The Long Tail | Chris Anderson | Talks at Google

Got bad, stopped making films. Sorry, Ben Affleck. That's a good guess. Ben Affleck's responsible for just that one right there. Yeah, exactly. They ran out of screenshots. This is where the carrying capacity of the American Megaplex Network is about 250 films a year. That's where they ran out of screens. They didn't stop making movies. There's 13,000 films shown in film festivals every year and fewer than 300 make them to mainstream distribution. The films don't get any worse. The films don't switch into Spanish. They don't all become documentaries, they just run out of screens. And so this right here, this is the dark matter of the marketplace. This is the latent demand that we can't see, we can't measure, we can't tap because of an inefficiency in distribution, because we have this one size fits all distribution method for films called the Megaplex, where it's got it, it can't just be popular, but it needs to be popular in the geographic area around the theater and popular enough to get about 3,000 people, people in over a 10 day run. Or else it doesn't get out there and they have a virtual monopoly on distribution or did until recently. And you end up with this crazy distortion of what should be a straight line. So this is the missing market. And Netflix, by not having a limit of screens, by being able to distribute via DVD and the web and YouTube by being able to do the same, don't suffer from the scarcity effect and do see this straight line effect. And as a result they're able to tap this market in a way that Hollywood traditionally hadn't been able to. So what we see right here is an excellent exercise when you're looking at long tail markets. This is the way we do it. We do this in business schools and elsewhere. Plot your. Take your market, rank the products or services or people or whatever it is you're ranking in your market and rank them in terms of popularity, sales, plot them, log log. And when you see the inflection point, that's your opportunity. This is where the bottleneck is. And if you can get rid of the bottleneck, you can tap this market. So let me give you an example of how that works. This is Amazon's sales. Forgive me for the crappy resolution. And once again we have Amazon sales rank. The dotted line represents Amazon sales rank of three years ago. This is basically the copies proportionate to the sales rank. So for example, sales rank 1000 would be selling about 100 copies per week and so on, all the way down to sales rank 1 million. And about three years ago, it looked like this. There was a straight line until there was this inflection point. And now just a month ago, we did it again. And now you see that it's getting much more niche heavy, that the low sales rank down Here in the 10 thousands are selling more than they used to, but still it's a straight line until you get this inflection point and then it falls off. So you're now trying to figure out, okay, we've clearly seen that it's a straight line for a long way, and we've clearly seen that it's becoming more niche heavy, that the small sellers are becoming more popular relative to the hits as time goes on. But you still have these bottlenecks. There's still an opportunity to do there. What might those bottlenecks be? Well, let me give you an example, just randomly chosen this morning. This is my book. And this is screenshots of an hour ago. I just want to happen to mention that little number four right there. Incidental, not why I brought it up. What you see is that you see a cover, you see a review, you see a blurb from this Eric Schmidt guy and fully populated metadata fields. But it turns out that, and this is the way it looks up there at the head, but it turns out there's this other book called the Long Long Tail. And oh, no, it doesn't have a cover, it doesn't have any reviews, it doesn't have any blurbs, its metadata fields aren't fully populated. And that, I think, is what happens. That's the reason that you see that inflection point. You see the inflection point because not because there's a limit of availability, there's a limit of findability. At this point, somewhere down the tail, the metadata falls apart, the covers don't appear, the reviews aren't there. If this thing was as well populated a listing as mine is, it would sell better. And this is a fantastic opportunity because this book probably did get reviews when it came out. It probably did have a cover when it came out. It probably did have metadata, and it's in the publisher's warehouses some way. It's in the publisher's records. And if they can only get them out of those records and onto this listing, this book is going to sell better. So let me give you another example that I don't have an answer for. I'm going to leave this to you as an exercise. This is Netflix's sales again, plotted. So what you see here is marketing. So right here, the top 10 in any given month still benefit from a lot of marketing. And that probably explains this inflection point. But down here, there's another inflection point. And I didn't plot this out all the way, but something's happening down here. Now, we could guess, and I don't know what the answer is. It could be that it switches at this point. Films are a foreign language. That would be a reasonable guess. That would be a perfectly reasonable scarcity function. And there's nothing you can do about it. That just is what it is. But it might be something like metadata or it might be reviews, or it might be that the liquidity of the recommendation market fails at this point, that basically there's enough people watching these films that the collaborative filtering works. But down here, these films are so thinly rented that the filtering mechanisms don't work and the recommendations aren't generating them. I don't know the answer, but I do know that if I'm Netflix and as I intend to scale this for the next to the next 60,000, that if I can figure out what's causing the problem here and if it's