Jacques Mattheij

Technology, Coding and Business

Dalton Caldwell

Startup School Transcript October 16 2010 Dalton Caldwell - Founder, Picplz; Founder, Imeem Link to video: http://www.justin.tv/startupschool/b/272178844

Jessica Livingston: The next speaker is Dalton Caldwell. He was one of the founders of Imeem and now he is the founder of a new company that is working on a new iPhone [edit: and now, Android] app called Picplz. Download it now, it’s a photo sharing app for your phone.

We all think it has to have someone come to Startup School and talk about a company that failed. So, Dalton is here. We’ve asked him to talk about Imeem, because he came to Y-Combinator and he talked about it.

Dalton Caldwell: Hello!

Jessica Livingston: He did such a great job, it was one of the best talks we had at Y Combinator, so we’ve asked him to touch on some things about Imeem.

Dalton Caldwell: Alright! It’s a pleasure to be here. I know probably half of you already saw me on the list and were like “Who is this guy? Didn’t his company just blow up about nine months ago?” Yes, that is me, and Imeem was the company I started, and yes, it did blew up a few months ago.

Actually, Paul specifically asked me to come talk today because there are still so many people that are doing music startups and doing things that require licensing. I’m going to talk about the real stuff, I’m not going to lie, Eric will appreciate that. I’m going to tell the truth about music startups and, you know, I have no strong emotional thing about it, I’m just going to talk about the economics of it. This will be a little bit more of a detailed talk than a lot of the other ones today, I hope you guys are ready.

I graduated from Stanford, happy to be on my alma mater here. I’ve actually never been in this particular auditorium before. I did symbolic systems, did a lot of programming, and I also got a B in psychology, which helps me a lot. Imeem was a six year startup. I started when I was 23, right out of Stanford. Six years is a lot of time in startup years. You know, it’s like dog years. I raised $50 million, I was a Sequoia CEO for most of that time, I acquired three companies, including Snocap, which Ron was talking about quite a bit, so in a weird way, I acquired Ron’s company to be started.

That seems very strange to me. So I acquired Snocap, started by Shawn Fanning and Ron Conway, and still my company blew up. I got my revenue to $24 million monthly runrate, that’s about $2 million a month or so. That is something I’m very proud of, when I look back at the entire experience, actually making something monetize at that scale is very important. Imeem got very big, we got to Alexa #75, 25 million uniques, our widget hit 108 million a month uniques, which is a very high percentage of the Internet. So, that was a lot of fun, it was a great trip.

I was a little afraid about doing this talk because I was afraid it would be like “This guy’s got sour grapes, he’s going to attack people in the music industry.” Let me just start by saying this: this “Us” vs. “Them” thing is biased. I think a lot of people behave a certain way because of the situations they’re in, and actually their personalities are far less important. There’s been a psychological research about this. So, you tend to project on “this guy is a jerk, this guy is mean, or this guy is greedy”, when in fact people are just optimizing the situation they’re in. This is something to really remember vis-a-vis the music industry. Knowing the music industry very well and knowing the heads of all the labels and all that, I can genuinely tell you there’s just as many smart people, and nice people, and good people, and bad people as there are in Silicon Valley, and I mean that. So, let that go on the record.

Alright, let’s talk about music startups. So, you want to do a music startup. Why does everyone want to do a music startup? Well, everyone is an expert, because they’ll have the best taste in music and know about all the great bands. Everyone loves music, everyone thinks that they have a messiah complex, so they’re going to fix the music industry singlehandedly, because these dumb music labels just don’t get it. Everyone realizes that this social music is this aha idea, and I should add music on Facebook, that will change the world. Frankly, it’s easy to get tons of traffic. If you listened to a lot of the talks today, the hardest thing when you’re very illustrated to just? getting traffic, if you put up free mp3’s on a web server somewhere, you get a lot of traffic and you’ll feel really smart. Wow, users love my product! But the fact is there is a downside to that, but if you want to get a lot of traffic really fast, just host a bunch of unlicensed mp3’s. Frankly, the users don’t care, you could build a terrible product and if you’re giving away free music, people still find a way to continue to use it. I was a big fan of the original Napster, that Shawn wrote, and I don’t know if you guys used it, but it was not great, it crashed a lot. I was a biggie Napster fan, I had a T-shirt and everything. So, you know, that’s why you do them.

