The Long Tail 4 -The Three Forces of the Long Tail

    This is a part of my critical reader’s companion to The Long Tail and discusses "Chapter 4 – The Three Forces of The Long Tail". Part 0 is here. The previous part in this series is here.

Chapter 4 is a really short chapter (6 pages, including 3 diagrams and a table) which sketches the three forces of the Long Tail (and also the six themes of the Long Tail age – what is it with numbers?) Subsequent chapters talk about each force in turn. So this is a really short post as well.

Given that it is short, this might be a good place to put the canonical picture of the Long Tail. So here it is.

The idea is that you arrange a set of products in decreasing order of popularity along the horizontal axis, and plot their popularity up the vertical axis. The first few are the most popular items (the hits) and as you go to the right you go past many individually less and less popular items. There are a lot of items in the yellow area, and while there is not much demand for any individual one of them, it is possible for some shapes of graph that there is more total demand (a bigger area under the graph) in the tail (yellow area) than in the head (green area). The Long Tail idea is that the Internet is pushing demand away from the hits in the green area and into the yellow area of the Long Tail.

Before goods can be put on shelves, they have to be made, and so Anderson identifies the first force of the Long Tail as "democratizing the tools of production. The best example of this is the personal computer, which has put everything from the printing press to the film and music studios in the hands of anyone." [54] In this definition Anderson fails to distinguish two separate aspects to cost. One is the cost of making the first copy of something – be it a tube of toothpaste or a blog post like this – and the second is the marginal cost of making subsequent copies. It may cost a few hundred or a few thousand dollars for a student to make a short film while it can cost over $100 million for a major movie studio to make a big production. But making the second copy of each one costs the same. Anderson focuses, here and in the next chapter, on those technologies that make it easier for many people to make cheap films (and so on) – hence that loaded word "democratization". But price is not determined by that cost, it is determined by marginal cost – the cost of that second (or thousandth, or millionth) copy. It cost me about as much to watch Jim Jarmusch’s shoestring-budget Coffee and Cigarettes as it cost me to watch Return of the King. We’ll see in the next chapter that this overlooked distinction makes a big difference to the story of production in the digital world.

The second force is "cutting the costs of consumption by democratizing distribution" [55]. It’s the part of the story that Chapter 1 focused on – how companies such as Amazon and Netflix can exploit the Internet to more effectively distribute goods. The Internet, he argues "makes everyone a distributor" [55]. But it takes more than a computer to get a little homemade video seen by hundreds of thousands of people, it takes You Tube. It takes more than an Internet connection to buy and sell an obscure piece of jewellery, it takes eBay. Not everyone is a distributor – in fact the economics of the Internet is likely to decrease the number of distributors rather than increase them. Chapter 6 looks at "democratizing distribution".

The third and final force is "connecting supply and demand" [55] – those recommendations, links, and so on that help us to find things we like on the Internet. Those recommendations, perhaps most effectively on eBay where trust is such an issue, have helped Internet ventures overcome some big obstacles. But when he says "The other thing that happens when consumers talk amongst themselves is that they discover that, collectively, their tastes are far more diverse than the marketing plans being fired at them suggest" [56-57] we can see that there is a big hole right ahead of him on the path Anderson is walking. People have, after all, talked amongst themselves for a long time. Does the Internet increase the level of consumer talk and so increase its usefulness as a mechanism for identifying niche tastes, or does it simply replace one forum for talking (face to face) with another (Internet-mediated)? Chapter 7 looks at tastemakers  on the Internet, where we will see if Anderson falls into the hole or not.

So that’s two short posts, pretty light on substance. Tomorrow I should have a longer one with a bit more to it.

The Long Tail 3 – A Short History of the Long Tail

Back from doing some actual work for a few days, this is another part of my critical reader’s companion to The Long Tail and discusses "Chapter 3 – A Short History of the Long Tail". Part 0 is here. The previous part in this series is here.

Chapter 3 is a complement to Chapter 2 and completes Part One of the book (although it’s not labelled as such). While Chapter 2 set out to chart the rise and fall of the blockbuster, Chapter 3 charts the rise and rise of the Long Tail in a short ten page trip from the Sears Catalogue the web and beyond. It’s a short chapter, and a digression from the main thread of the book, so I don’t have a whole lot to say about it. But I said I’d go through chapter by chapter so I will.

The first section describes the rise of the Sears Wish Book and the catalogue retailing business. The catalogue allowed people – especially those in small towns and rural areas (as piefuchs pointed out in comments on the previous part) – to choose from a much wider range of goods than ever before (200,000 in the 1897 Wish Book [43]) because the catalogue replaced physical shelves in physical stores. "With the heavy thunk of a single mail drop, the choice of available products increased a thousandfold from the typical inventory at the general store". Sears was able to do this because, like Wal-Mart many years later, it introduced new efficiencies into its management of the supply chain. Big warehouses in Chicago together with "virtual" warehouses that were its network of suppliers meant that they could stock items efficiently. It was a big change. The Sears Catalogue has several parallels to the online world that Anderson calls "the ultimate catalog". One that he does not mention is that it was not easily mimicked. Once there is a big catalogue, a second big catalogue doesn’t help much. The cost of keeping an item on the virtual shelf — whether that be a catalogue or a web server — is small, but the cost of building a virtual shelf in the first place is actually pretty big, and once someone’s built a good one it is difficult to compete with them.

Feeding the Tail [44-46] describes how supermarkets did the same thing as the Sears Catalogue had done years before – provided "greater variety, lower prices, and one-stop shopping" [45]: basically they exploited economies of scale. In fact, Anderson tells us the supermarket took off during the 1950s and 1960s: those years of "lockstep culture"[29] described in Chapter 2 did after all have another side to them and that side was variety – in the form of thousands of items on the shelves of the supermarket. It is worth reflecting on these parallel stories that Anderson is telling us. To characterize the 1950s as an age of mass culture and then to describe the rise of a culture of variety at the same time does raise questions. It is often said that the plural of anecdote is not data, and this chapter reminds us why. There are many stories in this world, stories that are true and which can illustrate any message you want to deliver. But stories by themselves are not enough to support broad conclusions. The chapter also reminds us that there are other sides to our current day culture in addition to the availability of lots of online books. We’ll see more of what those are in later chapters, but Wal-Mart is, as pointed out in the discussion of Chapter 1, just as much a creature of the modern technological world as Amazon.

