Death Of The Long Tail

[Lazz]

Do you remember “The Long Tail” ?

In a famous 2004 Wired magazine article, Chris Anderson borrowed the old post-war statistical concept of a frequency distribution with a long tail and re-made it in the form of a notion to illuminate internet media business successes. Low operation and inventory costs, according to Anderson, allow significant profit to be realized from the sale of small volumes of hard-to-find items to many customers, rather than large volumes of a narrow range of popular ‘hit’ items.

And the existence of the Long Tail, as a sustaining demographic purchasing a large number of "non-hit" items, thus became an article of faith in the new ‘indie’ ideology

This received wisdom, however, that diverse specialised items – though individually less popular - will together outsell mainstream ‘hits’, has now been profoundly compromised by actual analysis of one year’s music revenue in a study led by Will Page, Chief Economist for the UK’s royalty collections agency the MCPS-PRS Alliance.

Here is the sad grounded truth about the digital universe:

* Of the 1.23 million albums available on-line, only 173,000 were ever bought and paid for – the remaining 85% failed to sell one single copy.

* Of the 13 million individual tracks available by internet sale, only 20% were ‘active’, in that they managed to sell at least one copy.

* More than 10 million tracks failed to find one single buyer.

* Approximately 80% of sales revenue came from around 3% of these active tracks – numbering about 52,000 – the ‘hits’ which powered the industry.

* Only 40 tracks sold more than 100,000 copies, and accounted for 8% of the business.

(In the physical world, 40 tracks could make up just 4 albums).

The Two Main Lessons:

* The general rule for all inventory on the digital shelf is 80/0.38 – that is, 80% of revenue is generated by sales on only 0.38% of all available tracks.

* 80% of individual tracks sell nothing at all.

In their brief report on this research, The Guardian cheekily quoted Sturgeon’s Revelation (they called it ‘Sturgeon’s Law’) that “90% of everything is crap” – which I found pithy, appealing, and potentially perfectly apposite.

Read an interview with Will Page HERE.

[john]

Hey Lazz

Thanks for that piece of happy news!

Not a great surprise but useful to put numbers to the vague notions of before.

Cheers

John

[Steve]

Obviously we need to ditch the 80% that isn't selling and concentrate on the rest! We'll all be millionaires this time next year! Simple innit? Can't think why nobody thought of this before?

[Nightwolf]

in a recent poll it was proven that 60% of people don't believe statistics

Consider also the double dippers. I have a friend at work who uses bit torrent or some such, but if he likes the stuff he downloaded he will try to find it for purchase. Failing that he will spend that money on a show or merch from same artist. He likes the torrents for hard to find products as well as early releases. Often the obscure stuff is only available on the black market as mentioned in the article.

I myself downloaded the complete "Year Zero" NiN cd (from the NiN website), then bought the hard copy when it was released as well.

Am I to understand that they consider 2 years to bee a long tail?

[Lazz]

So far, I notice only one very brief media reference to the study.

I had to search and burrow for further detail.

Could it be that this bad news was intentionally buried by releasing it Christmas week ?

A cynic wonders who would do that ?

And why ?

Neither do I find reference on the MCPS-PRS Alliance website - and it's their research, as far as I know.

Perhaps it's available only to their members.

Has this been reported elsewhere at all ?

Anyone notice ?

"Do they consider 2 years to bee a long tail?"

John - I have no idea - in the context of what exactly - can't see where that fits.

[Nightwolf]

Using this approach our team constructed log normal intervals and plotted an unprecedented amount of digital music data over a significant time period.

based on an unprecedented analysis of digital music sales data gathered over a year

just doesn't seem like a very "long tail" to me I'd like to see a 5 year study at least (give me a chance to finish my album )

[roxhythe]

*I* think they're wrong. I do not question their data, but rather their measuring technique.

It is reminiscent of the famous poll by Literary Digest magazine in 1936, which predicted the Republican candidate for U.S. President, Alf Landon, was going to win by a big margin. LD's poll was nationwide; it was conducted according to the best scientific procedures in existence at the time. It had only one flaw. The poll was conducted by telephone. This was the middle of the Depression. Who had telephones? People who still had money and still had jobs. The people who didn't voted for Roosevelt--in big numbers.

I think the conclusion one is likely to draw from the information presented is that the recording industry has done a very good job of keeping outsiders out. If that makes the recording industry complacent, great. If they die from their complacency, well, they can join the other dinosaurs.

