soup before the song is even on the radio outside of NYC.

Nobody “owned” the Harlem Shake or the Chicken Noodle Soup dance styles. They did not have any authors. Both emerged collectively in Harlem, emanating out through the larger hip-hop community and into the mainstream. This process mirrors hip-hop in general, which has depended heavily on sampling, remixing, sharing, and collaboration between artists and borrowing from other genres. Understanding the collaborative and networked spread of authorless dance crazes helps us understand the similar, but much faster, transient, and more global spread of dance memes and other memetic phenomena in the digital era, as well as the growing tension between open cultural production and the potential for massive corporate profits predicated on intellectual property claims.

Baauer’s EDM song “the Harlem Shake” itself depends heavily on sampled music. The two most crucial samples which form the base of the song (and the dance meme) “do the Harlem Shake” and “con los terroristas” are from two different songs, sampled without attribution or permission. While answering questions in a Reddit “Ask Me Anything” (AMA) forum, Baauer was asked where he got the “con los terroristas” sample. He responded as if he didn’t know, remarking as shown in Figure 3

Despite Baauer’s ignorance (feigned or not), Redditors quickly figured out the origination of the sample, posting to his AMA thread (Figure 4).

It is curious whether Baauer really had no idea where the sample came from or whether he was intentionally feigning ignorance as he knew that now that his song had gone viral along with the meme, there would be greater scrutiny of its copyright claims. As Redditors pointed out, the “con los terroristas” sample comes from a 2006 song “Maldades” by Puerto Rican reggaeton artist Héctor “El Father” Delgado. In a radio interview, Delgado noted that he only learned about this Harlem Shake meme when a friend phoned him to tell him and that he was planning on suing Baauer for the unlicensed use of his track (UrbatonMusic, 2013). The other key sample comes from the 2001 song “Miller Time” by Plastic Little, a relatively unknown rap group. Due to the unlicensed sampling, Baauer’s label Mad Decent (a small independent label) had to cut deals with both Delgado and Musson (Zeichner, 2013).

The commonplace notion of “authorship” as either an individual or group of individuals laying claim to a work, already on shaky ground with EDM music, seems to fall short when attempting to encapsulate the large collections of digital labor that go into Internet memes. Internet memes are rapid, ethereal, produced by often anonymous nodes through networked practices that transform as they replicate. Unlike the polished, finished pieces of authored and produced creative works at the heart of corporate cultural commodities, memes are not the end of a process, and their power is drawn from the process of sharing and replication itself. When it comes to Internet memes and memetic digital culture, Patrick Davison (2012) notes, “with no documented authors, there exists no intellectual property” (p. 132). Despite the lack of authorship in the production of much digital culture, Davison points out that many of the popular platforms of the Internet “preserve and extend a historical prioritizing of attribution and authorship” (p. 130). The distributed and networked nature of authorship for digital cultural production, and memes in particular, runs against the legal premise of contemporary intellectual property.

While video memes on YouTube may be largely authorless, content on YouTube is governed by copyright policies designed for traditional notions of discrete authorship and ownership. YouTube’s copyright matching system, Content ID, automatically scans every uploaded YouTube video for copyrighted material. Copyright holders submit copies of music tracks, films, television shows, and other media to YouTube’s Content ID for matching. When a match is found, the system instantly applies the copyright claimant’s preference for the matched content: it can be blocked, tracked, or the far more popular option, monetized (via pre-roll and overlay ads). YouTube splits ad revenue with and the rights holder, each receiving somewhere between 40% and 50%. The system is highly sensitive and can match even brief snippets of audio or video, or barely audible background music in a home video recording.

Content ID was designed and implemented in response to increasing pressure from media corporations concerned about the unauthorized sharing and broadcast of their copyrighted media content, culminating in Viacom’s extravagant US$1 billion lawsuit against YouTube in 2007 (which was ultimately settled in 2014 without financial payout). Despite being legally protected from responsibility for the hosting of unauthorized copyrighted material as an intermediary by the Digital Millennium Copyright Act (DMCA), YouTube in its early years was seen as a threat to media content industries. Having purchased the unprofitable YouTube in 2006, Google infused YouTube’s early Content ID system with “tens of millions” of dollars to build out and improve the system, gaining the confidence of copyright holders.⁹ By 2015, monetization through Content ID has been immensely successful in not only assuaging fears of media companies but also, more importantly, establishing a new revenue stream: the ability to profit from the massive and growing pool of creative (and free) work of users. YouTube reports having paid out more than 1 billion in revenue to claimants through Content ID since its inception.¹⁰

Uploaders with matched content have little say in the automatic decision of Content ID. While YouTube does provide for users to file a wrongful claim against the match, or argue their use of the music should be protected as a fair use, the process is unclear and complicated. Even in clear cases of fair use, it can often require months as well as legal help and expert knowledge of copyright law to achieve a successful fair use claim.¹¹ Regarding Content ID, Niva Elkin-Koren (2014) explains, “In the case of erroneously blocked content, a user whose content has been blocked may have no effective recourse” (p. 47). All the while, YouTube uses threatening language about “strikes” and “blocked accounts” to dissuade users from challenging the automated decision-making process. For most, it is probably easier to simply comply with Content ID’s regime of control and monetization.