The Touchstone Customer [45-46] is a digression on how new technology — toll-free calling and credit cards — led to a resurgence in catalogue shopping, again by making it easier for sellers to present a wide selection of items to their customers, but it is The Ultimate Catalog [47-49] that brings to the real point of the chapter, describing Jeff Bezos building the Amazon enterprise. The discussion shows how the book market has some properties (a number of products too large for even the bulkiest paper catalogue, a mature wholesaling and distribution system, a product that doesn’t decay in the warehouse) that made it the ideal place for Internet commerce to start. There are, it suggests, particular things about books that make them special when it comes to online selling. To the extent that is true, the success of Amazon may not indicate "the future of business". The success of Amazon is real, of course, although its profitability has been erratic. It’s an interesting story. But if a book is put on a web page, does that make it a product? Given Anderson’s failure in the early chapters to show a really significant shift of purchasing patterns between the offline and online worlds to a new Long Tail world, the story is interesting, but what lessons we learn from it are not so clear.

The final section of the chapter, Long Tails Everywhere [50-51] is, like the Tails Everywhere section of the Introduction, a bizarre leap that stretches the idea of the Long Tail to breaking point and beyond. Getting carried away, Anderson writes (without further justification or description) of open source projects as "the Long Tail of programming talent", offshoring as the Long Tail of Labor, microbreweries as the Long Tail of beer, customized clothing as the Long Tail of fashion, online universities as the Long Tail of education, internet porn as the Long Tail of pornography, and al Qaeda as the Long Tail of warfare. Superficially attractive, such examples have no unifying mechanism underlying them and often little justification in terms of actual change. Was the Spanish Civil War any less Long Tail, whatever that means in this context, than 9/11? If offshoring is the Long Tail of labor are suburbs the Long Tail of cities? As for the resurgence of microbreweries, my explanation of that is here, and it has nothing to do with the cost of shelf space. Anderson provides no explanation of how microbreweries fit into his ideas, what shelves are longer and why. I know from experience that when you work on a project for a long time it is tempting to see the whole world through the lens of the ideas you are working on, but in this section Anderson needed an editor to haul him back from the edge.

Chapter 2, particularly in its final pages, opens questions about what Anderson really means by The Long Tail. He gives several different definitions throughout the book from narrow to broad, but, getting ahead of ourselves for a minute, perhaps the best summary sentence is that at the beginning of Chapter 4, where he says that the Long Tail is "shifting away from a focus on a relatively small number of hits (mainstream products and markets) at the head of the demand curve, and moving toward a huge number of niches in the tail." One thing about this definition is that it identifies the Long Tail as an economic phenomenon – it’s about supply and demand, buying and selling – and as a shift – a change from a previous narrower market to some newer market. Interpreting the Long Tail in a non-economic and static sense as "variety" or "large number of t
hings" (or "The Long Tail is nothing more than infinite choice" [180]) makes it so broad as to lose any useful meaning. Casual soccer games are not part of the Long Tail of Soccer just because there are a lot of them being played in different places; my desk is not a Long Tail of Paper just because it’s a mess of many different pieces of paper; and saying that "offshoring taps the Long Tail of labor" [50] adds nothing new or useful to discussions of changes in the patterns of international labour.

By the end of this chapter, Anderson considers that he has shown that the Long Tail is a real phenomenon and has both a history and a future. He’s ready to move on to the mechanisms that give it shape in the next chapter. The thing about a book is, if you disagree with the setup (and as I hope I have shown in these postings, there are enough weak points in Anderson’s description to make the reality and significance of the Long Tail highly questionable) it’s difficult to go on. Discussion of mechanisms assumes a real phenomenon. It will be challenging to write about that without becoming repetitive.

The Long Tail 2 – The Rise and Fall of the Hit

This is part of my critical reader’s companion to The Long Tail and discusses "Chapter 2 – The Rise and Fall of the Hit". Part 0 is here. The previous part in this series is here.

A book that paints a big picture, as The Long Tail does, and especially if it is paints using an informal, storytelling brush, can be difficult to refute. Point out the problems with any one story and you can be accused of splitting hairs — that one story is just a little piece of the picture, and if it’s a bit inaccurate then it doesn’t affect the basic thesis. On the other hand, if you make big claims and say that "it’s all wrong" you lose as well, because that’s unconvincing. So if you really want to refute a book (and yes, reader, that is what I want to do) then you’ve got to go through it page by page, example by example, and make your case.

A big picture really does have to be backed up by real data and real mechanisms somewhere, and the onus is on the author to provide that data and those mechanisms. Anderson fails to do so in The Long Tail. If you construct a theory (and Anderson calls the Long Tail a theory, even though I’d argue it’s not) and that theory is to hold water, then sooner or later there has to be something solid to it.

By Chapter 2 I was beginning to wonder where the solid foundation for The Long Tail could be found. We’ve seen how, in the Introduction and in Chapter 1, the book flits between engaging stories (Touching the Void, the 98 percent rule), ad hoc comparisons that hide as much as they reveal (Rhapsody vs. Wal-Mart, the coming-of-age teenage choice comparison) and generalizations. And so goes Chapter 2. It’s not that it’s completely wrong, it’s that enough of it is inaccurate or misleading to make the big picture melt away like a snowman in the sun.

The main argument of Chapter 2 is that, as the final sentence tells us "we are turning from a mass market back into a niche nation, defined not by geography but by interests" [40]. Yet the content of the chapter shows only that the past contained a mass market culture (among other things) and the present contains niches (among other things). This is not the same at all. You cannot demonstrate a transition by listing all those aspects of the 1950’s that were mass market and all those aspects of today that are niche – you have to ask if there were niches back then and if there are mass markets today. And that’s where Anderson fails. He just lists examples of mass-market 1950s things and 2007 niche things. It’s not enough.

So, now to the page-by-page look at the chapter.

