To give some background, last year ViralManager.com did a study where they indexed a “large number” of YouTube videos, and counted how many views they got in the first month. If you just look at the graphic they provided, it appears to be a compelling long tail story – it drops of really quick through the top 10%, then has a thick “long tail” that extends out through the next 90%. However, if you look at it closely, you’ll note that the scale on the Y-axis does not increase arithmetically, but rather geometrically. In other words, each mark is 10x higher than the previous (i.e. it goes from 10 to 100 to 1,000, and so on). So I did some quick math to give an indication of what it would REALLY look like on an arithmetic scale:

Ok – so I didn’t actually make the graphic, and it’s easy to see why not. If I did, and I wanted a view count of 10 to be represented by one centimeter on the chart (so it’s viewable), I’d need the Y axis to be about 10 kilometers long. Once you hit the top 1% marker (i.e. more popular than 99% of other videos), you’d be around the 500 meter mark. At the top 3% marker, you’d be around the 25 meter mark. At the top 10% marker, you’d be around 1.5 meters. At 50%, you’d be around 10 centimeters.

Or for another way to look at it, as I wrote about on my blog last week, it would appear that the top 1% of videos are viewed several hundred more times than the bottom 50% combined, and about fifty times more often than the bottom 80% combined.

To me, that looks like a blockbuster model – and based on the viewing habits of people I know that go to YouTube, this makes sense (many simply check out whatever is most popular, which becomes a self-perpetuating cycle). But what do you think – am I missing something here, or is the long tail really not that important no YouTube?

One of the papers I found most interesting lately was Revealing the Long Tail of Office Conversations, by Michael J. Brzozowski & Sarita Yardi. What the authors were interested in exploring tied was how social media tools (blogs, wikis, etc.) could breakdown geographic distances, work group boundaries, and organizational hierarchy in the organization. More importantly, they wanted to look at what motivated individuals to “invest their own time in creating content for public consumption.”

In order to do so, HP decided to conduct a study of themselves – the design and use of Watercooler, their social media platform. Included in their study was a survey of 144 users, 12 months of log files from the internal blog server, and interviews with approximately 96 blog users. The paper is broken up into sections on the design of Watercooler, results of the experiments, perceived benefits for users. The paper is also quite short- only about four pages – so it is well worth the read.

What I really liked were the readership and commenting network “social graphs”. I know they’re kind of hard to see on the screen, but in the first each node is an individual user color coded by business group, and each arrow represents someone having read someone else’s blog at least 3 times. It is notable that a couple of authors clearly emerge as “hubs” in the network, and also that 55% of links went outside of one’s own business group. It is also clear that the two green dots in the upper left corner are having an affair of some sort (just kidding).

The second figure focuses on comments instead of views – and the results are even more striking. One individual (J) emerges as the center of the giant component, due to a bit of a virtuous circle – she comments a lot on other people’s posts, and they in turn comment on hers. It’s also notable that there are several “disconnected clusters” where people are chatting away. And those two green dots on the side are still clearly having an affair (just kidding again).

I don’t want to rehash all of the arguments in their entirety, but rather focus on carving out a middle ground between the two POVs.

One one hand, technology is enabling a such a rapid increase in the volume of “units” being produced in certain categories (think: digital music tracks for sale) that what was previously the long tail is now being pushed into the “head” / blockbuster category based on one commonly used definition – the top 10% or 1%. I don’t think such units really belong in the “head” category.

On the other hand, I also don’t think they fit into the long tail anymore (which Chris defined in relation to availability in bricks and mortar outlets, a definition that needs to evolve in markets where b&m is rapidly decreasing in importance). This leaves them somewhere in the middle. Since the body is in between the head and the tail, I decided to go with that (after an embarrassing foray with the term ‘middle tail’) – could the body end up being more important than either the head or the long tail?

However, I realize what I just wrote might be clear as mud, so let me provide a numerical example that builds on the research.

According to the article, the entire music inventory of a typical Wal-Mart store is equal to about 10,000 tracks, which for the sake of simplicity we’ll say represents all the music options available to consumers in a pre-digital age. Let’s say you define the “blockbusters” as the top 10%, and the long tail as the rest – that would be 1,000 blockbusters, and 9,000 in the long tail, and you could map out the sales accordingly. Alternatively, you could argue that that this entire group of 10,000 represents the “blockbusters”, and the long tail is what was sold via concerts and indie distribution outlets. It really doesn’t matter too much, as you will see in a second.

Jump forward to the digital age. The paper in question used Rhapsody for their digital music analysis, which offers more than one million tracks – let’s just say an even million. Even if you redefine “blockbuster” to only account for the top 1%, this now equals 10,000 tracks – or the entire music selection previously available at Wal-Mart. If you use top 10%, it’s approximately 100,000 tracks, and thus captures 10x what was previously available to “ordinary” consumers.

This leads to a rather obvious observation: as the sheer volume of content available to consumers has increased so quickly, what might have previously been considered the “long tail” can now – statistically speaking – be counted as part of the “head”, if you use the percentage method. The HBR article notes the top 10% of Rhapsody singles account for 78% of plays (and the top 1% account for 32%), which sounds highly concentrated, until you remember – as Chris notes in his response – this accounts for 100,000 songs (or 10,000 for 1%). Does anyone really think there are that many “blockbusters”? Does the long tail really not begin until the 100,001st most popular song?