What I didn't see being measured here is how many records are sold *at gigs* by independent, not-very-famous artists. I have a feeling it's quite a few. No, nothing on the scale of the favorite children of the record companies--they have at their disposal what Joni Mitchell called "the Star-Maker Machinery," and it counts for a lot. But I myself have just about gone through four "pressings" of my last album in the last two years (true, they were small pressings)--I have only a handful left. And I have sold nearly every one of 'em at gigs. Next album, I'll see if I can generate some online sales. I don't know if it's possible, but it's a small investment and worth a try. And I'm small potatoes. There are people who are doing a lot more of this a lot better than I am.

My own impression, for what it may be worth, is it's a *very* long tail. It's just also very *thin". Think of a Norway rat. Big body, very long, thin tail. (Considering some of the people who populate the recording industry, a rat is not a bad analogy.) The rat probably does not think much about its tail. If the tail had independent thought, the tail might think more about the rat than the rat thinks about the tail. The rat, folks, is old and not in good health. The question I think *we* need to be asking ourselves is what happens to the tail when the rat dies. I'd like to see the tail survive without the rat, myself--maybe turn into something else rather than growing a new rat.

Joe

[Nick]

Fascinating article and some really interesting links to other things (like the Radiohead experiment). I'm a Guardian reader normally as we get a copy in the office each day - apart from over Xmas - but missed the article. When was it in the paper?

Edited December 29, 2008 by Nick

[Nightwolf]

WELL DONE

*I* think they're wrong. I do not question their data, but rather their measuring technique.

It is reminiscent of the famous poll by Literary Digest magazine in 1936, which predicted the Republican candidate for U.S. President, Alf Landon, was going to win by a big margin. LD's poll was nationwide; it was conducted according to the best scientific procedures in existence at the time. It had only one flaw. The poll was conducted by telephone. This was the middle of the Depression. Who had telephones? People who still had money and still had jobs. The people who didn't voted for Roosevelt--in big numbers.

I think the conclusion one is likely to draw from the information presented is that the recording industry has done a very good job of keeping outsiders out. If that makes the recording industry complacent, great. If they die from their complacency, well, they can join the other dinosaurs.

What I didn't see being measured here is how many records are sold *at gigs* by independent, not-very-famous artists. I have a feeling it's quite a few. No, nothing on the scale of the favorite children of the record companies--they have at their disposal what Joni Mitchell called "the Star-Maker Machinery," and it counts for a lot. But I myself have just about gone through four "pressings" of my last album in the last two years (true, they were small pressings)--I have only a handful left. And I have sold nearly every one of 'em at gigs. Next album, I'll see if I can generate some online sales. I don't know if it's possible, but it's a small investment and worth a try. And I'm small potatoes. There are people who are doing a lot more of this a lot better than I am.

My own impression, for what it may be worth, is it's a *very* long tail. It's just also very *thin". Think of a Norway rat. Big body, very long, thin tail. (Considering some of the people who populate the recording industry, a rat is not a bad analogy.) The rat probably does not think much about its tail. If the tail had independent thought, the tail might think more about the rat than the rat thinks about the tail. The rat, folks, is old and not in good health. The question I think *we* need to be asking ourselves is what happens to the tail when the rat dies. I'd like to see the tail survive without the rat, myself--maybe turn into something else rather than growing a new rat.

Joe

[FinnArild]

There's lies, there's damned lies - and then there is statistics. (Benjamin Disraeli)

[FinnArild]

What I didn't see being measured here is how many records are sold *at gigs* by independent, not-very-famous artists.

One could question if they even considered sites like CD Baby? I have sold a few there and elsewhere - and I'm sure I'm not in the top 15/20 % of worldwide albums/songs.

Think of a Norway rat.

Now, there's no need for namecalling ... ...

[Lazz]

just doesn't seem like a very "long tail" to me

I think I get it: looks like you are misinterpreting the term, John.

It’s not ‘long’ in the sense of time at all (‘though it used a year’s worth of data) just in terms of the tail’s plotted ‘length’.

All it refers to is the graphic shape of the data measure on a graph representation.

Most ordinary distributions (the usual ‘lognormal’) are the same ‘bell-curve’ shape where most everybody occupies the regular lumpen mass of the picture with a relatively small curve outwards at the bottom where all those who deviated from the norm hang out – but they don’t hang out very far.

The ‘long tail’, by comparison, is shaped more like a hockey stick where the bit that hits clusters up around the vertical axis while the handle stretches away low and long on the horizontal to accommodate a more diverse range of musical choices.

Anderson’s ‘long-tail’ picture of what’s going on is pretty much purely theoretical, though.

While Page’s preference is to be evidence-based at all times.