The first few pages of the chapter are a caricature of a cultural history of the USA: of how culture has changed from a predominantly local agrarian structure (pre-industrial revolution) through the growth of urban "mass culture" in the 20th century, to the "end of the hit parade" as the twenty-first century opens. It describes the 1950’s as "the ultimate in lockstep culture" [29] – an image of grey, authoritarian conformism — and there were indeed aspects of the culture that showed such a shared and "lockstep" trend – "by 1954, an astounding 74 percent of TV households were watching I Love Lucy every Sunday night".

Even here we should remember that in 1954 TV was in its infancy – a third of households did not have one. And an accurate picture of the 1950s would include the fact there was another aspect of society that declined from the 1970’s on, and that was the "collapse of American community" documented so well by Robert Putnam in Bowling Alone. A whole set of diverse community activities was available in the 1950’s that is not present now, and Anderson does not mention this. Putnam "draws on evidence including nearly 500,000 interviews over the last quarter century to show that we sign fewer petitions, belong to fewer organizations that meet, know our neighbors less, meet with friends less frequently, and even socialize with our families less often." Anderson says that the TV data show that "watercooler effect, the phrase describing the buzz in the office around a shared cultural event"  peaked in the 1950s and 1960s, because "it was a safe assumption that nearly everyone in your office had watched the same thing the previous night". But this neglects all those other social activities documented by Putnam, which were more prevalent at that time and which were also markers of diversity.

So a more accurate picture might be that there was the growth of one form of mass culture (TV) at a time when other variegated activities were still present. I don’t know how the arithmetic works out (and there’s an apples vs. oranges quality to it of course) but the onus here is on Anderson to prove his case, and he is being selective in his evidence. It is easy to point at the 1950’s as lockstep conformism if you focus only on those aspects of culture that were common and if you neglect all those that were diverse. And this is what he does.

Even as he goes on into the world of music [30] he draws the same caricature-like portrait. He says how "The [hit-making] machine [that was radio] hit its peak in the form of American Top 40, a syndicated weekly radio show started by Casey Kasem in 1970. It began as a three-hour program that counted down the topforty songs on Billboard’s Hot 100 singles chart… For a generation of kids who grew up in the seventies and eighties, this was the carrier signal of pop culture."[30]

Now I grew up in the UK, not the US, so maybe things were a little different, but (as I’ve already touched on in the previous chapter) things were not this monolithic. We too had out Sunday evening Top Countdown (hosted by Alan "Fluff" Freeman, pop-pickers) and my brothers and I listened to it regularly, as did a lot of other people. But was it "the carrier signal of pop culture"? No. There are two carrier signals for pop culture – the media and word of mouth. The cool kids were all listening to John Peel from 10 to midnight or to Radio Luxembourg and the rest of us listened to the charts and then got the pointers from our cooler friends as to what’s the newest on the alternative side of the music spectrum. So whether you were a Genesis lover or a Black Sabbath follower or whatever, even in those early 1970s there was a variety of tastes around. Anderson is suffering (have I said this often enough?) from a convenient selective vision when he paints his picture of the past.

The End of the Hit Parade [31-32] charts the turn away from mass culture. It is built around a graph that shows the number of hit albums (gold, platinum, multiplatinum and diamond) in each year. The graph shows shows a fairly steady rise from 1957 to 2002 followed by a rapid fall-off after 2002. This turn is explained as follows: "Between 2001 and 2005, the music industry’s total sales fell by a quarter. But the number of hit albums fell by nearly half. In 2000, the top five albums… sold a combined 38 million copies. In 2005, the top five sold just half that; only 19.7 million copies. In other words, although the music industry is hurting, the hit-making side of it is hurting more. Customers have shifted to less mainstream fare, fragmenting to a thousand different subgenres." [32-33]

Like so many other figures and paragraphs in the book, this sounds convincing but does not hold up to closer inspection. Anderson is right that "so
mething happened" in the first years of the century, and that the something was, as he says, digital music. One consequence of the shift to digital, we all know, was a move away from buying albums and towards buying individual songs. Here is a paragraph from today’s (February 18, 2007) Sunday Toronto Star:

"Individual tracks are clearly driving the business," was the Entertainment Marketing Letter‘s assessment of the music industry’s health at the end of 2006. "Album sales were down by 4.9% in both digital and physical formats to 588 million from 619 million in 2005, while 22 tracks were downloaded more than 1 million times each versus two the previous year."

By showing only albums on his chart, Anderson cannot distinguish between two effects of digital music. The first is a shift to individual songs from albums, and the second, to be proven, is the shift of purchases towards a long tail of niche markets. He claims the change is all about The Long Tail, but the quotation above makes it clear that the shift away from albums may have a big effect too. I can’t say which one is the more important, except to note that the evidence from comparing HMV and Rhapsody shows little "shift to the tail". It is a topic that Anderson returns to later, and I hope present more evidence when I get to those chapters that the Long Tail effect appears to be muted, at least. And anyway, to be frank, the onus is on Anderson to prove his case and yet again, by his choice of comparison, he has failed to do so.

Who Killed the Hit Album? [33-35] is another section about trends in music; this is a digression on the factors behind the shift to the Long Tail, a shift which is yet to be proven. It is a descriptive section with few data except for music sharing among music fans (which do seem very diverse) and a glimpse into the subgenres that Rhapsody promotes. The remainder is a plausibility argument about people having many different tunes on their iPods. There’s not a whole lot to be said about this couple of pages.

Broadcast Blues [35-38] moves the discussion from music to the real problems of radio, and thence on to other areas of culture. It claims to show that "What’s happening in music is paralleled in practically every other sector of mass media and entertainment. Consider these statistics from 2005:

  • Hollywood box office fell by 7 percent…
  • Newspaper readership … fell by 3 percent…
  • Magazine newsstand sales are at their lowest levels… in more than thirty years.
  • Network TV ratings continue to fall."[37]

But these figures do not parallel the shift away from hits that he claimed to have seen in music (albeit without good foundation, as I have shown). The drop in newspaper readership and magazine newsstand sales do not mark a shift from a hit parade to a niche market, although they do mark a move away from physical media to the Internet. The two are not the same: if I stopped buying This Magazine and instead got my social commentary from CNN Online I would be moving towards, not away from, a mass culture. It is clear, to repeat, that the Internet is important and is changing things. What is not clear is whether the shift from hits to the Long Tail is real, or in which cases it is real, and what the magnitude and importance of such a shift (if it exists) would be.

