Menu
Menu
Long tail theory is, in a word, bollocks

Long tail theory is, in a word, bollocks

The worst thing about really long tails? They are an absolute bugger to skin properly, even with a sharp knife, and they don't have much flesh on them.

Zip-a-dee-doo-dah, zip-a-dee-ay, my oh my, what a wonderful day! Hello, and welcome to yet another edition of This Wonderful Life,

coming to you this week direct from the soon-to-be-closed Rabbit

Skinning and Pan Handling room here at the Digital Life Labs.

We opened the room a little while back, when it became increasingly

apparent that, what with the recession on the one hand, and the

predicted demise of mainstream media on the other, we needed to hone

our depression-era skills to give us something to fall back on. And

when it came to honing new skills, what better place to start than

honing our rabbit-skinning knives?

But over the past week, two separate pieces of analysis have given us

cause to hope. One is the International Monetary Fund's prediction

that the Australian economy will not go backwards as a result of the

blah blah blah.

The other, much more exciting analysis comes from an economist with a

music royalty collection agency in the UK, who has crunched some

numbers and decided that the blogosphere's famed long tail theory is

in fact nothing more than a long (shaggy dog) tale.

Zip-a-dee! No more kangaroo pie for us!

The long tail theory, you may recall, is the one espoused by Wired

magazine editor Chris Anderson in 2004, predicting (among other

things) that mainstream media outlets such as Fairfax (which pays the

Digital Life Labs' bills) would eventually be swamped by millions upon

millions of bloggers, each of whom might have only a tiny market

share, but who collectively would dominate the market, owning more

market share than the big publishers and broadcasters.

Long tail theory also suggests, for example, that fully digital music

stores, no longer constrained by storage problems, would make more

money from selling unpopular tracks than they would from selling hits.

Sure, the new Pink CD might sell (for the sake of easy maths) 1

million copies, but for every Pink there might be 100,000 smaller

artists, each of whom only need sell 11 CDs to become the dominant

segment in the record industry. The title of Anderson's 2006 book

neatly summarises his theory: The Long Tail - Why the Future of

Business is Selling Less of More.

To bloggers and struggling musicians and authors, it's an appealing

theory, and one that appears to have been borne out, at least

anecdotally, by book sales at Amazon.com. A former employee of Amazon

once told Wired that he explained the long tail effect by pointing out

to his fellow employees: "We sold more books today that didn't sell at

all yesterday than we sold today of all the books that did sell

yesterday."

The problem is, however, that the long tail theory just might have a

tiny flaw. It just might be, in mathematical terms, bollocks.

Under long tail theory, sale of digital goods should follow a "Pareto

distribution" curve in which the sum of the sales of all the unpopular

goods exceeds the sum of the sales of the popular goods. In other

words, the skinny tail of the distribution graph, representing the

unpopular goods, goes on and on for so long that, by the time you

reach its end, the tail weighs more than the meaty, popular stuff at

the start of the graph.

But according to Will Page, an economist at the music royalty

collection agency MCPS-PRS Alliance, digital sales don't do this. They

follow a more conventional log normal distribution curve, which barely

has any tail at all. In the log normal curve, the bulk of sales go to

a select few goods, just as they always have, and the bulk of goods

get no sales whatsoever.

According to a report appearing in the technology news site The

Register, Page and some colleagues created a spreadsheet with 1.5

million rows of data, taken from tens of millions of transactions at

an unnamed online music provider.

Not only did 80 per cent of tracks receive no sales at all (the

complete opposite of the long tail theory), but the bulk of sales were

far more concentrated than even the economists predicted. Sales were

so concentrated on just a few songs, in fact, that even to list the

Top 40 songs, as the music industry often does, was misleading: it

should be a Top 14, Page suggested.

If Will Page is right (and for obvious reasons, we hope he is), it's

not just we here in the mainstream media who can breathe a sigh of

relief. Traditional brick-and-mortar retail stores are also off the

hook, because his study suggests that such stores have not been

distorting the public's consumption patterns the way long tail theory

suggests.

Anderson's long tail theory says that, because real-world music or

book stores can only carry so much inventory, the public is forced

into buying only the popular inventory. If the niche stuff attracts

few or no sales, it's not because it's inherently crap, but because it

just isn't on the retail shelves. Obversely, if Pink is popular, or if

everyone is nuts about Harry Potter, then it says more about the

constraints of the retail environment than it does about the public's

tastes.

But, according to Page, even when the niche stuff is available online,

the public still adopts its herd-like mentality and leaves the niche

stuff on the virtual shelf.

It's even possible (and this is our theory, not Page's, so it's

probably just as bogus as Anderson's, but in the opposite direction)

that internet technologies such as social networking and

user-generated recommendation engines could create distribution curves

with a tail that is even shorter than ever before, given the way such

technologies can create online herds that have a size and speed never

seen in the pre-internet era.

We certainly hope so. Because, you know the worst thing about really

long tails? They are an absolute bugger to skin properly, even with a

sharp knife, and they don't have much flesh on them.

Fairfax Business Media

Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

Join the newsletter!

Error: Please check your email address.

Tags new technologiesWeb 2.0

Show Comments