Here is a cue for you guys. Who knows what this is? Yeah, that’s a much higher percentage.
This was designed by one of my former employees, who worked for me for a very brief time before he went to work for some Harvard kids. I think he did a great job on it, you can see the little W and the H in the nose, there’s a lot of subtleties, it’s a very playful logo. Had those Harvard kids actually decided to stick with this, that summer they moved to Silicon Valley, and when they were trying to sell off their other site for like $1 million, I believe is the story, they were trying to pound it off for $1 million, so they could really focus on the music startups, because that music startup was the big one.

Let me show you another thing that the same designer did more recently. It is also kind of playful, there’s a hidden meaning there somewhere. That’s what six years looks like guys. You want to know what six years looks like? That’s it. I mean this sincerely, I’m glad they didn’t do that music startup. I think that speaks for itself.

What I’m trying to talk about is opportunity costs. You know, Chris Dixon talks about this stuff a lot, I know you guys probably read him. Yes, there is an opportunity cost when you decide to do a music startup, and it’s important to think about it.

Honestly, I feel kind of stupid spending six years building something and making it to that scale, and then standing here, being the sacrificial comedy relief for this particular conference. I think if I would have spent that time building something else, I think it probably worked out better for me. So, there’s such a thing as opportunity costs. Let’s talk about it. I’m going to go through these, we’ll see how the timing goes.

So, you want to build the tools for artist startup. Radiohead, ((Director Fan)), we don’t need labels, these ((very common branded startups)), so artists are not going to pay, they don’t have any money, there’s a lot of saturation there. TuneCore is really good, research them, they have a disruptive business model. They just charge a flat fee to put your music in the music store, they don’t take a percentage of sales, so you might want to look into that one. Topspin is really good, the CEO manages the Beastie Boys, and already has a huge [buying? buy up?] doing a direct consumer-artist proposition. Let’s talk about margins, no one has talked about it today. Doing a store like this, at best you’re in 20% margins. You have to sell $250K of indie band content every month, so your $50K are going to start up on. I don’t know if anyone if familiar with numbers here, but selling a quarter million dollars of unknown bands in mp3’s is the hardest thing that I can possibly imagine. So, watch out, this is the one startup that everyone does, don’t do it.

In general, let’s say, that’s just indie artists, let’s do download stores. If you do deals with labels, you have to pay advances, and that’s a lot of money upfront, I would say $3-4 million dollars you have to have prepared just to get these licenses. iTunes already has +90% market share and it’s really hard to get people to switch from buying music from iTunes - I don’t recommend that either. Amazon is selling mp3’s, Google is about to launch a whole big music store. So, you have to pay $5 million in, you’re going to have a 20% margins, so you have to sell $20 million of Justin Bieber downloads one at a time to ever make your paid in capital back. This is a really hard business, I don’t necessarily recommend it.

Ad-support, this is what I did. I was on the forefront of doing ad-supported licensing and pushing it through the music industry because, I thought, if you did free music that was ad-supported, you can get a lot of the traffic that exists in peer-to-peer land and get it back into a way that’s actually monetized and that compensates artists to do so. That was my grand utopian vision for Imeem. Everyone focused when they were dogging on Imeem about per-play minimums, and it’s true, per-play minimums are very challenging, every time a song is played, regardless if you show an ad or not, you have to pay the labels, so you can get way upside down really fast. But the other thing that no one writes about is that these quarterly minimums and advances are deadly. Quarterly minimums mean, no matter what happens, even if your site goes down, or you name it, you’ve got to pay the labels $X million every quarter. No matter what happens, you have to pay them this money, it’s burned, there’s nothing you can do about it. You can’t pivot when you’re attached to something like this. You can’t pivot out of deals like this. You’re boned. So, doing deals when you’re contractually obligated to pay out millions of dollars in a very consistent time line, is something that I just do not recommend. In general, labels won’t even do ad-supported stuff anymore. Just to cut to the chase, the reason is now that we have free music on mobile devices, it completely takes out the reason to ever own music again and this whole music cloud thing just messes up their economics even more, so I’m not even sure that they’re going to license ad-support anymore. Maybe this won’t be a problem for you.

Subscription startups are the thing that’s getting a lot of people started, man, this is a hard one. Here’s how you do a subscription music startup. Spend two years in stealth doing deals, just raise $5-6 million. Not too hard, right? You’ve got to give the labels 10-30% of your company. Let’s just talk about the actual business model: say you charge $10 a month, you’re going to face 10% monthly churn of people unsubscribing from your service, it’s kind of like the AOL effect. You’re left with about $3-5 of that $10 - and I’m being generous - to actually run everything on headcount, bandwidth, you name it, customer acquisition. Plus, you have the quarterly minimum thing that I was just talking about earlier, that happens regardless of the subscriber numbers. And there’s this whole international licensing thing, so every country has a different body that you have to license with, so to actually do more than one country for your music startup is insanely difficult. And can you imagine what any kind of social site would be like if it was US only? It would be very difficult to ever get that thing going, without getting into those network effects. If you’re doing a legal startup, you’re going to be facing this. Also, there is a matter of Napster, Rhapsody, Spotify, Rdio, Thumbplay, MOG, and the $300 million of free ads that Rhapsody has on MTV and VH1. So, yeah, a two persons startup, go for it, sounds good!