The final section is A Hit-Drive Economy is a Hit-Driven Culture [38-40]. It argues that the structure of current media corporations is built around hits. Drawing on the well known fact that predicting hits is difficult (or, as Grant and Wood say "nobody knows") it argues that "existing media and entertainment industries are still oriented around finding, funding, and creating blockbusters" [38] with bad effects on the quality of the end product (a favouring of safe sequels over risky products) – a claim I have no argument with. As Anderson says, this fixation on hits has "leaked outside of the Hollywood boardrooms into our national culture" [39] (ours too, in Canada). It’s not that I want Anderson to be wrong – I’d love a world of "groups bound together more by affinity and shared interests than by default broadcast schedules" [40] and I’d love it is we were "turning from a mass market back into a niche nation". It’s just that wishing it to be so doesn’t make it so, and it is far from clear that the operations of, for example, Rupert Murdoch’s News Corp. (owner of the social networking site will take us there.

At the end of this chapter, are we any further ahead than at the beginning? Not really. We have seen that there is a real shift from physical to digital and that some shared "blockbuster" experiences of earlier decades (network TV in particular) have become diluted. But we have not seen evidence of a real shift away from a mass culture to a niche culture, partly because Anderson ignores those aspects of times past that were richest in niches (civic engagement, local organizations, and so on as documented in Bowling Alone) and partly because his data (musically with album charts) do not demonstrate that people are really moving to the niches as they move online. It may be so, and there are forces pushing us in that direction, but as we shall see in later posts there are forces pushing us back the other way too, towards a hit-based, celebrity-based culture. Which one wins in what cases is up for grabs.

Postings will probably slow until the weekend, when I can spend more time on this project again.

The Long Tail 1 – The Long Tail

This is part of my critical reader’s companion to The Long Tail and discusses "Chapter 1 – The Long Tail". Part 0 is here. The previous part in this series is  here.

Chapter 1 starts with the story of the mountaineering book Touching the Void being lifted from obscurity a decade after its publication by online recommendations at It’s a nice story, and while I haven’t read the book, the film of Touching the Void is gripping and moving. But what does this story tell us? "By combining infinite shelf space with real-time information about buying trends and public opinion, they [online booksellers] created the entire Touching the Void phenomenon… Unlimited selection is revealing truths about what consumers want and how they want to get it." [16]

But this is, after all, just a story, and this is not the first time a product has been lifted from obscurity by a sudden word of mouth. In fact, the opening pages of The Tipping Point feature a story that is very similar – the return of Hush Puppies into style after an extended period in the commercial wilderness, as a result of word of mouth in the New York club scene and then fashion scene. And the Hush Puppy story has nothing to do with infinite shelf space – it may have more to do with second hand and unfashionable clothes stores where the first few people could get their Hush Puppies, before they became trendy again. It is tempting to start thinking "yes, but maybe those tiny stores represent a Long Tail of fashion" but that would be to miss the point of the book, which is that the phenomenon is new and coming from digital technologies. To interpret the Long Tail as "variety" leaves us with nothing but an appreciation of the diversity of the physical world (whose tyranny the infinite digital shelf space will release us from). The world does have a lot of variety, but that is not exactly a new observation. So as in the story about 20 kinds of flour in his introduction, we have to be careful about drawing lessons from Touching the Void, however fine the film is.

The Tyranny of Locality [17-18] is the beginning of this first outline of what The Long Tail actually means. And it is clear that there may be something here. There are limits to much retailing as a consequence of "the need to find local audiences". Geographically dispersed individuals with a taste for something non-mainstream (such as Bollywood movies in the USA) may find their tastes unmet because there are not enough of them in any one place to make it commercially viable to sell to them. The online world allows them to be sold without regard for geography (within a single country anyway, because of shipping) and so may allow more such non-mainstream products to find an audience. Another reason that geography is so limiting is that in the physical world there is "not enough shelf space" to carry everything; not enough screens to show all movies, not enough channels to broadcast all the TV shows [18]. The digital world promises to free us from this: "This is the world of scarcity. Now, with online distribution and retail, we are entering a world of abundance. The differences are profound."

This, then, is the beginnings of the hypothesis. Let’s think about two things that might also be going on, not discussed here. One is that there has been a long tradition of getting round some of the limits of the physical world. Anderson quotes The Triplets of Belleville as a movie that was critically acclaimed and yet "opened on just six screens nationwide" [17]. Yet the movie did appear right here in the small city where I live, in physical space, on the local arts cinema. So here is a mechanism (the smaller independent cinema) for diluting the tyranny of space, and it is a mechanism that may actually be threatened by the onset of digital distribution. Or may not – some have struggled and others prospered. The existence of independent arts cinemas in the physical world is a reminder that that world is not as homogeneous as Anderson portrays it.

The second phenomenon is to do with fixed against marginal costs. I’m not going to say much about this here (we’ll come back to it in a few chapters) but there are things about digital distribution that are different from the physical world in addition to the length of the shelves. Will these other things mess up the Long Tail hypothesis? We’ll see. (hint, yes to a degree, although not entirely). For now, again, let’s just remind ourselves as we read that this is a hypothesis, not a theory.

Markets Without End [19-24] talks more about the world of abundance, painting a picture of the long shelf of the new digital world against that of the old constrained physical world. Anderson looks at the Rhapsody online music retail service owned by RealNetworks, whose CEO blurbed the book and compares it to Wal-Mart’s CD offerings.

Wal-Mart carries about 4,500 unique CD titles, but Anderson shows by looking at Rhapsody’s download data that demand for titles outside that number continues out to at least the 800,000th title (downloaded a few times). "Individually, none of those songs is popular, but there are just so many of them that collectively they represent a substantial market… From the perspective of a store like Wal-Mart, the music industry stops at less than 60,000 tracks. However, for online retailers like Rhapsody the market is seemingly never-ending." [22]

The comparison is not valid, for several reasons.