So rearranging those numbers a bit, consider the following argument. Let’s say that right now the top 1% represents the head – that’s 32% of sales. The bottom 90% capture 22% of sales, which I’ll call the long tail. In turn, that 9% jammed in the middle is capturing 46% of sales – which I’ll call the body. While I’d need further research to verify it, the cost of developing and marketing that 9% is likely significantly lower than the costs associated with the top 1%. Might this body of work end up being the most important (read: profitable) of all? Perhaps even representing the content that is so good it sells itself?

There’s obviously still an underlying definitional problem here – as content is constantly added and nothing vanishes, what is the top 10% today becomes the top 1% sometime in the future, and the dividing lines might have to be moved again. However, the underlying concept might be an attractive one – maybe it’s not the head or the long tail that ends up mattering so much as what’s in between the two.

I like the strategy for a number or reasons – with being able to skim his book for free at the top of my list of course. I emphasize skim because it’s highly, highly unlikely I’ll read the entire book on my laptop – if I want to read the whole thing, I will probably buy the “real thing”.

From a promotion standpoint, each of the 20 people Dave allowed to exclusively offer the link naturally wrote a blog post the book, which is a great way to generate… wait for it… word of mouth marketing the The Word of Mouth Manual. Seth Godin’s (What Dave just did) was my favorite of all the related posts, but many of them are quite interesting… and the “side” benefit of good publicity for BzzAgent might be is most likely worth more than all of the potential book royalties anyway.

From a publishing standpoint, the fact that it’s exceedingly unlikely any real publisher would allow this approach to happen is but one of the reasons he opted not to have one – you can check out the others in his great June 13th post “The Book Distribution Experiment”. My favorite quote:

My first book, Grapevine, was published in 2005… by a major publisher and I saw the trouble with the publishing industry machinery first hand. For example, when my media company polled 15,000 BzzAgents about which cover we should choose, the publisher opted to go with the lowest ranked choice. Their reason? Because the people polled didn’t have “100 years of collective book experience.” Ignoring the opinions of real consumers so blatantly is one reason their marketplace is starting to crack.

We see that all the time around here – the mentality that it’s OK to ask customers what they want, and give it to them, as long as what they want is what you wanted to give them in the first place, otherwise screw them because they don’t know what they’re talking about. Seriously – is there a worse idea out there than engaging a community to get an opinion about a book called Grapevine and then doing the exact opposite of what they said?

So check out Dave’s book (the price is right!) - and if you really want to learn more about word of mouth marketing on the web, check out  each of the 20 blog posts and their associated comments…

Setting aside the fact that I have a slightly more positive view of the long tail, it’s an interesting idea – any “creator” simply needs to find their niche that will buy any and everything that they make. It appears that 1,000 was chosen just to make the numbers simple – i.e. make the point that at $100 a pop that grosses to $100,000 a year, which most people would consider a great living.

The post then goes on to put in a number of caveats and distinctions – groups will obviously need more fans than singles, it will be different by media, etc. It’s a great post from start to finish, and what makes it even better is the two follow-ups that have recently been posted on the site: The reality of depending on true fans and The case against 1,000 true fans.

In the first, Kevin opens by noting that the “1,000 true fans” idea was basically just a hypothesis – there was no hard data/findings behind it. In turn, he started digging around to see if is hypothesis played out in the real world – and this post is primarily a thoughtful response from musician Robert Rich, who agrees with Kevin’s thesis but tosses a little more “realism” on the topic. My favorite quote:

I remember telling myself when I was about 15, “If I can move one person deeply, that’s better than entertaining thousands of people but leaving nothing meaningful behind.” That’s the long tail talking. I suppose when you multiply this idea by a thousand, you have your thesis.

The most interesting quote in relation to the hypothesis:

Thanks to the internet, I am making more money now, selling directly to 1000 True Fans, than I was during the days on Hearts of Space selling 20,000 – 50,000 copies. But had I not benefitted from the immense promotional effort that it took for HOS to sell those albums, I probably wouldn’t be surviving today as a full time artist.

In truth, I could have copied about 7 or 8 different quotes above from Rich – his thoughts are well worth the read for anyone interested in this subject. Search for “Catch 22″ in the text to find a very interesting paragraph on the “True Fans Trap.”

The second post is the case against true fans – and here Kevin notes that while the hypothesis sounds good, and is similar to what people like Brian Austin Whitney postulated a few years back, he can’t find much evidence of anyone actually supporting themselves with 1,ooo (or 5,000, to match Whitney) true fans. In turn, he concludes:

What my research tells me: there are very few artists making their entire living selling directly to True Fans. The few that are, are selling high-priced goods, like paintings, rather than low-priced goods like CDs. But there are many that partially fund their livelihood with direct True Fans.