Let’s talk about running an unlicensed or sorta-licensed site. Oh, we’re DMCA compliant, every time we’ve got a take down, we take it down, right. It’s legal. I mean, I do not recommend this one either. If someone says or declares jihad against you in the press, it’s not fun. I’m not trying to mock this, I’ll be talking about this in a second. I have been sued myself, but it’s not fun, you’re not going to want to come into the office, you’re going to wonder why you’re doing it. I’ve talked to a lot of startup founders that have asked me for advice that go to the decision to shut down their company rather that they continue to face additional litigation. It’s just not something you want to go through. And it’s not badass, maybe it seems in an abstract way. Man, I feel like a badass, I’m getting sued, I’ll be taking on the music industry. But, when I have a knock on my apartment door at 6AM when a process server came to serve me litigation, that’s not fun. You don’t want to go through that. It’s not badass, it’s not cool. Finally, ((DMCA radio)), which is what Pandora is doing, actually seems pretty good, I like Pandora, they’ve done a nice job. They’ve had a lobby congress to keep it going and they did it successfully, and they do about $100 million in annual revenue, and still are almost profitable. So, yeah, I don’t know if I recommend that to very small startups. It seems incredibly difficult. And, yes, I know I’m missing a bunch of other music startup angles, I just tried to hit a few of the highlights in this talk.

I want to talk about why this is so hard. To me, it’s not about the people, it’s about the structures. I think the real problem here is all about the legality, and I think there is a lot of similarity to software patents. If you’re familiar with software patent stuff, patent trolls go out and say I invented… If a user clicks on something on your site, you’ve got to pay us $100 million in licensing fees. So, there’s always people doing patent trolls. Very similar to this.

There’s no actual defined rules to run a music startup in the international sense, so there are always gray areas where lawyers say you could do it or you cannot do it. That’s your call, if you feel lucky today. So, that applies to mashups, there’s publishing which, you know, just Google music publishing I could spend three hours talking about. But that’s in addition to paying the labels, you also have to pay publishing. And iTunes gets sued all the time for copyright infringement. Everyone does, if you mess with music, you’re going to get sued, it’s just a fact of life. It’s just like patent trolls, we got sued at Imeem by patent trolls as well, just like everyone else. If you have any kind of streaming content, patent trolls will come in and sue you if you have enough money.

The other thing is people think they get successful, oh, I got all this traffic, I raised all this money, that’s when you get sued. I’ve talked to a lot of startup founders who were like, you know, no one has contacted us yet, the labels seem to be OK with what we’re doing, we never really got an e-mail and I think they’re excited about our music discovery platform. Look, they’re not going to come and talk to you until you have money, until you’re successful. That’s the way it is, they have this whole queue of startups that they track, and they take screenshots - I’m serious - you just never want to get to the top of that queue, because when you do, you’re first in line and then they focus on you.

Another thing to really talk about is M&A, it’s basically impossible with music startups. If you don’t have deals, no one is going to buy you. They’re not going to buy some huge legal liability, because that means they’re buying themselves, that big company is now theoretically on the hook for a lot of money. If you have deals, no one is going to buy you because the deals are actually not transferable. The way every label that I’ve ever seen does, is the labels have a right to first of refusal to pull the deal if they don’t like the acquisition. So, everybody is talking and worried about board members, what if I don’t want to sell? Imagine that, in addition to your board, you had to get approval from all the four major labels to get your company acquired, and they’re going to look at the economics of the deal and decide if they’re being fairly compensated for that said deal. That sounds like fun, right.

Either way, it’s going to be tough, and the way to think about it is, this applies to more than just music, to any kind of large content company. If you’re dependent on somebody else’s content to survive, you’re effectively not in control of the asset, it’s not actually your company. You’re a proxy through the company, but if the guys over here that control the content don’t like what you’re doing, or want to take you down, or destroy you, they can. That’s really not a place I would recommend to be in.