Despite the strong arm of copyright law, particularly the strengthening of it by the DMCA bearing down on UGC, Guy Pessach (2013) notes that YouTube actually was originally able to gain market advantage through the DMCA’s safe harbor provisions, which protected sites from getting in trouble for user-uploaded content so long as they complied with DMCA takedown requests because it “enabled the hosting and public provision of endless amounts of popular copyrighted cultural materials,” (p. 51) and because it was (and still is) the charge of the copyright holder to generate the takedown notice. In essence, YouTube’s ability to respond to and act quickly (through their automated systems) to DMCA takedowns combined with the safe harbor protections allowed YouTube’s rise to the dominant video hosting service it is today.

YouTube achieved this market dominance in no small part through taking advantage of copyright law that allowed for YouTube to adopt new business models “based on collaboration and revenue-sharing with creators and rights owners, only now from a completely different negotiation (or one may say, coercive) position,” as not only does YouTube hold a “significant portion of audience attention” but is also “partially shielded from legal liability” for hosting copyrighted materials (Pessach, 2013: 52). These factors together mean that “authors, creators and performers have very few options other than agreeing to YouTube’s terms and conditions or vanishing from audiences’ awareness” (p. 52). In short, YouTube harnessed its ability to host copyrighted materials with limited legal liability into achieving market dominance and then harnessed its market dominance into the ability to force licensing on both content creators and copyright holders.

Despite the fact that YouTube’s automatic Content ID system allows copyright owners to immediately claim monetization of UGC with potential infringement, Ronak Patel (2013) argues that Internet memes could be protected from copyright liability (and therefore from control through the Content ID system) through a Fair Use argument because they “effectuate cultural interchange” and that “protecting memes responds to a market failure” (p. 256). However, this argument becomes increasingly difficult for content creators to implement effectively due to YouTube’s control mechanisms.

Although fair use has been around as a concept for many years, it was finally outlined in the Copyright Act of 1976 (under 17 USC § 107) as a way to allow use of copyrighted work without infringement. However, “Fair Use” is not specifically defined as much as it is outlined as a way to determine how “fair” a use is utilizing four factors: (1) purpose and character, (2) nature of copyrighted work, (3) amount of the portion related to the whole work, and (4) the effect of the use upon the potential market value. Furthermore, “Fair Use” is always a defense against a claimant, expecting the user of the material to be able to understand this seemingly complex system to defend themselves rather than being employed as a proactive stance (at least in YouTube videos, where there is no “Fair Use” button during upload). Utilizing “Fair Use” has often been illustrated as difficult to understand and best to avoid, particularly in “YouTube Copyright School” where the “Happy Tree Friends” cartoon explains to authors that “Fair Use” is unwieldy and should be avoided (YouTube Spotlight). YouTube’s stance against utilizing a fair use defense becomes more problematic as not only does it threaten a permanent ban on the user’s account if they find the user in violation of copyright three times but also it does so through a largely automated process.

According to Patel (2013), a fair use argument can be made for Internet memes, along many of the factors outlined in 17 USC § 107. In factor 3, the amount of an original work used in memes often is only a small piece of the copyrighted work, which reduces the possibility of infringement. Additionally, Patel argues that the first factor weighs in favor as well due to the function of memes as cultural interchange rather than displacing the intended purpose of the infringing work (p. 252). However, automated content identification systems lack ability to distinguish fair use from actual infringement, as the “fair use doctrine” is outside of a machine’s ability to discern. This creates serious concerns over who monetizes the work because although “a technological match might actually be a fair use . . . the copyright owner [can] obtain licensing revenues on works which should not warrant it,” and “authors who uploaded the fair use will be denied opportunities to tap into the advertising revenue generated by their original work” (Sawyer, 2009: 386).