First, let’s think about using Wal-Mart as a "physical shelf" basis for comparison. It’s not a specialist store, and although it sells a lot of music by volume it does so by doing just what it does in other areas — picks the top sellers and limits its stock to just them. Wal-Mart sells far fewer titles than a specialist CD store, just as it sells fewer book titles than a specialist bookstore and fewer toys than a specialist toy store. But it still drove Toys-R-Us to the brink (at the last time I checked) of bankruptcy. If you look at specialist stores the comparison is not so stark and may even go the other way. For example, HMV in the UK reported in 2004 (PDF file – see page 5) that "Whilst the 200 best-selling CD albums account for approximately 56% of the market, these titles in HMV UK are approximately 40% of sales. Similar ratios apply to DVD. So, with approximately 60% of HMV UK’s sales coming from the back-catalogue, the resilience of our model was proven during a year in which it faced the sternest of competitive tests." So the physical world consists of those like Wal-Mart who cherry pick the hits (shortest shelves) followed by the big chain stores like HMV (longer shelves) followed by specialist stores that stock just classical music for example, whose shelves may be short but whose stock complements that of HMV and Wal-Mart. To compare Wal-Mart against Rhapsody is misleading as a portrayal of the physical vs. digital worlds. In fact, numbers from a more recent weblog posting by Anderson shows that Rhapsody makes approximately 48-50% of its sales from the top 200 albums (it shows data for 100 and 1,000 so this is a rough interpolation), compared to HMV’s 40%. Rhapsody, by this measure at least (and there are oth
er measures), is more hit driven than HMV.

There’s another problem with choosing Wal-Mart as a basis for comparison. Anderson is looking to compare old models against new (remember that "preview of 21st-century economics" subtitle in the introduction?), but Wal-Mart is just as new as digital technology. Its growth in the US throughout the 1990’s parallels that of the Internet. It’s no old dinosaur being supplanted by new technology, it is just another face of new technology.

We can think of another model for "the Future of Business", which you might call the Long Vice, in which specialist retailers like HMV and (even more) its smaller independent brethren are being squeezed from both sides. On one side is Wal-Mart, using its economies of scale to provide the hits and cherry-pick that revenue from the specialist retailer. On the other side is the online retailer who can pick up some of the smaller selling items. Whether this leaves us with more choice or less (or a bit of both) is up for grabs, and we can talk more about the forces that drive down choice in the modern economy later on when Anderson goes into "Long Tail Economics" in a little more detail. But at least let’s acknowledge that the Long Tail story of old physical Wal-Mart vs new digital Rhapsody is fundamentally oversimplified and fails to make Anderson’s case.

Anderson spends most of this section on Rhapsody, and then does similar comparisons in other industries; Borders vs. Amazon, Blockbuster vs. Netflix. I’m just going to talk about Borders vs. Amazon here.

Having set the stage plausibly (but incorrectly, as we have seen) with the Rhapsody story, it seems to be just confirming evidence for the Long Tail hypothesis when he says that Amazon makes 25% of its business from books "not available in offline retail stores" or "products you can’t find anywhere but online" [23 – diagram]. But there are two problems. First is that misleading phrasing. We have three new-book bookstores in my town. Chapters sells more books, but Words Worth (our independent book store) sells an overlapping set, and A Mysterious Affair sells a different set too (and then there are Christian book stores and probably other ones I don’t know of). So our total number of books "available in offline retail stores" is much bigger than the stock of Chapters. Phrasing the comparison between Borders and Amazon as a comparison between "the offline world" and "the online world" is misleading, just as the Wal-Mart vs Rhapsody comparison was misleading. This is especially so because (again, we’ll come to this) there are reasons for thinking that while the physical world is populated by a variety of stores – even though some smaller ones are being threatened by the Wal-Mart/Borders model and (ironically) by Amazon and Netflix — the online world has forces that encourage natural monopolies (quick, name a pure-play new-book online retailer that isn’t Amazon) or at least oligopolies. At the beginning of this post I hypothesized about the role of small out-of-the-way stores in the revitalization of Hush Puppies. It is possible that the new world may have less variety in it than the old.

The second problem with the story has been pointed out elsewhere and discussed (with honesty, I should add) by Anderson here. He says that "One of the most quoted statistics in my original article was the data point that 57% of Amazon’s book sales are in the Long Tail, defined as beyond the 100,000 books available in the typical Barnes and Noble superstore". Well, that estimate came from an analysis of Amazon’s proprietary data (it does not publish detailed breakdowns of its sales). A revised estimate, included without comment in the book, is that the Long Tail accounts for 25% of sales. This is a downward revision of over half, and yet it has not affected Anderson’s thesis or his winning enthusiasm for it at all. You have to wonder about a hypothesis that is so little affected by a big change in its underlying data from the real world.

The chapter finishes, as so many of them do in this book, with a ringing call towards a bright future. "as demand shifts towards the niches, the economics of providing them improve further, and so on, creating a positive feedback loop that will transform entire industries — and the culture — for decades to come" [26]. At this stage of the book, we just have to say that the jury is out, but if you are like me, you’ll be feeling increasingly suspicious and frustrated by the slippery way in which Anderson is putting together facts, stories, and comparisons to create a misleading picture that, wouldn’t you know, supports Anderson’s hypothesis. There are forces pushing towards an increased reliance on hits (see Wal-Mart above) as well as pushing out the other way. Some people say that online stores are just as hit-driven as bricks-and-mortar. The burden of proof is on Anderson, especially given the evangelical language he uses to describe what he is talking about. And in this chapter he dramatically fails to make his case.

The Long Tail 0.2 – Introduction

This is part of my critical reader’s companion to The Long Tail and discusses the Introduction. Part 0 is here. The previous part in this series is  here.

After taking two posts to get past the cover, you may be very glad to know I have nothing to say about the dedication page or the acknowledgments. This exercise will be long-winded enough without that. Yes, I’m ploughing on, turning pages recklessly past the Table of Contents and all the way to the Introduction. But I’ll stop there, and look at it closely.

The point of an Introduction is pretty similar to the cover. It’s to set out the main themes of the book, to sketch the argument, and so to draw the reader in. Some writers call it Chapter 1 (that’s what I did) and others don’t call it a chapter (that’s what Anderson does), including it as part of the front matter for the book.

Unsurprisingly then, some of the things I have to say here will repeat what I said about the cover. Anderson starts off by talking about the world we are in now, and this world is "the world the blockbuster built" [1]. (Remember, numbers in square brackets are page numbers in the book, not footnotes) But that world is "starting to tatter at the edges" [2].