He then goes on to discuss a challenge from Jaron Lanier, a musician doing research in a similar space. What they are looking for is a new musician (i.e. not one leveraging popularity developed in the “old” model) that’s making a decent living (more or less sufficient to raise a child) from sources friendly to a world of open, massive, unregulated file sharing (thinks concerts, ad sales, merchandise – any thing that doesn’t rely on “old, declining media.” So far, they haven’t found anyone yet – if you know of one drop them a note.

It’s quite an interesting discussion, and one that I hope KK continues on his ever thought provoking blog. Do wikinomics readers have any thoughts on the 1,000 true fans debate, or perhaps the very notion that the long tail is something that creators must try to “escape”?

“…the application of the long-tail concept to language runs into problems perhaps similar to other attempts to apply economic analysis to languages (as) people don’t move “down the tail” to niche markets with language in the way they might with music or books ….. With language, the most prominent fact is that people live in the long tail, as it were, and there are some incentives to move up the tail to dominant languages. Part of the issue is how new technologies facilitate not abandoning the linguistic home in the long tail when dominant languages are learned and used.”

As Paul highlights in his post there are several tools and applications that, in theory, faciliate learning, or given Don’s take, not leaving, the long-tail. But in a world where the language of economics is conducted in one, perhaps two, and in the future maybe three languages, can a combination of technology, ethno-nationalism and culture trump trade and economics?

I highly doubt it. Ethnologists believe that over 500 languages are now considered to be near extiction, and ultimately there’s a reason for it. As countries migrate through the demographic transition, and subsequently become increasingly urbanized, there’s an inherent move towards common languages in order to faciliate the trade of services and goods. And so while there will always be cases such as the  recent decision in France that highlights how nationalism will periodically ensure linguistic staying-power, this still only applies to major global languages who compete to be in the top 10. Those much further down the long tail, like for example the Breton-speaking French, will continue to see their numbers diminish, with technology perhaps serving only as a tool to ensure their capture for the sake of linguistic history.

Does this spell doom for the other 7000 or so languages that are in use today? As I have written before, machine translation is one solution. However, this is imperfect even between similar languages such as English and German. Accurate translation between the different language families is still a distant dream. Another way of survival for less used languages would be for more people learn them. Here the internet can help by connecting students with native speakers and language teachers around the world. While it is known that human interaction is the best way to learn a language, getting a good Hausa teacher in your city might be a challenge. Virtual worlds and large online multiplayer games already facilitate spoken communication using VoIP in an interactive setting. A more formal teaching tool is being developed by a startup called Myngle. The company is developing a platform that will use a combination of VoIP, collaborations spaces and podcasts to allow students and teachers to self organize around learning different languages. The goal is to allow anyone to learn any language. While the unifying forces will likely continue to consolidate the languages used to communicate online, the long tail of language learning will at least give the others an opportunity to be exposed to the world.

The ‘free’ business model topic is discussed by Chris Anderson (author of The Long Tail) on the latest edition of the GigaOm Show, where he speaks about certain topics covered in his upcoming book entitled “Free: The past and future of a radical price”.

It has become obvious that in order to have a chance at becoming big on the internet these days, you’ll need to be free in one form or another. There is very little reason not to offer at least one service for free since technology has made startup costs low and maintenance costs such as bandwidth practically free. Taking a glance at Web 2.0 heartthrobs such as Facebook, Digg, YouTube and Flickr, two distinct business models have surfaced and Chris describes them as: Give away the product and sell the audience (ad supported like Digg, Facebook and YouTube) and give away to 99% of the customers to get 1% to pay (premium service such as a pro account on Flickr).

The best part about Chris Anderson’s upcoming book is that it too will be available for… that’s right… free. The book will be available for free as an audio book in MP3 format, in an eBook format as he believes those who own eBooks are influential, in a typical online format, and surprisingly he is working on getting his book ad supported and ideally it too will be free. Personally, I think that last one is going to be a tough sell to the publishers.

Now what would be the reasoning behind a free book? Well, he argues when you give the book away for free, the book becomes a marketing vessel for the author. More people will read the book and become aware of the author Chris Anderson. Chris would essentially be building his personal brand. By doing so, he will see a spike in demand for his other ‘products’ such as speaking engagements, which unlike downloading an MP3 has become more expensive due to travel costs. The economics behind this one is simple, there is only one Chris Anderson, so supply remains the same, but demand will increase significantly, which makes the scarce resource of his time more valuable.

Click the image below to view the short interview with Chris Anderson, taken from The GigaOm Show Episode 10 – TechCrunch40.

There are also other great interviews you can see in the full version of the show that includes interviews with Marc Andreessen, David Sacks and the guys from Engadget. This isn’t the first time I’ve mentioned this show – I’m a big fan of its an awesome look into the business side of Web 2.0.

It’s worth reading – they’ve got some pretty interesting opinions and insights into the long tail, and a lot of fascinating numbers around TV usage, AA ratings of top TV shows relative to number of channels available, the long-term fall of the Big 3 networks, and many others.

Their survey also has some juicy tidbits – for example, an online video survey found that for “all respondents” user generated content (UGC) lagged movie trailers in terms of popularity, but came out just ahead of music videos and news… but if you look at only males 18-34 (any advertisers interested in them?), UGC blows everything else out of the water.