As memes like the “Harlem Shake” can serve to increase the market value of the work (and absolutely did in the case of Baauer’s Harlem Shake song, as evidenced by the Billboard and iTunes rankings), they may be considered “fair” as far as factor 4 is concerned. This, of course, does not apply to all memes, but as Patel (2013) points out, according to market failure theory, “an otherwise infringing use should be permitted where it serves a function that, while socially desirable, could not exist in the market but for fair use” (p. 254). Essentially, even if there did not exist a positive market value influence for the copyright holders, it could be considered fair use due to the first factor, which utilizes the copyrighted material for cultural interchange.

YouTube’s Content ID system might have seemed at the time as a big improvement over the status quo, as it led to a shift from outright blocking to monetization. However, systems like YouTube’s Content ID pose larger issues for digital culture when it comes to considering the space in which this labor exists and who is allowed control and profit from it. The fundamental problem seems to stem from copyright laws that do not really “fit” the way we think about artists and creation, especially in the digital age. The whole idea that Baauer has, or should have, the exclusive right to be a kind of benevolent copyright decider on YouTube seems problematic when the phenomena is made clear. Not only is Baauer’s song a remix of prior samples, work, genre, and culture, but the creative force behind the memetic phenomena has equally, if not more to do with the likes of Filthy Frank, TheSunnyCoastSkate, Redditors, and everyone else who contributed to the collective formation of the meme, not to mention the tens of thousands who created Harlem Shake videos, reproducing and sharing the meme. A meme, due in part to its seemingly accidental, collective creation, should not be owned or “authored” in the same way as a song. Copyright law thinks of texts as either individuals or pieces of individuals that make up a whole, but a meme, as Patel noted, is more than the individual snippets, but they are effectuating “cultural interchange.” The creation and distribution of the meme move beyond pieces of texts, yet copyright law remains steadfast in its assertion of ownership because textual snippets remain tangible within the Content ID matching system. Just like the original Harlem Shake dance, the Harlem Shake meme was a collectively produced cultural phenomenon. In short, the memetic nature of digital culture is driven by the process of production and sharing, rather than a finished, authored product. This process also brings to light the massive networked labor and, in the case of the Harlem Shake meme, the accumulation of hundreds of thousands of hours of work that go into the production and reproduction of popular phenomena.

Part of the problem is that while systems like YouTube’s Content ID offer some form of a compromise compared to the past, it does not address the root problem of digital cultural production: amateur and noncommercial remixes, mashups, fan videos, and so on are, outside of mounting a complicated fair use argument, still considered acts of copyright infringement, still a form of “theft.” Except now, rights holders on platforms like YouTube can decide whether they want to block all creative production, or permit and profit, establishing a form of “benevolent ruler” model.

As the Internet and digital culture transform further into the “walled garden” model of social applications and platforms, harnessing of vast pools of free labor for profit increases without bound. Nicholas Carr (2006) first referred to this technique of monetizing “free labor” as “digital sharecropping” in 2006: “Web 2.0 provides an incredibly efficient mechanism to harvest the economic value of the free labor provided by the very, very many and concentrate it into the hands of the very, very few.”

Clay Shirky (2010) challenges Carr’s concept, arguing that critics like Carr are mistaking “labors of love” for “labor,” suggesting that the critique of digital sharecropping is not entirely appropriate for many of the Web 2.0 platforms, as users who create the value for Yelp, Amazon, Facebook, and so on are spending their leisure time to create and share material based on their interests (pp. 56–60). He argues that the term more adequately applies when a platform allows for the financial exploitation of noncommercial and amateur creative work, which YouTube, especially in this case, seems to be profiting handsomely from.

The coercive nature of YouTube’s “new bargain,” however, redefines Carr’s notion of “digital sharecropping” into something with a bit more punch. As Michael Sawyer (2009) notes, Carr’s notion “lacked a coercive analog to actual sharecropping,” but now, with systems like YouTube’s Content ID system, “the licensing option could coerce users” to either sign over their rights or lose access (p. 386). YouTube’s own frequently asked questions (FAQ) on Content ID makes it clear: “In most cases, you can’t monetize a video that has a Content ID claim. Instead, the copyright owners can choose to monetize your video.”¹²Alexander Galloway (2013) argues that “[i]t is impossible to differentiate cleanly between nonproductive leisure activity existing within the sphere of play and productive activity existing in the sphere of the workplace” (p. 135). In this arena of monetizing activities of indecisive spheres, the distinctions between commercial and noncommercial, professional and amateur that characterize cultural production on the web remain as blurry as ever.

With high-quality video cameras on most smartphones, as well as digital production tools available previously only to professionals now available for free on most personal computers, creative production formerly seen as a labor of love can now easily cross over into a variety of styles of amateur production, blurring distinctions yet again between work and play, as well as amateur and commercial. YouTube’s refusal to allow producers to monetize thousands of videos (and millions of hours of work), instead granting rights to monetize (or ban) solely to a copyright claimant, seems, at the very least, problematic.