I want to spend some time on a paragraph on page 2 where he backs up this claim that blockbusters have had their day. It’s not the only place he does this, and we’ll return to the question of whether blockbusters are "starting to, gasp, rule less" later, but this is the first time he attempts to bolster the claim. He writes that "Most of the top-fifty selling albums of all time were recorded in the seventies and eighties" [2]. (The end notes make it clear he is talking about the USA here. As elsewhere, most of his figures and data refer to the US.) Interesting, but what about other areas of culture? A parallel statement would be that most of the top-grossing movies of all time were in the same time period. But the top grossing film of all time is Titanic (1997) and even adjusted for inflation, the top box office hits are Gone With the Wind, Star Wars, The Sound of Music, ET, The Ten Commandments, and Titanic – one each from the thirties, fifties, sixties, seventies, eighties, and nineties. Among the top 50 are 2 from the 1930’s, 4 from the 1940’s, 5 from the 1950’s, 9 from the 1960’s, 11 from the 1970’s, 7 from the 1980’s, 7 from the 1990’s, and 4 so far from the 2000’s. So there is perhaps a bit of a peak here, but it is not a dramatic one, especially as the figures are biased towards pre-1980 films which have had multiple releases (Disney animations in particular). So movies do not seem to show the same waning blockbuster effect as music. Instead, Anderson points to the waning total theatergoing audience, which seems to me a different phenomenon entirely.

On TV he has a stronger case – the waning of network TV in the US means that "the number one show today wouldn’t have made the top ten in 1970". But what about books, the other area of entertainment/culture that he talks about a lot? He doesn’t say. The bestselling books of all time are books such as the Bible and Quotations from Chairman Mao, which don’t tell us a lot about book-buying habits in the sense he is looking for. A best-selling book can sell for years, decades, or more, giving Beatrix Potter’s Peter Rabbit a head start over Asterix. So although I do feel sceptical about the fall-off of blockbuster books (Da Vinci Code and Harry Potter anyone?) I’ll give him a break here.

Returning to movies for a moment, there are other sides of the movie industry which tell a different story from the one that Anderson is telling. Peter Grant and Chris Wood’s 2004 book Blockbusters and Trade Wars explains how (p 76) the era of the blockbuster started with Jaws. Before then, most movies launched in a "rolling opening" screening in major cities first and smaller centres later. Jaws opened on 460 screens – a big change. By 1995 however, "virtually all of the 150 or so movies released by major studios in the United States and Canada opened in more than 800 screens. By the turn of the century, 2,000- to 3,000-screen openings were commonplace. In 2001, Harry Potter and the Sorcerer’s Stone opened in the United States on a breathtaking 8,200 screens — nearly a quarter of all the theatres in America. On its opening weekend in Canada that year, Goldmember managed to seat an even more astonishing 52 per cent of all admissions". Here is a different story – if blockbusters are waning, it is not because we have more choice but because they are living a shorter time, rolling over ever more urgently to make way for the next big opening. Anderson skips over these other consequences of the increasing returns to scale of digital production.

Anderson’s conclusion, then, is that hits "are not quite the economic force they used to be". More dramatically, he goes on to say "Where are those fickle consumers going instead? No single place. They are scattered to the winds as markets fragment into countless niches." At this stage in the book, the reader can only say "well, we’ll see." He hasn’t proven his case yet, that’s for sure.

iTunes Killed the Radio Star [2-6] contrasts Anderson’s own upbringing in the ’70s and ’80s with a modern teenager. The difference is simple: "The main difference between Ben’s adolecence and my own is simply choice. I was limited to what was broadcast over the airwaves. He’s got the Internet." Now let’s be clear: the Internet is a big thing. It is changing much about this world, including adolescent lives. It’s going to change more. Some of the changes will be profound, some banal, some specialist, some pervasive – and it’s only just starting. It is the multifaceted nature of the changes it is bringing that make comparisons dangerous. When Anderson says "He plays video games with friends, mostly online" he is confirming that one change is that play has shifted from the streets, from friends homes, to the computer screen and networks. So to compare Broadcast TV to "The Internet" is misleading. This is something I’ll return to in later chapters, but what he is describing is not so much an increase in choice as a shift in choices. Some choices (playing in the street) have basically been lost, while others have risen to take their place. Just because these new options travel over the same network as online newspapers does not change the fact that it’s just another form of playing with friends.

What’s more, Anderson’s description of his mass culture youth, hemmed in by the small number of rock stations, and by an inability to explore outside the mainstream, is just not plausible. I mean, I was there. I grew up in the ’70s on the edge of Leeds, a middle-sized provincial city in the UK, and went to school in a town of 13,000 people. This was not a cutting edge environment, but even within that I was behind the times – unfashionable, with a narrow taste in music and a conformist taste in clothes. But even I managed, despite my best efforts, to be exposed to non-mainstream culture. The biggest band of the mid-’70s was Led Zeppelin, who thrived despite never being played on radio. My own particular favourite was Pink Floyd (ditto) but there were minor bands a plenty (Wishbone Ash anyone? Hawkwind?) My friend Matthew was into German techno-rock: Tangerine Dream and Kraftwerk. And that was before punk shook everyone up. So to portray this world as devoid of anything but the mainstream is, frankly, ridiculous.

But let’s leave my teenage years behind quickly and go back to the book. A section called "The 98 Percent Rule" [6-10] describes the beginnings of the book. The rule is based on a conversation he had with the CEO of Ecast, a digital jukebox company. This CEO pointed o
ut that 98% of all their 10,000 albums had sold at least one track per quarter (three-month period). Anderson is stunned at the variety of people’s tastes when unencumbered by the physical limitations of old jukeboxes. This 98 percent rule is that just about anything you can offer for sale will find a buyer, so if you can (in a digital world) offer everything, then you can make money off all those items that sell a few here and there. Hits still matter, but "What if the non-hits–from healthy niche products to outright misses–all together added up to a market as big as, if not bigger than, the hits themselves?" [8].