Let’s talk about why the labels are so irrational. This is something I did not understand whatsoever until I spent a lot of time with them and I totally see where they’re coming from. If you went from an industry with a $50 billion annual gross revenue and cut it in half and just watched it every year get smaller and smaller, your job was to fix it. We’re going to pay you a lot of money, you’re an SVP or an EVP, go be innovative, go figure this out, you can fix this, right? They’re facing a very challenging thing, because they’re trying to solve an equation, and the equation means bringing in several billion dollars every year. Frankly, they’re a little emotional about it because they’re worried about their jobs, all their friends just got laid off, the labels have lay offs every few months, it’s terrible, I feel for these people that have been working for the music industry their whole lives. These are real people with families, and their entire industry is being burned down.

So, they’re trying to fix this, they’re emotional about it, and then people come in, whether a ramen-profitable startup that might be able to generate a few thousand dollars next year and are like “We’re just going to get big first on users and then maybe monetize later, we have Google AdWords in the mean time.” If you come in there and say something like that to them, I think they justifiably, you know, it’s not even worth their time or what their salaries cost, frankly, to even have that conversation. So, they’re trying to solve for these big wins, they’re not trying to license a bunch of little people, they need to find the touchdown pass, the buzzer beater, so the whole thing doesn’t get burned down. Again, I relate to that, that is what the world looks like to them. It is hard, I do not want that job, but, you know, this is a lot of the reasons that people tend to demonize the music industry so much, they’re dealing with economic realities and, you know, there’s not a lot of options.

Frankly, there’s also a bunch of structural cash incentives here. So far, over the past few years music startups have effectively been money transfer schemes where a lot of money goes into VC funds, they take their fees, the money goes to music startups and to labels. This has happened quite a bit, it has also happened with video content sites, you name it. It’s been really good for the bottom line, the margins are really good on these kind of big cash things, you don’t have to pay the artist, or anything, when labels take those advances, and the people that worked with labels are compensated for short term cash. Very much like Wall Street. I’m really interested in what caused the bubble to burst in real estate and why all this irrationally happened. The reality, in my opinion, is there was a lot of people optimizing for their own short term benefit and doing what made complete sense that would give them big bonuses at the end of the quarter and at the end of the year. Everyone was like “I’ve got to make more money.” And then, the entire system that was building this money just fell off of the cliff because there was absolutely no incentive to be thinking about the long term strategy. This structural reason I would suggest is why, how we ended up here, where the labels are trying to find the billion dollar deals and throw the touchdown passes, instead of licensing a bunch of three persons startup, which at this point in time I agree it doesn’t even really make sense.

Whenever I talk to people of this who are like “Well, Dalton, what do you think, what’s the answer?”, I’m actually going to try to give that one today, because I figure someone will ask me about that. I think you need broadly available, legal music APIs that you could use as a developer to actually innovate off of. I really tried to do this at Imeem, I got this license, we had a music API. It was not successful because Imeem’s API is not on anymore and so people can’t really build their company on something that doesn’t really exist. But I really saw this as the most important transformational thing that I could bring to the table in my company. I think that Congress has to get involved, as much as they did with Pandora and DMCA radio, you need to have statutory licensing so that there’s a very clear way if you want to do a music startup, you have to check the file in four boxes, here’s what the rates are, it’s squeaky clean, it’s statutory, so even long tail content owners that are not owned by the labels, for instance The Beatles, that tend to sue on their own, those guys need to get in line as well, because even if you do deals with all the labels, you still can get sued by these long tail people that own their own publishing and their own masters. Additionally, you need to get that entire thing done internationally because, like I said, every single European country has a different licensing body. Can you imagine, they all want their advances, and they all have their own different deal structures, it’s very challenging. To me, I think the solution is very similar to software patents, quite frankly, where the government has to get involved. Sorry!

This is kind of how I felt when it was over. You can take it for what it’s worth, but it’s similar to what Reid was saying earlier, you just need to know when you’re running against the brick wall, and you know when to pivot. At the end of the day I think that there are some things that you’re probably better off not even getting involved with. That’s certainly how I feel about this, I got a lot of great experience from it. I’m pretty sure that’s the only winning strategy with music startups.

So, take away is here. Don’t be cannon-fodder, it’s not good for anybody. Work on things you love, you know, your life is short. I started all this when I was 23, and I’m 30 now. So, here I am, I’m working on something I’m really excited about. I have a lot of the best team that I worked with previously and we did, you know, as the previous talk that I was talking about, working on startups that scale. I’m applying all that to my new startup, it’s a lot less stressful, it’s a mobile application, I think it would be great if people had a chance to check it out for iPhone and Android. And, you know, talk to me on Twitter about all this, that’s pretty much it. So, let’s do Q&A. I don’t know how I’m doing on time. 5 minutes, alright. Yes!