Not to say that the copyright holders do not have a stake in the monetization, but as we have illustrated, the Harlem Shake meme has been composed of more than simply a collection of things “found on the innerweb,” and those amateur laborers deserve recognition, as well as some form of protection for the maintenance of creative, open cultural space. Of course, this is not a new story by any stretch of the imagination, as Andrew Ross (2013) notes: “each rollout of online tools has offered ever more ingenious ways of extracting cheaper, discount work from users and participants” (p. 22). Content ID, while at first glance seems a step in the right direction, only assists in extracting additional profit from free labor.

Although it might be difficult in the current copyright climate to mount an argument that UGC utilizing copyrighted content should be able to monetize without sharing revenue with the copyright holder, we believe that due to the compelling argument for a “Fair Use,” there at least exists a productive argument for what we call “fairer use” for UGC where revenue is divided, donated, or minimized in a more amicable manner, recognizing the importance of the creative labor of content creators. The current system seems to turn its back on the importance of UGC, leaving us without a dance hall of our own to “effectuate cultural interchange,” which fails all of us.

This, of course, remains difficult within the current climate due to the “coercive” arrangements currently in place and the inability of technology that cannot accurately make decisions about content.

Although Baauer and his label Mad Decent were praised for allowing the widespread use of Harlem Shake in the creation of derivative works, as if they donated the song for Internet philanthropy, those creations were directly responsible for an incredible generation of revenue for Baauer and Mad Decent. Not only did the meme bring global exposure to the song Harlem Shake, eliciting hundreds of millions of streams through YouTube and audio streaming sites, but also prompted sales of over a million digital downloads from iTunes (Grein, 2013). In addition, the timing was perfect for Baauer, as Billboard had just revamped their metrics for the Billboard Hot 100 chart to include YouTube views. Due to the explosive popularity of the meme, Baauer’s Harlem Shake rocketed to the No. 1 position in the Billboard Hot 100 chart and stayed there for 5 weeks (Trust, 2013). The value of being able to derive both global fame and revenue from the Harlem Shake meme was so great that Diplo, head of Mad Decent, stated that the phenomena “saved the label,” which he claimed had been headed for financial ruin (Cubarrubia, 2013).

YouTube’s Content ID helped Baauer and Mad Decent even further. Although it is unclear exactly how much money Mad Decent and Baauer generated from Harlem Shake videos on YouTube, using the information available, we can reasonably estimate the revenue generated. Taking into account only the Harlem Shake meme videos on YouTube with 1 million or more views, there are roughly 250 videos that Mad Decent claimed through YouTube’s Content ID system; adding up all of the views from those 250 Harlem Shake meme videos with over 1 million, we get over 1,382,000,000 total views.¹³ It is difficult to estimate how much revenue was generated per video, as we don’t know the exact CPM (cost per 1,000 views) rate which advertisers were charged, given that CPM often varies given the country of origin and video (users watching in wealthier countries typically are priced at a higher CPM). Generally speaking, many YouTube videos receive a CPM of around US$2. At US$2, these 250 Harlem Shake meme videos would generate roughly $2,764,000 total in ad revenue. Given that YouTube takes about 45% of revenue, this would leave around US$1,520,200 for Mad Decent. According to MSN Money, however, Harlem Shake videos received a CPM of US$6, which would generate roughly US$8,292,000, of which US$4,560,600 would go to Mad Decent. Given that thousands of Harlem Shake meme videos were flagged for revenue, far more than just these 250, it is incredibly difficult to know exactly how much revenue has been generated. In addition, in early 2012 (prior to the meme), Mad Decent had inked a deal with INDmusic, which functions as a kind of “Vevo for indies” on YouTube, to monetize their videos through pre-roll advertisements (Hampp, 2013). According to Billboard, more than 4,000 Harlem Shake videos were tagged for revenue through YouTube’s Content ID system and the INDmusic deal. We can say that Mad Decent’s ability to harness the meme through their copyright claim and privileges netted them, in direct revenue from YouTube alone, many millions of US dollars.

From the perspective of Mad Decent and YouTube, the Harlem Shake meme was a potent and harness-able force. Content ID allowed Mad Decent to harness millions of hours of creative free labor. The millions of hours of creative labor that went into producing the tens of thousands of Harlem Shake meme videos were a creative undertaking the scale of which not even the largest record companies could muster. While no one person created the Harlem Shake meme, the ability of Mad Decent to control and profit from the phenomena through Content ID gave them a kind of de facto ownership of the collective production, as they not only remain the majority profiteer but also control whether or not to block the video.