One reviewer who has challenged this part of the book is Wall Street Journal’s Lee Gomes. Gomes argues (here ) that whereas Anderson claims the 98% rule to be "nearly universal" citing iTunes, Netflix, Amazon and others ("And so it went, from company to company"), Gomes himself found something different:

But when I looked online, I was surprised to see what seemed like the opposite. Ecast says 10% of its songs account for roughly 90% of its streams; monthly data from Rhapsody showed the top 10% songs getting 86% of streams.

Bloglines, the widely used blog-reading tool, lists 1.2 million blogs; real ones, not computer-generated "spam blogs." The top 10% of feeds grab 88% of all subscriptions. And 35% have no current subscribers at all — there’s clearly no 98 Percent Rule in the blogosphere.

At Apple’s iTunes, one person who has seen the data — which Apple doesn’t disclose — said sales "closely track Billboard. It’s a hits business. The data tend to refute ‘The Long Tail.’ "

Other economists, of course, are looking into these same questions, though some seem to be reaching far more restrained conclusions. Harvard’s Anita Elberse, whom Mr. Anderson said was a consultant during his two-year research project, studies the video sales market, both online and off.

She said in an email that her work to date shows a "slight shift" toward the tail. But she also noted "a rapidly increasing number of titles that never, or very rarely, sell," which suggests "it is difficult for content providers to profit from the ‘tail.’ "

It would be wonderful if the world as Mr. Anderson describes it were true: one where "healthy niche products" and even "outright misses" collectively could stand their ground with the culture’s increasingly soulless "hits."

But while every singer-songwriter dreams from his bedroom of making a living off iTunes, few actually do, mostly because so many others have the very same idea. And to the extent that Apple is making money off iTunes, thanks go to Nelly Furtado and other hitmakers. Indeed, you can make the case that the Internet is amplifying the role of hits, even in relation to misses, not diminishing them.

Anderson responds on his weblog but while there is dispute over other portions of the book (and we’ll get to those in a few posts time) he doesn’t challenge this part so far. The 98 percent rule seems to be a rule that is often broken.

Anderson himself replaced the phrasing of "98 percent rule" with "The Long Tail" as he continued his work. The point to make here is that in the introduction, despite some intriguing-sounding stories and some big claims, he has not demonstrated that a "Long Tail effect" actually exists. This matters.

Tails Everywhere [10-11] is one of two places where the book goes seriously off the rails. As we’ll see, for most of the book the meaning of "The Long Tail" is fairly clear, but every now and then Anderson seems to get carried away and treat it as just another word for "variety". So here he says "People often ask me to name some product category that does not lend itself to Long Tail economics. My usual answer is that it would be in some undifferentiated commodity, where variety is not only absent but unwanted. Like for instance, flour… Then I happened to step inside our local Whole Foods grocery and realized how wrong I was: Today the grocery carries more than twenty different types of flour… There is, amazingly enough, already a Long Tail in flour." [11] This passage has two large faults. First, it clearly implies that Anderson sees the logic of the Long Tail, most of which is about media and entertainment industries, as extending to pretty much all of economics — "The Future of Business". Also, he sees it as stretching beyond the digital world. Yet here (and again we’ll come back to this when he explains in more detail what he means by a Long Tail effect) he seems to be saying that the mere presence of a variety of goods is sufficient to indicate a change. You have to wonder, does 37 brands of toothpaste reflect a long tail in toothpaste? And what if (something he doesn’t mention) they all come from two companies? If he is going to argue that Long Tail simply means the emergence of niche markets then he has to demonstrate that a change is happening, whereby niche markets are becoming more important. Once more, at least in his introduction, he does not do this.

A Preview Of Twenty-First-Century Economics [11-13] sets out how Anderson sees his book. First, he sees it as an economics research project (among other things). He’s doing economic research that "very few economists" [12] are doing, and venturing into "uncharted waters". "Some of the data in this book" he says "has never before seen the light of day". Anderson claims his role is to "synthesize the results [of the efforts of online entrepreneurs] into a framework. That is, of course, what economics does: It seeks to find neat, easily understood frameworks that describe real-world phenomena". So this is an economics research book, he says.

Let’s sum up what I’ve seen in the introduction. It has some nice sentences, some big claims, but is already showing at least three potential problems. First, he is so far unable to substantiate his claim that the era of blockbusters is waning and that the "mass-culture era" [2] of the ’70s and ’80s has given way to a world of niche cultures. Second, his 98 percent rule seems dubious. Third, he is claiming to have a book of research that sets the stage for the future of business and twenty-first century economics — he is setting the bar high. Can he clear it? Let’s see…

The Long Tail 0.1 – The Cover

This is part of my critical reader’s companion to The Long Tail and it discusses the cover. Part 0 is here.

I know, I know. Never judge a book by its cover. But books sell themselves based on their cover, even in the digital world, so covers matter. The cover is the author’s chance, and the publisher’s chance, to tell us in a few words and images what this book is all about before we put down our dollars or euros and buy it. At some point we all judge books by their cover: what does The Long Tail‘s cover tell us to expect in this book?

At the top of the cover on the edition I have (probably the Canadian edition) is the phrase "The New Economics of Culture". On some other editions, like the one shown above, the phrase is "How Endless Choice is Creating Unlimited Demand". Either way the message is clear, and it sets the tone. New Economics of Culture, Unlimited Demand, Endless Choice: these are grand phrases proclaiming big ideas about dramatic change.

The title comes next, and a good title it is. The Long Tail. Even if you haven’t read the book, there’s a good chance you’ve heard that phrase, because it has become a catch phrase for many different developments in the digital economy. It’s not quite up there with The Tipping Point, but it’s close. Getting a good title is an important thing. To summarize an argument in a pithy, evocative phrase is not just smart, it’s good communication and good communication is what books are all about. Chris Anderson used the title first a few years before the book, in an article in Wired Magazine, of which he is the editor. I’ll come back to that, because although the article forms the basis for the argument of the book, and much of the article forms chapter 1 of the book, an article and a book are two different things. Still, this is a promising start.

Below the title is a simple graphic. It’s an Enter button from a computer keyboard, which both welcomes us to the book and tells us that this is a book about the digital world.

