Supernova 2005: Long Tail Panel

What follows is my transcription of the last 20m of the Long Tail panel at Supernova 2005. Amazing stuff with Dave Goldberg, VP Music at Yahoo!; Jeremy Allaire, CEO of Brightcove; and David Hornik, a VC at August Capital.

Amazing factoid from Goldberg: they're closing one rock radio station per week. Audiences for rock couldn't get what they wanted from rock radio, now getting it from other places.

People discovering stuff online is driving offline physical sales of CDs, which reenters radio and MTV and driving plays. Too late, though. Y! has a thesis: music will disappear from terrestrial radio within ten years. Don't know implications for preferences, but will change way artists get invested in and marketed.

Major record labels and movie studios have controlled distribution. When you take away that distribution, they have to be good at either marketing or investing. Right now they're good at neither. All these things will change at the same time.

Q: Would be good to have data you have about long tail availability and how it's acted on. Put that out changes the entire debate about piracy and distribution. You have in your hands the way to effect the policy debate.

Yahoo!: The tail is impacting the head content, it's happening in real time.

Q: How can that kind of distribution be stable? Over time shouldn't it flatten?

Y!: What happens in our filters is that we see stuff start to occur and it starts to spread out among the community. People like things or dislike things. Even though we have billions of ratings, the average rating is almost always exactly average. Music is a very binary thing: people either love it or hate it. Our ratings, even though people we give people 0-100, most are either 0 or 80-100. other content is very difficult from content. narrative content, people have a range of opinions.

Anderson just put up a slide of offline vs online sales. two very similar long tail curves. Soundscan data for cd sales and online rhapsody data. niches sell twice as well online as offline. Different graph: Top 50% of cd's account for 80% of offline but only 20% of online

David Hornik: looking at online communities like Friendster, a lot of value is the potential to become the new influencers, replacing radio and mtv. People are going to have to bet on what they think will drive sales over the long-term.

Audience: Radio created popular music, not the other way around. With long tail, filters became just as important.

Audience: If you look in other markets like Amazon book market, which is larger and a longer time market, people buy variety. In the long tail, someone comes online, finds through filters additional things that are additional to them, and they buy additional quantity and greater variety. Lots of data about that. Suspect we'll see the same in music.

Audience: Question the idea that radio created popular music. To me, radio was the foundation for which it was possible for things to become popular. Traditional media coopts what's popular and amplifies it, but doesn't create it. Adam Curry got a job with Sirius. SF radio station that went to podcasts. We see that over and over again, media good at smelling what's new. Problem they have this time is that what's new is not under their control. For first time, separation of content. Text no longer in web pages (RSS), audio no longer on radio (podcasts), video no longer on TV (BitTorrent). Read Murdoch's speech to senior employees a month ago, berating them for how bad News Corp has been at getting into the new media. Unlocking library a way to make money, but they can't control content after that.

Audience: Some of this has to do with where the capital investment and marketing dollars are going. Nobody makes a movie for $40M for that they think suck, but there are a lot of movies made for $40M that suck. The balance of where the capital investment is to where the return is. The long tail enables a cheaper form of distribution to create revenue for things that would be too expensive to hold in retail space because of the costs of that retail space.

Audience: At the same time, payola will also help.

Hornik: There is this other idea, the long tail fo software. To create software that isn't the big enterprise multimillion dollar sale that is still critical. If you spend $40M making software, hard to distribute it for $50 on the Internet. But if you only invested $1M, the options are different. Some businesses build a long tail business and VCs offer them money and they say "no, we built a business around not needing your money".

Audience: Don't you think you ultimately reach the point where efficiencies in the model drives the cost down almost to performance royalties at fraction of a penny, needing micropayment infrastructure that doesn't exist if you draw out long tail to its maximum efficiency.

Anderson: Most people don't actually want to make money from their music. Most would like to give theirs away for free, some would be happy to give it away if they could track usage, and only some want to be paid. Perhaps you're not paying for song, you're paying for access to filter.

Hornik: Ahem, saying "most people dont' want to make money" is a misnomer. Most don't expect to. If they could, they'd expect to make money. If there was a system in place, who will say "I don't want that quarter?"

Anderson: most content producers, given choice between smaller paid and larger unpaid market, will choose larger unpaid.

Hornik: rhubarb rhubarb

Goldberg: Free or subscription models. TV today, capability exists to have free (ad supported), subscription-based, and pay-per-vew. PPV available to 40% of US population on TV sets, but content not available. Not rights, not technology, a business model. Transaction and acquisition costs. Costs you a lot of money to get someone to do something one time. Film business couldn't justify amount of marketing to get you go into theatre and pay $10. Justified because acquisition cost spread out over entire life cycle of the content. Aggregators like us say we can't figure out how we would ever make money on a pay-per-use model when what you care about is maximizing $$$ from customer in a givne month. Avg consumer spends $7/mo on music. If I can get that $7 without acquisition or distribution costs, I'm motivated to do that. Everything moves to subscription model because of acquisition costs.

Allaire: music, perhaps. Textbook publishers will not take blog adwords for small audience high-dev-cost works. video industry don't make money from ads while the movies play on tv.

Hornik: $20 DVD is new. Knows the person at Disney who suggested $25 DVD. Disney rubbished: DVD rental place pays $60. But she convinced them to try it, and it revolutionised DVD. What we don't have on video yet is the portability to make the particular piece of content in video form sufficiently multipurpose you'd pay the price you'd need for on-demand. These things are much more complicated than single issues, often driven by technology costs and time.

Audience: Have you explored tagging to contextualize preferences engine?

Y!: I was going to say this before, but we're still very early on in finding ways to do filtering. We've figured a coupl eof things that work. We look at it as: these aren't the only things, there should be more. We're constnatly looking at startups, IBM research, etc. Most of the time they don't scale or they don't work. Y! just bought Flickr. Music different from photos, but how we get metadata is one of the hardest problems in this.

Audience: I've spent time in the business. Good panel, because you have source of capital and distributor. Cost of production of film > TV > music. Studios have provided capital for a basket of inventory, same way you'll basket of companies in same industry. TV puts a lot of shows in preproduction, see's what's there to put capital in. Long tail can disintermediate, but without VCs to fund it, you will have a collision somewher eon how you'll fund it

Y!: We don't believe we're replacing that function. We're disintermediating the distribution. Someone will provide capital and marketing.

Audience: Studios provided capital. Controlled distribution to try to guarantee a return on that capital.

Allaire: Studio as bank is strong analogy. Even cable and satellite networks are financiers. As spectrum widened, # people who could finance and bring products to market has grown. On internet scale, # people able to viably finance and distribute gets larger. Doesn't go away.

Y!: But not necessarily existing people. Record business today is huge destroyer of investment capital. Huge. Look at public record companies. Separate their catalog from their new release businesses. Pretend it's half and half, though in some catalog is more than half. For most, spending on new release is 300%. Is that the best use of catalog cash? Will shareholders want this? Time-Warner didn't.

End of panel.