At the end of last year, 8tracks CEO David Porter announced that the popular music-streaming service was shutting down. Citing heavy competition and a lack of funding, the company would be folding after, as he calculated in a blog post, 137 months. The surprising announcement struck a chord with people who remember 8tracks’ launch in 2008, back when the MP3 reigned, and streaming media and recommendation algorithms barely existed. The service was largely powered by its users, DJs, who curated playlists based on taste and mood.
Over the last 12 years, a lot has changed in the online-music industry, and those shifts eventually left 8tracks without many options except to call it a day. 8tracks was, for a time profitable, even as music royalties threatened to eat up much of its revenue. Yet finding funding became almost impossible from the get-go as investors sought unicorn-level valuations instead of reliable businesses with lower returns. Porter spoke with Intelligencer last weeks about what went right and what went wrong.
How was 8tracks conceived?
The idea goes way back. The original impetus or idea came from my time in London in the ’90s. I worked for an accounting firm called Arthur Andersen, back when that existed. The thing that I thought was so interesting about the music scene there was, for electronic music, it was really focused around the DJ rather than individual artists. DJs who weren’t even producing music. DJs who played at clubs, who put out compilation CDs. The emphasis was on the curator rather than the creator.
This paradigm of the DJ as focal point made even more sense in this world where there was information overload. You really needed a great filter to help separate signal from noise. With that concept in mind, I wrote a business plan for a class in the fall of ’99 and the idea that I put into this plan was a platform where people could maintain a profile and the main type of content that people could share was playlists. I didn’t have my head fully around the copyright and royalty requirements, but I knew a service that allowed people that were passionate about music to share with others would be super compelling.
There was a specific feature of Napster that painted a vision of that. If you were downloading music that was not so mainstream and a particular user’s name kept on cropping up, you could click on that user and then click a button labeled “Hot List” and it would show you all of the MP3s on their hard drive. It was this unorganized form of social music discovery at a grassroots level, and I thought that was really cool. I thought maybe there was a way to add some organization to that platform.
[After some work at other music start-ups,] I put together the business plan for 8tracks. We couldn’t get funding. I pitched Fred Wilson at Union Square Ventures in 2006, and he let me know pretty quickly and in a very nice way that at this stage, VCs don’t fund business plans, they fund businesses. The world had changed and you needed to bootstrap your way out of the gate and then once you had some traction then you could begin discussing funding with VCs. So we bootstrapped and launched in the summer of 2008.
This was a couple of years before Instagram rolled out, and it was a cool way to create a work where the whole was greater than the sum of the parts. You had this cover art and a description, and it was almost emulating the mixtape concept. There was also this freeform tagging structure that was useful for music. You could label your playlist by a genre or by an artist, but we found that users started tagging in the human dimension, like with an activity or a mood. We were sort of the pioneers in programming in that way.
So people could listen to those playlists on 8tracks directly? Were you hosting the music yourself?
At that point it was an iTunes-dominated world. People would upload MP3 or AAC files and fashion a playlist. It had to have at least eight songs, and if you do the math at about four minutes a song, the idea was that would provide a natural breakpoint where we could put an ad or something. And we would stream together a set of similar playlists, but each of those playlists was curated by a person. People at the micro level, and algorithms at the macro level.
[User uploads] was at that point the only way we could function legally. We operated under the Compulsory License for Webcasting, which is a set of rules that makes sure it’s similar to a radio experience and thus is promotional of music sale and not substitutional. You were limited to how frequently you play tracks from a particular artist, there was a limitation on how many skips you could have, you couldn’t rewind, you couldn’t see the track list in advance.
What body governs the Compulsory License?
The Compulsory License was established by the Digital Millennium Copyright Act of 1998. It allows an entity to stream in the U.S. and the rates for doing that weren’t set initially. They put in place a process that would set the rates, where a copyright royalty board would review existing deals and testimony from the record labels and the webcasters and come to a decision about what the rates ought to be. Those rates get reset every five years. That was the regime under which we operated for the sound recording royalties. There are also music composition royalties that are paid to ASCAP, BMI and SESAC. Those royalties go to the songwriter.
At the time you launched, those royalties were not excessive?
At the time we launched, those royalties were definitely cheaper. What was unique about our situation was that just the year prior, in 2007, the rates came down. There had been a lot of pushback from Pandora and its listeners that Congress strongly encouraged an alternative set of rates. One of the new categories was the Small Webcaster license. The rates are generally a per-play rate so you pay a certain amount for every track streamed irrespective of how much revenue you’re generating. The rates under the Small Webcaster arrangement were a percentage of revenue or a percentage of expenses, subject to a minimum. What was nice about those rates is that they grew as you grew. In the early days when you’re not of sufficient scale to be generating premium CPMs and are relying on programmatic advertising, you could still pay the bills. That was great for us because it allowed us to scale up in a sensible way. 2013 was the last year we operated under the Small Webcaster license. Had that not existed we would’ve run out of cash a lot sooner.
How did people find 8tracks?
At the outset, it was just a lot of my friends that lived in downtown New York and Williamsburg. We did get some good early press. There was a piece comparing us to Muxtape, which was another kind of hipster Williamsburg favorite of that era. They had just gotten shut down by the labels, so we were the better legal alternative or something. And then we just steadily grew. The nice thing about the crowd-curated model is that, in addition to providing this deep programming model that emphasizes music discovery, every time a DJ creates a playlist they share it as widely as possible. So there is this inherent social quality. The people making your programming are also your marketers. A few months after launch we were getting a lot of love on StumbleUpon. There were two mixes in particular, one was a mixtape of mashups between classic hip-hop tunes and the soundtrack from Zelda called The Ocarina of Rhyme. The other was our guy Tyler, now the CTO at Bustle, made a playlist called “Songs to make you feel better.” Happy music for recessionary times, I guess, since it was early 2009. Those two mixes blew up. We jumped from 30,000 unique visitors in a month to 300,000 and that’s when we knew we were on to something.
So what changed? It seems like music licensing was a real sticking point for you?
Things changed when we went for a small webcaster to a large webcaster, because we had hit thresholds for revenue and monthly streaming hours by the end of 2013. We were being charged on a per-play basis. That made things much more expensive.
We did have this halcyon period where everything was up and to the right. The big uptick in traffic in 2009 begat even more traffic. We did raise a round from Andreessen Horowitz and others and that allowed us to hire more people. We thought we were on the right trajectory. By 2012, 2013 we were the No. 3 internet radio service in America, and we were gaining on iHeartRadio in the younger demographics.
Probably even more impactful than the royalty arrangement, in December 2013, Spotify rolled out their free mobile offering. That basically gave people a pretty good user experience for on-demand listening in their pocket. If you plot our monthly active users and our total streaming hours — which is probably our best measure of activity — those both started to fall off in March or April of 2014. At that point we were 8 million MAUs, and 32 million monthly streaming hours, and that began to steadily decay. I forget the timing on this, but their lean-back offering [streaming service slang for algorithmic recommendations] got way better after their acquisition of The Echo Nest, which was arguably best-of-breed music-recommendation technologies. For a lot of listeners, it was good enough. I don’t really think we’ve lost any meaningful listenership to any other service, I think it’s almost exclusively been Spotify.
The third thing was the inaccessibility of more funding. When we went to try and raise the Series A round, investors passed. One potential investor said our trajectory was fantastic, and what that signaled to them was that we would likely be acquired by a bigger player, and that would cap the investors return. That did make it more challenging to raise from other venture guys. We ended up raising a Seed 2 round, not enough to constitute an A Round but figure out our next step. At the end of 2014, Google tried to buy us; we turned that down because it seemed like we had an A Round lead on an investor, and the offer was more of an acqui-hire.
In 2015, we decided that we should try and sell, but ultimately there were no takers. We then decided we would try Regulation A crowdfunding, which is different from Kickstarter, because it’s equity-based. It’s also different from Regulation C crowdfunding in that you don’t have to be an accredited investor. It made a lot of sense because so much of our product is created by our community. There were a lot of hiccups along the way. We ended up setting a cap of $11 million, I thought we’d raise maybe five or six million. But because of a lot of issues, including the inability to take payment via credit cards, we ended up raising only $2 million. We couldn’t really get ourselves back to the growth that we needed to.
Do you regret not selling to Google?
I do now. It wouldn’t have been an amazing outcome but personally it would’ve been good. Maybe I could put a down payment on somewhere in San Francisco, but it’s a mixed bag. Under those circumstances at that point in time, given what I knew, I still think it was the right decision. Everything was up and to the right, it looked like we were getting new funding in, there was no indication whatsoever that we’d see a dropoff. I felt like we were sufficiently differentiated in the market.
As someone who is not well-versed in finance and works in the ailing media industry, I’m always sort of baffled every time I hear about VC funding. Any business that operates in the black sounds good to me. Was it frustrating to hear that your profitability and growth wasn’t enough?
It’s certainly frustrating. I guess I’m looking at it from my lens now a number of years on but it was frustrating at the time. Because we hadn’t really taken much money yet. We’d been living on a shoestring — that initial funding was only $1.2 million — and yet, with almost no cash, we were able to goose growth to get to five times and profitability. We were clearly superserving a segment of the market that no one else could touch. It seemed to be that we could do a heck of a lot more if we had more funding. It’s hard to know what kind of ROI a VC could see, but Spotify had raised a lot by that point in Europe, and the big boys had endless amounts of money to put into a loss-leading music service.
The music industry is challenging. There’s fierce competition on the demand side [streaming services] and the supply side [record labels] are an oligopoly, strictly definitionally. I understand their investment thesis; they’re looking for home runs. That’s just the structure of that model. They’re looking for highly risky bets that really pay off and subsidize everything else. I do think there is a middle ground that’s being under-served today. The question that we always have is “What’s wrong with a $100-million company?” That seems like a reasonable thing to back but it does fall outside the venture model.