[Edit: questions are almost inaudible, if someone can do a better job of digging out what was actually asked, then I’d be grateful.]

Audience: I ((ran)) a music site for a while. It was (( )) never even making any money.

Dalton Caldwell: I don’t know. [laughter]

Audience: Cool. So, I shut it down. It was called Hip Hop Goblin, by the way. Anyway, what about starting your own label?

Dalton Caldwell: That’s something that one of the guys that started Vimeo did. It’s hard to make a lot of money fast, and that is not a startup because you’re directly tied to the revenue stream of that particular artist. And also, one of the things that I learned to appreciate is that the people that listen and find artists and do production are freaking geniuses, people like Jimmy Lovine and other guys like that are really special talent and I had no respect or recognition for what that was, and those are the people that start labels. So, I would not try to raise funding for a label, but if you have great connections, you have a good ear for music, and that’s something you want to do, heck yes, I think…

Audience: How about crowdsourcing for funding ?

Dalton Caldwell: I would use Kickstarter, man. Kickstarter has a little bit of that.

Audience: I’m going to quality control for artists (( ))

Dalton Caldwell: I think it’s an interesting idea. I can see a bunch of friction points there, and I would be happy to talk to you about it afterwards. But it makes sense to me, I just see some friction to ever get into that scale, basically.

Audience: Thanks.

Dalton Caldwell: Sure. Yes!

Audience: (( ))

Dalton Caldwell: Excellent question. Are the labels going to collapse? I think it’s already happening in slow motion, and I don’t think the labels are going away, they’re just getting smaller. It’s just like any other industry that reaches some kind of saturation point, that’s not a technology company where… These are really valuable copyrights, owning the copyrights on The Beatles is profitable, and you’re going to make a predictable future revenue off of that from doing commercials and all that. So, the labels are getting smaller and smaller. When you think about it, labels are completely the same thing as VCs, where they find talent, they write them checks, they help use their connections to get a really cool production, you know, like “Hey, Rick Rubin, come produce this track.” You know, you bring in other artists and then use all your connections to get their stuff on the radio. Then, when the artist makes money, you make money. So, I see incredible parallels between what labels actually do and the VC industry. So, there’s going to continue to be labels, they just cannot write checks anymore. During the 90’s they were signing all sorts of bands and writing them multi-million dollar bonus checks. Now the size of the check is much smaller, the number of bands that they sign is much smaller, so that’s the net effect, it’s happening, it’s just going to continue to happen, in my humble opinion. Yes.

Audience: Hi. Do you (( ))

Dalton Caldwell: Sometimes.

Audience: (( ))

Dalton Caldwell: You know, I would not be… If I would have just started some other startup and it wasn’t successful and I wasn’t a Sequoia CEO and all that, I wouldn’t be here. So, I have a lot to credit it. Think about what Ron what saying earlier about Fanning, and about Parker, they didn’t make any money from Napster, but they built a personal brand and they built a network. So, yeah, I’m not going to look a gift horse in the mouth and say I totally regret it. I got tremendous value out of it, but on the other hand it does suck to see something you build blow up. Yes.

Audience: If you think that the labels are getting smaller and smaller, what is (( ))

Dalton Caldwell: I think everyone is going to continue to make music anyway. Let’s talk about… I think it’s similar to the VC thing. Let’s pretend that for whatever reason VC firms got a lot smaller, which is happening anyway, what’s happening with the Angels. The way that things are financed, the way music is financed is going to change dramatically, and I can’t possibly predict if music will become higher quality or lower quality. I don’t know, but the fact is people are still making music, everyone now has laptops with pro tools on them so they can make studio quality stuff. So, a lot of the reasons that labels existed to do physical distribution, as well as recording, no longer make any sense. So, yeah, I don’t know if music is going to get any better or worse, but there’s going to be less $5 million advance checks for bands that sound like Nirvana the week after Nirvana breaks on the radio, right. That’s how the music industry, it was all people following trends and funding people that sounded like other people.

Audience: (( ))

Dalton Caldwell: Yeah. It was good for some time. Last question.

Audience: What do you think happened (( ))

Dalton Caldwell: They bought it because Bill is really good, and their ((front-end)) guys are really good, and they promptly shut it down because of licensing reasons. I don’t actually think it’s coming back. From what I hear, the whole team is not even working on music stuff anymore.

Audience: (( ))

Dalton Caldwell: That is what I understand, and they’re really talented guys, so, that makes sense, but I don’t think they’re working on what everyone thinks they’re working on. Alright, thank you! <!– 100 –>

HN Submission/Discussion
If you read this far you should probably follow me on twitter: