The Death of the Music Industry

I stumbled on this interesting graphic the other day. They looked at per-capita spending on music in the United States and adjusted for inflation. I think this is interesting on a few levels. First, it shows how the move to digital sales hasn’t rescued the music industry (despite the occasional claim that ‘if you give people a chance to buy it digitally, people will pay’). The movie industry has also taken a hit over the past ten years (a slight decline in box-office revenue of about 15% or so, and a larger decline in DVD sales and rentals). While I think these declines can be traced (at least in part) to piracy, I do wonder why the games industry hasn’t seen a similar hit, though I can think of a few factors that might be affecting things: the increased danger and difficultly of pirating games on consoles (e.g. the XBox requires a mod-chip and you can still get kicked off the Microsoft network, the PS3 wasn’t cracked for years), the increased danger of pirating software (because malevolent software is far more dangerous, since it can infect your computer with viruses), the music industry was generally pretty nasty and unlikeable, the money spent on music wasn’t seen as going to the artist but going to the rich “suits” at the top.

See if you can guess where on this chart Napster was released.

This also makes me think of other issues, like: what is “enough” money needed to sustain an industry, versus when artists/creators are overpaid. Software is something that increases the number of workers (i.e. costs and functionality) as more money comes in. Movies are a bit similar to software, in that, the more money that comes in, the more money is spent creating elaborate special effects. It’s harder to argue that music gets better if a band is paid $1 million versus $20 million a year.

Fallout Soundtrack

Mark Morgan put together a free hi-quality version of the soundtracks to Fallout 1 & 2:

Mark Morgan’s “Vault Archives” is remastered and mixed full-bandwidth soundtrack from old-classic “Fallout” games (24 tracks).

I think he copied the original music and recorded it in much higher quality. For anyone who spent way too much time playing the Fallout series (like I did), this will bring back some memories. You can download it here or here, and listen to a few tracks using the embedded player here.

[Hat tip to Octopus Overlords]

Graphic: How Much Do Music Artists Earn Online?

InformationIsBeautiful has another interesting graph up. It compiles information about how much musicians earn through various music sales. I had no idea CDBaby was paying so much to the musician. Another interesting detail: musicians earn about 9% from each sale on iTunes or Amazon, and on the sale of retail CDs, labels earn 20% while musicians earn between 3% and 10% depending on the royalty deal. I assume this means that big-name artists can negotiate bigger percentages because they have more negotiating power. While some people get angry that the labels earn so much compared to the musician, I’m always hesitant to share their view because I don’t know enough details about the situation. For example, if a musician is earning $30,000 a year, and then he signs with a label so that they earn 100% of the music sales, but make him famous by pouring money into marketing, and the musician ends up making 10x as much money because his fame has increased his concert sales dramatically, then is earning a no money from each sale bad (assuming it comes with a dramatic rise in ticket sales)?

The other surprising thing about the chart is how little money anyone makes from streaming music services. In order for a musician (and this is a *single* musician, not a band) to make the equivalent of minimum wage by having his/her music on Rhapsody,, or Spotify, they have to get between 850,000 to 4.5 million plays per month – which is shockingly high. And, it’s not because the record labels are taking it. It’s because there’s so little money flowing to the record company or musician.

Click the image to see the full chart.

On a related note, one of the members of OK Go was on yesterday’s “Planet Money” podcast, commenting on a recent Op-Ed piece he wrote in the New York Times. He had mixed feelings about the record industry, but he said one good thing about the record companies is that they fund a lot of bands. 19 out of every 20 bands who get signed go nowhere. So, when 1 out of 20 make it big and earn money, the record companies take a lot of the sales revenue. Given the economics of the situation, they have to. Kulash argues that record companies are “risk aggregators” (or, at least, they used to be until they started losing so much money over the past ten years). What is a “risk aggregator”? It means that they spread-out the risks that bands take over lots of bands. It’s kind of like insurance. The bands who get rich pay money to support the less successful bands, except that it happens through the balance-sheet of a record company who only accepts bands into the group if they think they might be successful.

…It’s decisions like these that have earned record companies a reputation for being greedy and short-sighted. And by and large they deserve it. But before we cheer for the demise of the big bad machine, it’s important to remember that record companies provide the music industry with a vital service: they’re risk aggregators. Or at least, they used to be.

To go from playing at a local club once a month to actually supporting yourself with music requires big investments in touring, recording and promotion — investments young musicians can’t afford. My band didn’t sign a contract with EMI because we believed labels magically created stars. We signed because no banker in his right mind would give a band the startup capital it needs.

Record companies, on the other hand, didn’t used to expect that all their advances would be repaid. They spread the risk by betting on hundreds of artists at once, and they recouped their investments by taking the lion’s share of the profits on the few acts that succeeded.

At least, this was all true when we signed our deal in 2000. Today, as the record industry’s revenue model has collapsed with the digitization of its biggest commodities, companies are cutting back spending on all but their biggest stars, and not signing nearly as many new acts. If record companies can’t adapt to this new world, they will die out; and without advances, so will the futures of many talented bands.

In these tight times, it’s no surprise that EMI is trying to wring revenue out of everything we make, including our videos. But it needs to recognize the basic mechanics of the Internet. Curbing the viral spread of videos isn’t benefiting the company’s bottom line, or the music it’s there to support. The sooner record companies realize this, the better — though I fear it may already be too late.

Most people complain that the record companies are taking too much from the musician – but they’re only looking at how much they take from the successful bands, without looking at the full economic picture. I’m not really trying to be an apologist for the record companies. For all I know, they’re like ticketmaster – who unfairly harvests way too much money from ticket sales and does it’s best to drive competitors out of business or buys them to preserve the monopoly. I’m just a little skeptical about the whole narrative that gets passed around the internet. It’s also interesting to hear OK Go express the opinion about record companies that, while not all good, aren’t all bad, either.

Internet Echo Chambers

One of the obvious effects of the internet is that it’s allowed people to pick and choose news outlets that match and reinforce their own preexisting beliefs. It ends up creating communities of people reinforcing their own common viewpoint. Even worse, news stories that confirm their beliefs get brought into the internet site, providing additional “support” for their ideas. The problem is that there’s always a filter going on – the filter eliminates information conflicting with the community view, while popularizing information supporting the community view.

Now, as if that wasn’t bad enough, all information ends up getting filtered and reinterpreted – so that information which only tentatively supports a viewpoint gets reinterpreted as proving one viewpoint. I ran across a couple instances of this recently on the internet. A few cases:

Via Slashdot, I ran across an the article “Why my books are no longer for sale via Amazon”, which contained this paragraph:

Amazon still uses Digital Rights Management. This is not a customer-helpful feature. In fact, Apple iTunes has gotten rid of it, as have many other music sellers, and found that sales increase. If Amazon was doing this ‘for the little man’ then where are all their attempts to get rid of proprietary Kindle software and DRM? We know it increases sales to drop DRM, so why aren’t they doing it? Customers would love it. But Amazon would lose control.

Now, like a lot of creators, I have mixed feelings about DRM. But, what I was interested in was the quote “we all know if increases sales to drop DRM”. Fortunately, he provided a link. So, I’ll bite. For someone who wants to get to the facts, what’s the evidence for his statement that “we all know”? His link leads to a 2007 Crunchgear article, “EMI’s DRM-free Sales Are Working”:

Plenty of people looked disgusted with EMI’s decision to turn its catalog of music loose, free of DRM. Now it seems the tables have turned, as EMI’s DRM-free approach to selling digital music is working out quite well. According to reports, Floyd’s Dark Side of the Moon has seen a sales increase between 272 and 350 percent, while OK Go’s Oh No has increased 77 percent. Those are huge numbers and could mean the difference between a paltry paycheck and a big fat one.

So with increased sales and widespread praise from just about everyone, will other record labels follow suit? Is this what the industry needed to realize that DRM just isn’t the way to go anymore? Looks like it to me.

Okay, sounds good. What’s their source? A 2007 BoingBoing post “DRM-free EMI music outselling lockware”:

The Inquirer reports that in the short time since EMI went DRM-free with its music, its sales have skyrocketed: Since EMI ditched the DRM on iTunes it has seen sales of Pink Floyd’s Dark Side of the Moon increase by between 272 and 350 percent… According to Bloomberg, digital sales for other DRM free music increased by between 17 to 24 per cent. OK Go’s Oh No increased 77 per cent. Coldplay’s A Rush Of Blood To The Head jumped 115 per cent.

Hm. What’s BoingBoing’s source? The, “EMI has DRM free sales boom” (I bolded a few interesting sentences left out of the above “news reports”.)

RECORD OUTFIT EMI is reporting that the sales results for its DRM-free music are better than those with protection. Since EMI ditched the DRM on iTunes it has seen sales of Pink Floyd’s Dark Side of the Moon increase by between 272 and 350 percent. It is too early to tell if this is just a temporary rise as punters replace their old DRM infected tracks with those which are protection free. However it does look like punters have given the record labels the thumbs up for DRM free music. According to Bloomberg, digital sales for other DRM free music increased by between 17 to 24 per cent. OK Go’s Oh No increased 77 per cent. Coldplay’s A Rush Of Blood To The Head jumped 115 per cent. Sadly CD sales still are dropping down the toilet.