Given what you said about record labels being an oligopoly, an interesting thing about streaming start-ups is that the costs per user don’t really fall as a user base increases. The royalty rates always stay the same, right?
There was some talk recently from Jimmy Iovine about the lack of scalability because the royalty rate is fixed. That’s not new news, we’ve all known that. The real point here is it makes it hard for the business to work unless you have sufficient scale to sell a ton of ads with a direct-sales team and achieve premium CPMs. Spotify has 100 million users and a subscription plan and yet it still struggles to reach profitability.
No matter how big a music start-up gets, it’s not a curved graph. The expenses rise in a straight line.
I think that’s right. The way I would think about that is: what we’ve seen borne out in the marketplace is that in order to have a profitable model of streaming music, you have to be the market leader with massive scale. I think Spotify is close to profitability, but they are by far the market leader. There are markets where you can have competition, but given the economics here, it’s just not that possible. It’s unclear if the other big tech companies would be profitable on a standalone basis with their streaming services, or if they serve primarily as loss leaders.
Just recently, Spotify announced that they would start targeting ads based on the podcasts people listen to. The general think was that the reason Spotify has leaned so heavily into podcasts, and bought Gimlet, is because they don’t have to pay record-label royalty rates on podcasts.
A hundred percent. You can strike deals because the suppliers of the podcasts are not in a situation where they can call the shots to the same degree. One last point I would make on the music side is that I saw something recently that the most-streamed songs this past year had considerably fewer streams than the previous year. There was a similar study that came out about a year ago segmenting the the market — the top 50 streams of the year, top 500 streams, etc. — and comparing that year on year. What was fascinating is that the top two segments, as a percentage of the total, were going down. What this suggests is that there is a greater push down the long tail. We’re finally seeing that shift in music thanks to on-demand streaming and a subscription-based model. We’re still very much in an oligopoly and a star system today, I think there is going to be a steady push down the independent sector and DIY artists.
That sounds good to me.
It’s a democratizing trend. I think it’s also good for listeners because there’s more diversity.
Do you see any differences in user behavior between people who use streaming services and who use something more deliberate like 8tracks?
You’re asking, if you look at a music fan circa 2008 and a music fan circa 2020, how do they differ?
Yeah, what got me thinking about it was how 8tracks started as a service you had to upload files to. Very few people, only psychos like me, download music anymore. Is there a behavioral difference between people who have local files and those who rely on streaming?
I do think there’s a difference. The DJ community on 8tracks is only like one percent of the user base. So the people who are still uploading, those are still the collectors. I still think people amass libraries on Spotify, but I do think the people still making playlists fall into this collector bucket where they want to own their music. They don’t want to be dependent on a service that could go out of business or suddenly raise prices or lose music rights. They want to own that music in the truest sense. That aficionado market is still real.
Given that, what’s the response to your announcement been like?
It’s been lovely. I was blown away by how many people have sent us tweets and emails. It’s an outpouring of love and fond recollection of 8tracks. Even to this day, people are like “There’s never been anything like it.” There’s other services but there’s something special that still hasn’t been captured or emulated in the same way. There’s a feeling of loss among many. We’ve accepted this downward path for a long time, but there has been some new interest since the blog post, and so we’re entertaining some discussions about a possible buyer. I’m not sure any of those will come to fruition. Upshot is I’ve been pleased that so many people found so much value in the service. The quintessential is people who listened to 8tracks during college from 2008 to 2012 or 2013 and that was like the soundtrack to their life, it helped them at certain times. It was a special tool, not to sound cheesy.
What would you say is the special thing that 8tracks had?
To my mind, there’s probably three things. One is the depth of music discovery. There’s an element of serendipity and surprise when it’s a person instead of an algorithm. Maybe a song doesn’t fit the genre but there’s a contextual link. Thing two is this programming that gets really granular. A listener can pick a cue like “yoga” or “rainy day” and an artist and get this really thinly sliced taste and context that’s really unique. There are a lot of mood-based playlists on Spotify and others, but if you go to 8tracks, you can crosscut those with a taste cue like genre or artist. The third thing is a little bit more elusive, but it’s this sense that there’s a soul behind the music. There’s a narrative and a story communicated through the mix art, through the title and description. And you can interact with that person directly if you want.
What happens to the 8tracks archive?
We’re gonna keep the site up. We’ll eventually run out of cash if there’s not a buyer. We don’t stream from our servers anymore, but we’ve left the site up for now. We believe we might be able to get a buyer more if we leave the site up. We also want to give as much time as possible to export their playlists.
What’s next for you?
TBD! If we do find a buyer for 8tracks I want to help with the transition. I’m kind of thinking about whether there’s an opportunity that makes sense at a larger streaming company. The thing I’m excited about is a productivity and mindfulness app that I’ve been working on called NextUp. The goal is to help people plan and execute their day in a thoughtful fashion. I think of it as another tool toward being happy in the same way one might do yoga or meditate.