*I* think they're wrong. I do not question their data, but rather their measuring technique.

Hi Joe – sorry, mate, this seems a pretty spurious distinction.

Measuring techniques are what generate the numbers, so I fail to understand how you can accept the data while dumping the methodology. Especially as we know so little (it is not yet published) about how they did whatever it was they were up to. For something like this though – my guess is that it involved counting (which, as a technique, seems fairly reasonable).

Anderson agrees with you alright, but he has a career rep hanging on it. He is bitching at the moment because Page et al won’t release the full data so that he can tear into it. And neither is it known for sure from where and how they collected their data – they won’t say – but, with 13 million tracks as sample inventory, the smart money leans towards iTunes as the major or even sole source.

It is reminiscent of the famous poll by Literary Digest magazine in 1936, which predicted the Republican candidate for U.S. President, Alf Landon, was going to win by a big margin.

I see neither validity nor value in this comparison.

Maybe I’m missing something.

Care to elucidate ?

What I didn't see being measured here is how many records are sold *at gigs* by independent, not-very-famous artists. I have a feeling it's quite a few.

That’s like ordering a hamburger and bitching because it doesn’t have power windows.

(Sorry – you’re in the wrong shop.)

As I understand it, the long-tail theory as well as the surrounding debate has specific relevance ONLY for internet media business models.

(If you know of anywhere else it may be applicable, I would like to be made aware of it.)

The study was of on-line retail downloads – so (very sensibly, I think) that’s what they looked at.

They were not focusing at all, in any sense whatsoever, on live indie gig-sales of physical CDs.

The aim (as far as I am aware) was to subject this modern long-tail theory to testing.

When was it in the paper?

It was in both the Garundia and the Thunderer on December 22nd

The first public presentation of data was at the TelCo conference in November.

There's lies, there's damned lies - and then there is statistics. (Benjamin Disraeli)

Perhaps I should have mentioned this before, but I did predict there was a 99.98% chance that somebody would reference this quote – which I believe should be more fairly attributed to Mark Twain..

[TaoMannaDon]

The music industry is destroying itself. It chases away the best talent and turns its customers into criminals. Computer-driven radio playlists destroy any hope an unknown with talent might get heard on the radio. Inexpensive recording gear puts music production in the hands of the masses whether or not the "masses" have musical talent, so most independent music is more a labor of love than appealing creative expression.

The radio plays rap and crap.

The massive size of the Internet hides gems and raw talent from those who could appreciate independent effort.

What we need is a replacement for the DJ driven radio show. I can still recall the days when an excited DJ would find a new act or new record and play it on his radio show until the song became a hit. The DJs that found the best songs had the largest audiences. Those were exciting times for musicians and listeners. And I miss them.

On the other hand, there are independents on YouTube making lots of money from ads and all they do is talk. An enterprising writer might team up with others to create a music-driven video show with an interesting concept and find much more financial success than just posting songs to iTunes. All you need is a good idea.

There is a joy that comes from creating; if it is real and true, making money from your creations will be less important. Don't quit your day job. But don't give up your creative outlet just because you can't make a living from it.

Keep writing,

Don

[roxhythe]

Ooo. I knew that would generate some comments. Here's a few thoughts:

Lazz, my comparison with the Literary Digest poll was to suggest that depending on what one is measuring, and how one is measuring it, one can draw conclusions that are at odds with reality. I can see from my own and others' experience that we are selling CDs, at gigs and online--we're just not selling a lot of them. A statistical model that doesn't show that is missing something.

I am also a little distrustful of statistics that appear to come from the Industry showing that only a handful of people--people already in the Industry--are making any sales. That strikes me as more than a little self-serving. The Industry has had a deliberate program of preventing newcomers from breaking in for a number of years now; "information" that discourages people like me from trying to break in may have more than a little political taint. I don't know, but I do not intend to let it discourage me.

Two avenues for record sales that the Industry has mostly ignored are sales at performances and sales online. They are, I think, just starting to realize those exist. For people not already benefitting from the "Star-Maker machinery," those two avenues are likely to be the most productive avenues for sales--if and to the extent that we can take advantage of them, and prevent our exclusion from them like we were excluded from radio.

Finn, the Norway rat comparison was not to diss those of you in the Northland. I was after a particular *shape* of the money distribution curve, and the Norway rat comes close, I think. I live in a seaport, and we have those things here (I expect they're called "Norway rats" because somebody decided that's where they originally came from--I don't know if they even exist in Norway). They are the size of a medium-issue dog, have largish bodies, *very* ugly large teeth and very long, very thin tails. When I worked for the Pinkerton Detective Agency in Portland, which was a seaport, our guards on the waterfront used to carry .45-caliber pistols to defend themselves against the Norway rats. Very nasty characters (the rats, that is).