something I can do something about, that the opportunity to grow my market lies in getting rid of these bottlenecks. So let me talk a little bit about how power laws come to be. There's another book that there's actually two other books that I'd like to recommend. The first one is called A Fool by Randomness by Nassim Tlaib. And Nassim is a Wall street hedge fund manager and a quant. And he is focused, most of Wall street is focused on arbitraging tiny differences and inefficiencies in the marketplace, getting a penny here or there's spreading lots of bets and they tend to bet on what's going to go up. Sometimes they bet on what's going to go down, but it's largely within a relatively small range. Nassim decided that he instead was going to bet on extremely low frequency, high amplitude events. The very rare unpredictable things like September 11, like earthquakes, like just random crashes in the book, Fooled by Randomness says how much he believes randomness does play a role on Wall Street. And he was going to bet on, he was going to put basically all his chips on the low frequency, high amplitude events. And the consequence of this is quite demoralizing when you're in the business, because if you're betting on an earthquake, you can sit there for years upon years and it never happens. And you lose money every year because you've got a long on a position that isn't coming through. And that sucks to lose money every day while the other guys are making a little money every day. But when that moment comes, you make it all back and more so. Sorry. Well, there are other low frequency, high amplitude events that are good. I could probably think of some, maybe, I don't know, peace in the Middle east, But. So what seems about is the tyranny of the Gaussian. This is a bell curve, otherwise known as a Gaussian. And a couple things about bell curve. This is the way we've seen the world for our entire lives. It's all about averages and means and medians. And basically we sort of think of ourselves being mostly in the middle. And if you're a retailer or your company, you pretty much focus on one standard deviation. It's like, well, we're going to focus here on these middle four bars and let's not bother with the edges. And if you have to focus on one bit, you do. It makes perfect sense to focus on where most people are. But there's another way of looking at the world and it looks like this. And forgive me now for my conceptuals, but let's, you know, normally we think of this. Oops, sorry, let me go back one. So this, basically this is the way the world distributes when you have scarcity. So let's say that it costs a lot to do something. Well, you're probably, and the risks of failing are high. You're probably going to all focus on the middle stuff. So, you know, starting companies with no venture capital, you're all going to cluster here in the middle. You know, if you're, if you're making movies, you're going to cluster here around the formula. You're not going to try too many fringy ones. If the only way to be a successful artist is to get on top 40 radio, well, you're pretty much going to do what top 40 radio requires and woe be told, you know, woe to those who try to experiment too much along the fringes. But let's imagine that the cost of failure is zero. Let's imagine that the marginal costs of doing something essentially zero, like for example, writing an email or starting a blog, well then what you find is a very different distribution. Let's take blogs, for example. So now I'm using this fake word quality. I completely reject the notion of quality being meaningful in any way. So let's Just say quality is what you think it is. So if this quality means, you know, sort of the length of a post, that's fine. If you think it's the. If you think it's the sort of the number of syllables in the words, that's fine. If it's the language, fine. Whatever your definition of quality, we each have our own. And so what you find is that through most of these, you'll find is that because it's so easy to do something, you'll see most people, a few of them will get it absolutely right and most of them will get it wrong. And down here, these, of course, are automated, these are splogs and just random crap in MySpace pages. Not to suggest that they are in any way equivalent. So you see a very different thing. Basically, you see is that most of this stuff, that the bad quality stuff is actually most popular and that you have an inverted distribution. Now, this would be a bad thing. This is sort of the Internet, pre Google, right? So most web pages are crap, and it's very hard to find the good ones. However, we have this miracle of network effects. And what network effects do is network effects is like word of mouth. And measuring incoming links is another way of measuring word of mouth. And what network effects tend to do is they tend to amplify the good and. And suppress the bad. And so you take something like a search engine or like collaborative filtering or like just word of mouth, organic word of mouth through blogs or MySpace friends or whatever, and it turns that shape into that shape. It suppresses the, you know, in terms of the first page of results on a Google search, you've suppressed the bad sites, you've amplified the good sites, measured by, in this case, quality is relevance. And you've ended up. And you've ended up with something that looks like this. You now created a hit out of this overabundance of random stuff. By using network effects to discriminate and to filter that chaotic marketplace, you've managed to create a traditional power law simply through the power of network effects. So, oops, I was going to end up with one video and then take questions. Let's see if this goes. I'll start again. Sorry, I'll start one more time. Let me introduce this video. Peter Hirshberg from Tech and Roddy riffed off some ideas and made a video which is just fun. And it's a good way to launch into the next fun part, which is the Q and A. See if it goes.