And then the subtitle: "Why the Future of Business is Selling Less of More". Again, this is a big statement and I’m going to hold Anderson to it. He’s talking about the Future of Business. It’s a big subject. He’s not talking about the future of 15% of business, or the future of some companies in San Francisco, he’s talking about the Future of Business and he’s telling us what it’s going to be, which is Selling Less of More. Diversification, the end of the blockbuster, the promise of variety and choice, the dream of a more democratic, participatory, and diverse future: these are the visions that the phrase conjures up.

At the bottom of the page is a blurb by Eric Schmidt, the CEO of Google no less. "Anderson’s insights influence Google’s strategic thinking in a profound way. READ THIS BRILLIANT AND TIMELY BOOK". Coming from the chief of one of the quintessential new Silicon Valley companies, this is high praise. The blurbs on the back cover continue the tone. They come from other CEO’s – Reed Hastings of Netflix, Rob Glaser of RealNetworks, Terry Semel of Yahoo! – as well as other prominent commentators: James Surowiecki of the New Yorker Magazine, Lawrence Lessig of Stanford Law School and Geoffrey Moore, business author and founder of the Chasm Group. These are visible, prominent people.

It’s interesting to see where the endorsers come from in the light of Anderson’s claim in the book that the new economy eliminates "The Tyranny of Locality" [p17] or "The Tyranny of Geography" [162] or "the tyranny of physical space" [163]. There are eight people on the cover (including Anderson himself). Here is where they come from (look for the red dots):

There’s one dot in New York (that’s James Surowiecki) and one in Seattle (Rob Glaser of RealNetworks). The other six are grouped together, indistinguishably at this scale, in and around San Francisco. Here’s a close-up of the San Francisco Bay area with the six red dots marked:

So looking at the cover does tell us something. It tells us that this is a Silicon Valley book, written by someone who (as editor of Wired Magazine) is well connected with the movers and shakers of the digital world. Not only does he influence them (as Schmidt says) but he listens to them. He knows them and they know him.

There’s an obvious irony here. A book about how the digital world permits small players from the fringes to reach a bigger audience (in ways I’ll discuss later) and about how geography is fading as a force in our world comes to us from the centre of that new world, praised by big names from the high-class-neighbourhood of the techno-culture. It is promoted by and emerging from a network of culturally similar, geographically close (and very wealthy) people: The Long Tail is a reflection of conversations among a new elite. The Long Tail is joined to a head, and the head is still very much the place to be.

In his blurb, the CEO of Yahoo! says that "Technology and the Internet are making the world a smaller and more connected place." Moore argues that the new economy will have "more collaborative, participative, and idiosyncratic offerings… knocking the Old Guard on its backside in the process". Glaser says it’s "the biggest transformation of media since the commercialization of television fifty years ago". The cover tells us that perhaps this quirky and diverse new world  may have aspects to it that are not entirely unlike the old one. Is that simply an irony, or is it a marker of a problem with the book? I’ll argue in what follows that it is a symptom of some things that are wrong with the book itself.

The inside flap continues the extravagant praise for the book. It is compared to The Tipping Point twice. The Long Tail is "an entirely new model for business that is just starting to show its power". And it’s not just the digital world that will be changed: "a similar transformation is coming to just about every industry imaginable." "What happens when everything in the world is available to everyone?" This is setting the bar pretty high for the actual content. Next time, let’s finally get to the book itself.

The Long Tail 0 – Cards on the Table

My next several postings are going to be a critical reader’s companion for The Long Tail by Chris Anderson. And if you’re going to read these postings, there are some things you should know – and that’s what this introductory post is for, to make sure all my cards are on the table.

The posts talk about The Long Tail, but they are really about two books. Both were published in summer 2006 and both were about the changing face of individual choice in today’s economy and culture. If it wasn’t for The Long Tail‘s index being two pages shorter, both would be exactly 240 pages long. Both authors have a training in science, but neither are professional economists or cultural studies professionals. In fact, both work in the computer industry, one as an editor of Wired Magazine, and one for a computer software company. At the time of writing, The Long Tail is at number 310 in the Amazon best seller list. The other book currently stands at 806,127.

Obviously that second author is me. My book is "No One Makes You Shop At Wal-Mart: the surprising deceptions of individual choice" and it was published by Between The Lines of Toronto, a small publisher whose staff love books and love to publish provocative, left-wing titles with an academic slant. The Long Tail is published by Hyperion, of New York. As for the subject and the authors’ take on it, mine is encapsulated in that phrase "the surprising deceptions of individual choice" and Anderson’s in a chapter title "The Paradise of Choice". Couldn’t be more opposite.

So there’s a danger here, because I have a chip on my shoulder. The Long Tail and No One Makes You… cover overlapping territory, and have almost diametrically opposite views. I think that Anderson is basically wrong. Not completely wrong, but basically wrong. What’s more, and what’s more important, I think his book is fundamentally sloppy and misleading despite the fact that there are a few good insights – and he does have a great title. But you, dear reader, knowing that I have the aforementioned chip on my shoulder, might naturally and understandably think this is sour grapes – the envious grumblings of a bit player when a more talented writer has seized the agenda and won the argument.

And maybe it is. I’d like to say not, but who am I to say? Any honest person knows that we are not the best judge of our own motivations. I’m sure as hell not the best judge of mine, and that’s why I wanted to get this off my chest before starting. If you see bias in what I say, you know where it comes from.

But even with that danger of misdirected envy, there are still things I want to say, so all I can do is put down what I think and you can judge it on its merits. One option would be to take a five paragraph run at the book, but its success and its influence have been so significant that it needs more. I’m going to take a chapter-by-chapter run at the thing (with a couple of extra posts as well).

Just a few more preparatory statements before we can start.

First, this is a reading of the book, which is a bit different from the idea behind the book. There are only a few cases where I have gone beyond what is in the book to cite discussion and research that has happened since its publication (or before). A book should stand reasonably well on its own, or point to places for more reading if it leaves discussions open-ended.

Second, there are some things that are important to me in non-fiction books. Consistency is important. So is clarity. And so is logic. If they don’t matter to you – if rhetoric or eloquent phrasing or memorable stories matter more, then you and I are going to have different opinions.

Third, I reference page numbers in the book in square brackets, like [23] this.

I think that’s all. The next post takes a look at the cover (this project may take a while to complete).