Joe

[Nick]

It was in both the Garundia and the Thunderer on December 22nd

The first public presentation of data was at the TelCo conference in November.

Perhaps I should have mentioned this before, but I did predict there was a 99.98% chance that somebody would reference this quote – which I believe should be more fairly attributed to Mark Twain..

Thanks

There was a 7.43% chance of the person 'using statistics as a drunken man uses lampposts - for support rather than illumination' quote coming up.

[Lazz]

'using statistics as a drunken man uses lampposts - for support rather than illumination'

Great quote - and completely new to me.

Where's it from ?

I still appreciate the post Lazz..............and I'm not gonna shoot the messenger just cause I'm not particularly fond of the message. The data is what the data is, at least according to this one study. I'm guessing that none of us is particularly happy with the current state of the industry, but I still think it's helpful to have some realistic idea of where things actually do stand, even taking into account that the study doesn't consider every variable. In the old days, the industry didn't count the albums rappers were selling out of the trunks of their cars either. When things reached the point where the numbers were too big to ignore, the industry started counting. I'm guessing that the indie thing thing just isn't to that point yet?

It is my impression that the 'indie' thing is certainly getting counted - car-trunk rap sales make a worthy comparison point.

But of course 'indie' has been arounf for farkin ages though in one form or another, hasn't it ?

I mean. for anyone like myself largely involved in the world of jazz it's always the case.

One other big point to be made here is that in order for 'the industry' to 'officially' count your sales you are required to pay for the privilege.

It costs a small fortune to get registered with SoundScan and play with the big-boys.

Why would we bother ?

Oh - and I spent a long time on an earlier response before you posted your latest, so I thought I would let the following fly anyway.

Forgive me.

Here it is.....

Lazz, my comparison with the Literary Digest poll was to suggest that depending on what one is measuring, and how one is measuring it, one can draw conclusions that are at odds with reality. I can see from my own and others' experience that we are selling CDs, at gigs and online--we're just not selling a lot of them. A statistical model that doesn't show that is missing something.

Yo Joe.

Far canal

And bugger me gently.

This is very weird – like criticizing football ‘cos it ain’t hockey.

The study is explicitly about on-line retail downloads.

NOT CDs – whether by gig-sale or CD Baby – NOT CDs.

NOT bit-torrent either.

OK – so it would be interesting to get a comparative contextual shot at the entire landscape.

But that plain ain’t what this is.

I am also a little distrustful of statistics that appear to come from the Industry showing that only a handful of people--people already in the Industry--are making any sales. That strikes me as more than a little self-serving.

Will Page’s salary-paying constituency resides predominantly inside that notional long-tail – they are the majority of little publishers writers and composers we don’t get to hear about as well as the Chappells and EMIs and McCartneys. What can possibly be self-serving about telling your employers that the latest theory they’ve all been hanging on to doesn’t actually turn out to be very robust ?

The ‘long-tail’ came touted as the golden-goose.

Where’s the percentage in killing it off ?

And just exactly how would such a thing be self-serving ?

I need that spelling out to me, please.

Appreciate it.

Please note, also, that ‘the industry’ does not appear overly keen to disseminate this info widely.

Perhaps, from a PR viewpoint, that could be self-serving.

I too am distrustful of stats.

Skepticism about source is always a good move.

This source appears to have impeccable credentials.

This source appears to be personally and professionally committed to the making of policies which are evidence-based rather than the making of evidence which is policy-based.

This source appears to be at least as skeptical as you and me.

(Read the TelCo interviews – pretty deep and open)

As for what is being measured:

For the Literary Digest – it was voting intention.

For Will Page et al – it was download retail over a 12 month period.

Voting intention is something people can and do lie about just to mess with the research, and you can always change your intention any time you want right up until the point of actually doing it.

Download purchases can only be identified through the completed act of actually doing it.

That seems a huge difference right there.

As for how the measuring is done:

For the Literary Digest – their method generated a predictably very skewed sample in most basic socio-economic terms.

(Normally, such limitations would be stated right up-front – but this was the Literary Digest after all – hardly a paragon of objectivity or literary or journalistic standards, and certainly without a fully open agenda.)

For the Page mob – all we know is that it is based on actual sales records over a time-period.

We don’t know where these records came from, or how representative they might be of the universe on which they are commenting, or many other details at all about their sample.