No one would have believed in the closing years of the 20th century that our most pop regular media were being watched in a new way by a force that was quietly gathering strength. With blind confidence, we considered them our own. Our audience, our subscribers, our cuddly couch potatoes. With infinite complacency, we blithely segmented and sold them to advertisers. We learned to shape their habits, mold their desires, and give them the illusion of infinite choice. Yet from their side of the screen, with envious eyes, they studied us. And bit by bit they learned and linked and drew their plans against us wielding weapons we ourselves provided. And then early in the 21st century came the great unraveling. We offered free choice, but all they heard was free. We devised more powerful, more complete, more feature rich software, but they preferred to grow their own. They pounced and they blocked and they blew the house down. The world we knew was forever disintermediated, whatever that means. They have tasted power and there are a lot more of them than there are of us. High inside our glass towers, our greatest minds prepare to respond.

Well, can't we just start blocking back at them?

Well, yeah, but where's the money in that?

Thank you.

So I've got some time for questions, so have at me. Thanks. Over there in the block. So I definitely, I definitely see argument as a large opportunity in the more access. But I really skeptical of the end of common culture. I mean people still watch the Super Bowl, Sure. I get to have it both ways. I get to both entirely agree with you and defend my thesis. Actually, I've never said it was the end of common culture. It's the end of monopoly, of the hit. So you always have hits. We're a gregarious species. We like to do things that other people like to do. The problem is there's two kinds of hits. There's organic hits that really do resonate broadly across our culture. There's the World cup, there's American Idol, there's csi, and then there's synthetic hits. Garfield two or others. Synthetic hits are the ones are hits created by a machine that needs hits. These are hits that are essentially, you know, products of, you know, mediocre quality that have the full marketing force of the blockbuster apparatus around them. And I think what you're going to see is you're going to see these synthetic hits are going to struggle as word of mouth spreads so much more rapidly and they have to compete with a million, you know, a million smaller, often amateur, you know, products. So you'll still have, you know, you'll still have hits and you'll still have and you'll still have, and you'll now have niches. The difference is that you're going to have the organic hits at the top. You can have these bottoms up hits, these new kind of grassroots YouTube phenomena or MySpace phenomena, the Arctic Monkeys. Clap your hands, say yeah, kids, lip syncing video hits. And you've got the new hits from below. You've got the good hits top down from above. And then you've got the bad hits struggling and now having to face more competition than they've ever had to face before. Yes, here in the blue shirt with the scoping crossbones. Yeah, sure, sure. There's loads of them. Obviously Google, you probably don't consider yourselves a media industry, but Google's the long tail of advertising. So all those keywords are ads. It's the long tail of advertisers and it's the long tail of ad driven publishers, all those blogs. So all these things managed to scale the company scaled down to reach the market that traditional media couldn't approach. And this is my background is physics. And my hero growing up was Richard Feynman who said famously, there's a lot of room at the bottom. And the notion we all talk about scaling up all the time and you've got to scale up your infrastructure but scaling down your business to the huge bottom, that didn't surpass the economic test of the traditional industry is what you're able to do with these self service models like, like adsense. So there's kind of an adjacent space. Obviously ebay is the long tail of hard goods and the long tail of merchants with all those mom and pops as well. I used another example of a capital one, the credit card company. In the 1980s, there were basically two credit cards. There's a credit card for people with good credit history, a credit card for people with medium credit history, and no credit card for people with bad credit history. And then what Capital One realized is that there was this new abundance of information about people's personal credit histories and then processing power to be able to crunch that and to not have a one size fits all credit card for people with bad credit histories, but instead personalized credit cards with personalized rates for each one of them. And they were now able to offer credit cards to all those people who weren't able to get credit cards before and scaled down and basically double the size of the business by finding the customers who were sub economic before. This created a national phenomenon of debt, which we're now living with. But it was a hell of a business for a while. And I'll give you one last example, my favorite example of the moment. Last month Anheuser Busch started a new division called Long Tail Libations. What's the Long Tail beer? Microbrews, craft beer, regional beer, all the supply chain efficiencies that have made it easier to have more variety. There's been a doubling of variety on the supermarket shelves over the last decade. And all those stock keeping technologies that allowed them to have smaller number of products on those shelves allow you to go beyond Pabst Blue Ribbon and get into 40 or 400 regional and microbrew beers. And now Anheuser Busch of all people, has found a long tail in alcohol. Is this true for businesses as well? So are there more smaller businesses? Small, yes. I mean, so most of the growth in the US economy comes in small and medium sized businesses. And that's absolutely true. The barrier to entry to starting a company has never been lower. As we here in the Valley. I know very well there's 750,000 people in America who make at least part of their income off ebay. Another perfect example of this, but one of the other. Yeah, yeah. So this is your point. So we have a long tail of producers and a long tail consumers and a relatively short head of aggregators. You know, ebay, Amazon, Netflix, Google. And it seems. Oh well, we have a contradiction here. We have oligopolies. I mean the marketplace sense, I think that the story of search is actually quite telling in this matter. So Google dominates traditional search, but now we have vertical search, images, maps, et cetera, et cetera. Well, so video YouTube for blog search, that's probably technorati or somebody else. And it turns out that just because you dominate the canonical big marketplace doesn't mean you're necessarily going to dominate all the verticals. And I think the big opportunity going forward is to find, is to find vertical marketplaces for content and products that aren't well served by a one size fits all model. And I'll give you one example of this. If I was going to start a business right now, it would be a classical music service. Classical music is pathetically underserved in the bricks and mortar market. It's one of the largest areas in itunes. Not because the demographic, classical demographic coincides with the ipod demographic, but because people who are into classical are so poorly served. It's the only place you can get music. This is a special problem in the U.S. okay, it's a special problem in the U.S. but let me just finish the example and tell you just how bad a problem it is. It's a metadata problem. The presentational form for music is basically based on the lowest common denominator of metadata available in all the data sets that come into companies like itunes. And it's the pop model. You've got band, track, album, and sometimes label, but not always. Classical, of course, doesn't. That's meaningless. You want composer, conductor, soloist, lead violin, you know, town of the. Or you know, you can think of any number of ones and none of those are available. And as you know, one of the problems with MP3 players is they don't even have a space for metadata fields like that on an ipod. So classical, despite being popular on itunes, is actually quite poorly served from a presentational form on itunes. And you imagine what if they were, what were there on itunes that really was just set up for classical, or one that was just set up for jazz, for example. With jazz, it's not the band so much as it's the individual performers. On IMDb you're able to sort by, you know, you can see like, you know, all the cinematographer's movies or the screenwriters movies. Every one of those metadata fields is clickable, sortable. And you can reorganize the entire marketplace along that because it's populated, it's peer production populated. Those metadata fields are all created by fans of movies who submit the information. And the question is, where's the IMDb of music? Where is the wiki music? Where is the music sites that actually have all these abundance of information that is meaningful to some of us? I'm into the DC punk scene. Why can't I sort under DC 1980? That information is out there. And I think there's an opportunity for companies to build marketplaces that specialize in these niche verticals. Yeah, right here in front of how.

Do you think about the public browses of choice? As long as it increases, you have more and more choices.

Exactly.

There's been surveys, for example, supermarkets. Oh my God, there are 15 kinds of farms and flakes.

Right. So this is the paradox of choice question. Fortunately, I take a whole chapter to Savage. Barry Schwartz. Barry Schwartz wrote a book called the Paradox of Choice. And he uses the famous Stanford Jam example. You know, the researchers put six kinds of jam on a table and, and they asked people, and people would buy a certain number of them, and then they put 30 kinds of jam on the table and people would buy fewer of them and they'd be less happy with their purchases. Because why do I need to be an expert on jam, this is so confusing. I bought the apricot, I should have bought the strawberry. And this is called the paradox of choice. And there's only one problem with this theory. So his theory is more choice is bad, it's oppressive, it makes you feel bad, it's paralyzing all this sort of thing. There's a really easy answer to it which is, yes, choice is oppressive unless you have help. Now, the Internet is an abundance of choice unless you have Google, but you do have Google, you know, Amazon is an abundance of choice unless you can rank by best selling and ratings and reviews and price and all this kind of stuff. So I went back and actually looked at this and so I went to the supermarkets and I said, how many kinds of jam do you have? And I went, they had 320 kinds of jam. It doubled in the last six years since the Stanford study came out. So either they were stupid or there was something going on. Of course, it's not a blind choice. We have brands, we have our preferences, we have our kind of inbuilt experience. Things are ordered in certain ways. Some of that choice is not. Although it may be 300 kinds, it's actually 70 of them are organic, another 40 of them are light, et cetera. So there's lots of things we have built in to help us make that Choice. Amazon has 1200 kinds of jam, as it happens. You know, confused. Ranked by best selling, pick the top. It's as easy as that. And so Barry and I go, you're picking. I love this. Fortunately, I've heard this one before. So the question was, doesn't this contradict the long tail by picking best selling? Haven't you privileged the hits? Haven't you gone to the hits? OK, so here's the difference. So I hate top 40. I think top hundred in music is a completely useless qualifier because it's apples and oranges, it's hip hop and then an R and B and then an alt country and then a country and then a pop and emo, et cetera. Now go into a genre. Okay, now go within Afro Caribbean jazz. And now top 40 is perfectly sensible because it's apples to apples because you're within a genre, you, you've actually created a micro hit and a mini star way down the tail. So that jam. So when you go in the Supermarket, the top 10 sellers in the supermarket are going to be bananas and milk and bread, et cetera. And it's a mishmash. Apples and oranges within the jam category. And remember, this is.

Thanks.

Now go within the jam category on Amazon. It's not all food, it's not all products, it's not all jam. You're listing under Organic Jam. And now you've got the equivalent in Afro Caribbean Jazz. You've got the best sellers within Organic Jam. So yes, it's a hit, but it's a tiny micro hit way down the tail. And I think that at that point, the aid in making a choice, you know, vastly compensates for any slight hitism that might have crept in. Some more questions. I'll take one in the back with the. With a black shirt. Yes, please. Can I use the example of books? What'd that do? Okay, so I think one of the most sort of cheering and heartwarming stories of our age is the dramatic shift in the used book market. Ten years ago, used bookstores were a failing small business industry in competing with the Barnes and Nobles on one side and Amazon on the other. And the problem with the used bookstore is you go into a store and you're looking for a particular used book. You go into a store, first of all, you know where you got to find the used bookstore. You got to get to there. Who knows what they have? Then you have to navigate their crazy unique shelving. And then you find they probably don't have it anyway. And you go to another bookstore and you're disappointment, disappointment for disappointment. What happened is that they all took their inventories and they digitized them. They put them online in a, you know, in a common data format. And then they made it. And they put through Elibris, they made it available on Amazon. And now on Amazon, if the book isn't available new, you can see if it's available used. And they basically brought liquidity to the marketplace. As a result, used books are now the fastest growing part of the book industry, growing by about a third a year. They now represent 10% of the book industry. And most and most importantly, I think this is the part that really speaks to your point. They've made the concept of out of print obsolete. So what does out of. I don't think my children will have any idea what out of print means. Out of print means you've got to click this link versus that link. It's still available to you. Who cares whether it's in print? It's out there. Now, from an author's perspective and a publisher's perspective, you might worry that you're not getting your revenues from that sale. But from a consumer's perspective, it is a fantastic increase in choice to give You a couple data points. There are 32 million books in American libraries collectively. 32 million unique titles. About 1.2 million of them, maybe 2 million of them are really in print and easy to find. Another maybe 20% of them are out of copyright and could easily be scammed. And you guys are doing that. And the vast majority, about more than 70% of the books in the middle are unavailable, which is to say not in print, not out of copyright, ambiguous intellectual property state. This is why you guys are in court and fighting about this very subject. But those books are available in one way and that's available through used bookstores. So our shared cultural history is locked up in books that aren't available for most of us. And the fact that these used bookstores took information that was latent in the system and simply by digitizing it and aggregating it and making it available through a retailer like Amazon, Libris and others, they've unlocked that shared cultural history. And I think that's terribly exciting. Yeah, up here. What is your view of the long tail in politics? Oh yeah.