Lessig (2008) notes of YouTube that “No site—ever—has more quickly become central to popular culture” (p. 195). From the very beginning, YouTube provided a platform for people to post and share culture, like dance videos, styles, and techniques. Today the dance crazes of the world spread virally through the social media networks of the Internet, in addition to late night dance halls. In the digital age, people have gravitated toward participating in both the celebration and transformation of cultural phenomena online, and, as always, people want to dance.

YouTube is not a commons in the sense of a free space, service, or resource that is publicly owned or controlled, but it nonetheless functions in a similar fashion. Its reputation as a free, accessible, inclusive, and open platform drives its immense social and economic value. As Ross (2013) notes, “The underlying spirit of mutuality. . . has been surprisingly tenacious in the face of concerted efforts on the part of would-be monopolists to enclose, privatize, and commercialize the digital domain” (p. 30). Ad revenue sharing seems to make sense for many YouTube professionals who desire to monetize their videos. However, the majority of YouTube video creators’ social, noncommercial, free labor becomes increasingly subject to structures of enclosure and monetization by intellectual property holders through a system like Content ID. As numerous aspects of everyday living occur increasingly within and through social media platforms, this process poses a real threat to the cultural and social value of YouTube as a functional digital commons. Tiziana Terranova (2012) historicizes this process, arguing for “free labor” as a feature of the cultural economy has become an “important, yet unacknowledged source of value in advanced capitalist societies” (p. 46). Moreover, Terranova describes the digital economy as increasingly driven by mechanisms of internal business practices and structures to capture these growing pools of social knowledge and free cultural labor generated online.

This attempt to enclose and monetize poses a direct challenge to what Michel Bauwens (2012) calls the “peer-to-peer ideology of the cognitive working class” (p. 261). Terranova argues that “As wealth generated by free labor is social, so should be the mode of its return. This means investing this wealth in the reproduction of the common” (p. 71). The logic of Content ID naturalizes a system in which users are forced to bend to not only the structure of YouTube but additionally to the will of corporate rights holders, allowing their creative contributions to be controlled because of the inclusion, however small or inadvertent, of a copyrighted work. While some revenue is “reinvested” back into YouTube, Content ID automatically captures and diverts profits into the coffers of copyright holders. Given the monopolistic dominance of YouTube over video sharing and its owner Google over search indexing, mobile data, and other digital services, users who find this unfair or exploitative have few practical alternatives.

Transforming the platforms of the Internet into more closed systems of profit extraction threatens one of the basic bargains of culture: people celebrating and participating in culture without having their energy and “labors of love” exploited for profit. This basic bargain was, it should be emphasized, how YouTube rapidly rose to popularity and market dominance. YouTube’s Content ID system, or more importantly, the rationale behind the system, sets up a pattern for increasingly aggressive for-profit harnessing of noncommercial and amateur cultural creation, establishing a platform of aggressive digital sharecropping that changes the “basic bargain” of social and/or noncommercial culture exchange. This is a troubling new normal for the amateurs of YouTube: your content may be sharecropped or blocked.

Looking at a specific meme like the Harlem Shake remains useful to identify the discrete occasion or phenomenon of collaborative, authorless creation and spread of culture, as well as its relationship to new copyright structures like Content ID. However, focusing on the memetic nature of popular phenomena simultaneously masks the far larger collection of creative work and energy that is uploaded to YouTube every minute of every day and which is also governed and monetized through Content ID. Nevertheless, a clearer understanding of memetic culture, authorship, and amateur labor can help to reframe discussions about the Internet and social media within the tradition of articulating and protecting the rights of commons and non-market spaces and culture.

YouTube’s Content ID gives front-end control to copyright holders, leaving them in the position of cultural dictators (benevolent, if we are lucky). While Content ID does not seem to immediately hinder “amateur” content production, the underlying problem of giving exclusive control to authors and rights holders over popular cultural expression and content that they did not create (and whose exposure they already benefit from) further leverages the genuine production of social capital into a system of commercial exploitation. Digital sharecropping and the leveraging of cultural production for profit appear as an endless source of economic growth potential for this burgeoning industry. While the individual instances of sharecropped labor may be negligible, as the normalization of coercive digital sharecropping as the Web’s changing cultural bargain becomes widespread, those individually negligible instances amount to an incredible assemblage of unacknowledged labor which inscribes the core environment of the Internet as one that may be ignoring both current and historical labor systems.

The authors would like to thank the anonymous reviewers as well as Tarleton Gillespie and Hector Postigo for their comments on earlier drafts of this paper.