That’s exactly what Chris Anderson (‘Long-Tail’ author) is bitching about on his blog – that Page et al won’t give him access to the methodology or data-sets in advance of publication – but even he doesn’t have the temerity (advisedly) to question the integrity of the researcher

All we know for sure, however the data was generated, and despite Anderson’s irrelevant protestations that there are extant business models where his theory might be genuinely applicable, the distribution does not have a long-tail.

The Industry has had a deliberate program of preventing newcomers from breaking in for a number of years now; "information" that discourages people like me from trying to break in may have more than a little political taint. I don't know, but I do not intend to let it discourage me.

Dear Joe – with the greatest love and respect – and oodles of kudos to you for personal persistence and consistent delivery – but this reads to me as a shining example of being ‘self-serving’.

The industry has no such deliberate policy.

The guys who run their businesses are always trying to make money in the traffic of unpredictable fads and fancies. They are always obscenely hot for the next new thing, a pretty face with haircut and trousers they can market to the indiscriminately trend-conscious in exchange for their disposable income – which conveniently leaves guys like you and me way outside the general equation – although, in effect, the next new thing is merely some kind of re-incarnation of the same old new thing…. But they are always on the lookout for whatever they think they can sell. Nowadays, more than ever unlike before, the main reason they think they can sell stuff is because the stuff is already selling.

When and where they own resource-power, however, they do have a natural tendency to work as much market maximization and market maintenance as they can muster: i.e. they have power and money to play but little if any control over which horse is the best bet.

Two avenues for record sales that the Industry has mostly ignored are sales at performances and sales online.

Gobsmacked again, guv'nah.

Say it ain’t so, Joe.

It is totally untrue that the industry is ignorant of these things:

1) performance is the primary promotional point-of-sale

2) new media (largely owned and controlled by the same entertainment conglomerates anyway) is an active avenue for marketing and distribution.

The current popular shape of music contracts is an obvious reflection of those facts.

"Star-Maker machinery"

Yeah !

Good song from Joni, nice story, taut imagery – but maybe a concept appropriate for altogether different times.

Just around one handful of years back, Warner had big plans for a new young Irish boy. I can’t remember his name and never heard his stuff. But the big WEA honchos definitely dug it. And remember they are no novices at this game. They had a plan. And a promo strategy. They allotted a multi-million budget for the project. Significantly – and despite their best efforts and skills and business acumen – the boy ultimately sold a paltry 734 units.

That’s the power and control dichotomy: they have the power of all those marketing resources – yet they can never really control whether joe (excuse me) public actually picks up what they’re laying down.

I remember vividly also once watching the board at Island in the days when Frankie was around. Frankie wasn’t all they had that month, of course. Far from it. Same as every other cycle, there was a volume of other shit they had thrown against the wall at the same time. But it was Frankie Goes To Hollywood which stuck the hardest. So it was Frankie Goes To Hollywood who got all the heavy resource power market muscle and publicists and media mechanics lined up behind ‘em.

The rest just kinda slid off the wall and got ignored to death – including whatever the feck I was involved with at the time – but it wasn’t planned and no-one was in control. The world of commercial fad-driven opportunism doesn’t actually work like that on the ground. It has to be responsive for survival. The control delusion too often leads to the Warner-style woes of the previous example. Best to wait and see what’s shifting and then put the muscle behind the money-maker.

…. prevent our exclusion from them like we were excluded from radio.

Radio is a very different animal – and a bright pink halibut, too.

Radio, like print, is about selling an audience to advertisers.

So they max the audience through the lognormal playlist.

You know it makes sense.

For what it’s worth, by being selective and paying attention to programmers and programming, I have no trouble at all getting any of our stuff played in the land of non-profit radio. Problem is that packing and mailing and calling becomes an expense without compensation – because radio-play these days has absolutely zip impact on sales - and unless you have a cocaine and whores budget, as well as the product, to wangle your way onto the log-normal playlists of the bigger networks, there is an aperture for crossover which is firmly closed in our direction.

But who listens to radio anymore ?

That’s another story….

Innit.

More and more I am tickled by Sturgeon's Revelation that "90% of everything is crap".

[roxhythe]

Lazz, that's a lot of information. Thank you.

Joe

[FinnArild]

Perhaps I should have mentioned this before, but I did predict there was a 99.98% chance that somebody would reference this quote – which I believe should be more fairly attributed to Mark Twain..

Yes - I thought so too, until I researched it. Seems Mark Twain can only be entitled the "popularization" of the quote.

[Nightwolf]

Good read