Arbitrarily small laws and some people call it pork barrel.

Right. So this is, I get. Yeah, this is the politics question, which is the third rail of public speaking. So you know that Gaussy and I showed that bell curve. Well, we have reduced the diversity of political views in this country to two poles, you know, Democrat and Republican. But it's, you know, almost absolutely sure that we're a sort of a more even distribution than that there's a scarcity effect because we have representative democracy and because we have a party system and because it's difficult for people to become informed on individual issues, we end up with, you know, with the two party system. But you can imagine if you went to something with more like direct democracy where you could poll people individually and let's say you didn't go, you didn't vote for individual candidates, but you voted for issues. I think you might find that there would be a broader spread of political views and you might find, you know, maybe a third party or independents or you would find that there's actually that we, you know, that the two party system doesn't actually suit the true diversity of opinion in America as well as it does, as well as a more free marketplace of ideas. Now how to make that happen, I don't know. But I do would say that we are. That politics is an example of a system distorted by scarcity effects. You would get the California ballast, we would get the California ballots and we can Argue whether that's a good thing or a bad thing. But it's certainly a diverse thing. I'll take. Let's see, I've got another 10, eight questions. Okay, I'll take. I see four hands. I have four hands. So I'll start there, then there, then there and there. In the back of the white shirt. Yeah. Are there typical exponents on the. No. I mean, what we're trying to do is we're trying to quantify what the constants and the exponents are in the power laws. And by the way, it's not just the distribution between hits and niches. It's also, over time, you have a decay function. So today's hits is tomorrow's niche and you have a decay function over geography. Japan's hits are our niches, et cetera. So it's a pretty complicated 3D topology as you look over time and interest, between broad and narrow, and each market has different exponents. We try to, you know, sometimes in the case of just the decay function of music over time, we are quantifying that one. And, you know, I can come up to me afterwards, I'll show you some of our research. I think next to, next to him there was another question. There was another hand up there. Back there. No. Okay, I'll go to the front. There were two. Yes, here.

So I was wondering if in a lot of these cases you find a power law going the other way that constraints somebody's sake. For example, for Amazon, for the first 1,000 books, it's easy for them to find the metadata. The next million is maybe not quite so easy. Once you go past that, it's just impossible for things. Also, if you're selling books, Once they write 20 books, people are less interested. Once they write 50, they just crap because you've read it all before.

Well, to your first point, I mean, there are definitely metadata constraints. So obviously when we're looking at one of the big long tail opportunities, the archives, you know, television, for example, is the biggest one. Television's a crazy business. We spend all this money making content. We broadcast it for this tiny window, and then it vaporizes. A tinier fraction is syndicated for later rebroadcast, and even smaller fraction comes on dvd and the rest just is gone. It's gone into the archives where it typically is never seen again. In that case, the constraint is rights. You've got to clear the rights. You got to clear the rights to the music, you've got to clear the rights to the performance. You got to clear the rights to the script. The famous example is WKRP in Cincinnati, which was a 1970s or 80s comedy show set in a radio station. The problem with that show is that in the background there's music, pop music from the era playing all the time. There's 30 or 40 songs they have to clear for each one of those episodes. And if they can find a way to clear that efficiently, then you can, then you can tap our collective archives of television content and other kinds of content and be able to free up. But at the moment, at the time being, for the time being, that is a real barrier to tapping that and does create a kind of a focus on the bits where the rights are available. I think there was one question next to you there.

Yeah, I was wondering if, let's say somebody from the RIA or the MPA is looking at this book. Well, wouldn't they say it's not really that the bottom end is coming up, but that the top end is going down? Like, I'm not that I've ever done this, but there are some.

I think, you know, so what does the rise of the tail do to the head is the question. And yes, you know, now of course, itunes and rapsteading. The services we're looking at are commercial services and people are being paid and we are seeing a strong shift towards the niches because they are available. So you're seeing a generation who grew up with the ipod being very used to going through the commercial route. I mean, as you know, the peer to peer services have a lot of availability, but interestingly, they suffer from a little head problem as well. It takes something like BitTorrent. BitTorrent. Because the torrent seeds tend to be what's popular at the moment. The niche stuff actually tends to fall off the torrent seeds. And on some of the other peer to peer networks, you actually find it's quite hard to find some of the niche content because there's not enough liquidity in that marketplace. So I think actually the commercial services do have a place and the labels do have a place because they're the ones who own these back catalogs that have so much potential value. And as a matter of fact, the strongest sales at the moment are in the back catalog. So I think that, you know, that again, yesterday's hits are today's niches and that, you know, there is a marketplace for this stuff. We're starting to see just how big it is. And I suspect that piracy is always going to be, you know, a portion of the marketplace. But as long as they can offer something as additional value like this metadata and high quality and low prices, that it's going to be a net positive and it's going to grow the market rather than cannibalizing it. Do I have time for just one or two more? Okay, yeah, here. Great. Right. So the question was, it seems like a lot of the niche content is older stuff. What's the incentive for people to create new stuff? Because that's not scalable. Just to be clear, there's two long tails. One's the long tail of time where that's the archives. The other is the long tail of interest between narrow interest and broad interest. Broad interest is the hits and narrow interest is the. Is the niches. What's the incentive to create stuff of narrow interest? I mean, why do people upload their music to MySpace? Why do people upload their music to YouTube? Why are there 60 million bloggers out there? The reality is that money is not the main reason why people do this. They do it for other, you know, the whole notion of pure production, you know why? To make people make Wikipedia. They didn't do it for money. They did it because the tools of production were put in their hands. I'm a blogger. I also run a mainstream magazine. I blog for free. I run the magazine for a paycheck. I think this notion that people only, that the economics of the head, that people are only motivated by money is the real fallacy of our time. And we're seeing that people are motivated by reputation and expression and fun. I used to be in bands. People are in bands because it's fun to be in a band. It's a great way to meet chicks too. And what do you want? If you're a band, how many of you are going to become a hit? I mean, almost none. Almost no bands ever become hits. Does that make all the rest of us failures? No, we had fun. And if we can get people to come to our live shows, possibly by giving away music on MySpace or someplace else, that's fantastic. And that gives us a bigger audience than we would have had otherwise. And so what if we don't quit our day jobs? It's more fun. I think the rise of the non monetary economy and these incentives that aren't commercial is the real lesson of the long tail. And I guess I can take one more. Sorry. Okay, one more fast question in the black shirt back there. No, I think it's the oligopoly question again. Comes back to the sort of, you know, the irony that we have a small head of markets at this place. I think it's a short term. First of all, I think we're just starting these marketplaces and it's absolutely natural that the marketplace with the biggest liquidity and the biggest choice is going to dominate in the near term, I think over time, because that information is largely open and available to everybody. You're going to find specialists who do market out. So just in search, who went to vertical search. And I think that Google, which might have been seen as having a monopoly, is now, from a vertical search perspective, seeing as sort of having, you know, competing with a lot of vibrant startups. I think we're going to see the same in these other places. Ebay has by no means a lock on their marketplace, as you well know. Itunes does not have a lock on their marketplace. Netflix does not have a lock on their marketplace. All of them are going to have to fight with a lot of competitors going forward. As for the network neutrality thing, that's a longer question. We can talk about that offline, but at this point, I think I just wanted to say we've got books. I'll stick around and sign for a few minutes. And thank you so much for coming.