And their source for this information? They provide two links, but one is dead. The working link is a blog about the music business: Coolfer, “EMI Says Dropping DRM Showing Good Initial Results, But Questions Emerge”:

Bloomberg News has an article in which an EMI SVP said, “The initial results of DRM-free music are good.” Increased sales of Pink Floyd’s Dark Side of the Moon were singled out. That’s true. Digital sales of Dark Side of the Moon have averaged over 3,600 units since the launch of iTunes Plus and the availability of unprotected AAC files. In the 11 prior weeks, average sales were 830 units per week. That’s an increase of 272%. In the week iTunes Plus was released, digital sales of Dark Side of the Moon jumped 350% that week alone. (iTunes Plus, which enables users to upgrade tracks to unprotected AAC from the older, protected AAC versions for $0.30 per track, lauched May 30th.)

Okay. So, sales jumped from 830 units/week in the 11 weeks leading up to May 30, to 3,600 (units/week?) in the 2-weeks or 2.5 weeks following. Now, my first question looking at this is: Are these results typical for all music sales after iTunes Plus, or was there some other factors involved to drive-up “Dark Side of the Moon” sales specifically, which could undermine the claim that these increased sales were due to dropping DRM.

The blogger goes on:

Here’s the main question: Does an upgrade using iTunes Plus count as a scan? I can’t find out. If that’s the case, the increase in sales will actually a temporary thing. Once people who want to upgrade their tracks have done so, sales should drop and level off at pre-iTunes Plus levels (or, as EMI is hoping, above pre-launch numbers). My gut tells me SoundScan counts an upgrade as a sale. Those Pink Floyd numbers look to be more indicative of a technology-enabled sales jump than they are a sign of support for DRM-free downloads.

Yes, that’s true. If iTunes counts every $0.30 upgrade from DRMed “Dark Side of the Moon” to DRM-free “Dark Side of the Moon”, then does that really count as an argument against DRM?

What about other albums? Digital sales in the last two weeks for Smashing Pumpkins’ Siamese Dream rose 17% versus the 11 prior weeks’ average. Norah Jones’ Come Away With Me jumped nearly 24%. OK Go’s Oh No increased 77%. Coldplay’s A Rush Of Blood To The Head jumped 115%.

Notably, these numbers aren’t nearly as high. It makes me wonder if other things were going on with these bands at the same time that iTunes Plus launched. Afterall, if these sales increases are caused by eliminating DRM, then why are there such large disparities in sales increases? In other words, why is there a 272% increase for Dark Side of the Moon, but only a 17% increase for Siamese Dream? I also wonder if these sales increases were picked from the albums that spiked that week. They’re only showing five albums. Are there thousands of others that showed virtually no change? And did the announcement of DRM-free tracks cause a spike in iTunes sales in general – not necessarily because they are DRM-free, but merely because iTunes was in the news? Interestingly, the lower sales percentages were dropped from the downstream blog-posts citing this article. By the time you get to the CrunchGear article, only two albums are cited (Dark Side of the Moon +272%, and Oh No +77%) and then they act as if these results are typical.

Those are just digital album sales I’m talking about. But here’s something very interesting: CD sales of four of those five titles dropped sharply over the same period. CD sales of OK Go’s Oh No dropped 45%. CD sales of Coldplay’s A Rush Of Blood To The Head dropped 24%. Norah Jones’ Come Away With Me fell nearly 33%. Pink Floyd’s Dark Side of the Moon is down only 12%. Smashing Pumpkins’ Siamese Dream increased almost 15% — probably the result of the band’s new single and upcoming album hype.

Interesting, so sales of DRM-free digital downloads might be cannibalizing sales of DRM-free CDs. Even worse, CD sales are still much larger than digital sales. From numbers I’ve seen, CD sales in 2007 would’ve been bringing in 7x as much sales revenue as digital sales. This means, a 50% increase in digital sales won’t offset losses from a 10% drop in CD sales.

The blogger seems to agree with the cannibalizing interpretation:

“The most likely scenario is one in which iTunes Plus — and the emergence of higher quality, DRM-free downloads — has somehow accelerated the CD-for-digital substitution that has eroded CD sales.”

What’s the point of all this? When you look at people citing one source after another, the information was filtered and twisted into a viewpoint that confirms their preferred beliefs. The original article said that “We know it increases sales to drop DRM” – a bold and unequivocal statement, but the follow the chain of links down to the original source and you find a much more complex picture – one that’s far more tentative and unclear, with unanswered questions and conflicting information.

This chart shows who dropped which information from their blog-post. BoingBoing dropped all the negative information about the decline in CD sales and the possibility that this was a temporary increase. Tech Crunch dropped some of the lower sales increases in favor of the high-sales increases and suggested they were typical.

On that note, I’ll quickly cite a second example along the same lines. This is another BoingBoing distortion:

BoingBoing: O’Reilly drops ebook DRM, sees 104% increase in sales:

It’s been 18 months since O’Reilly, the world’s largest publisher of tech books, stopped using DRM on its ebooks. In the intervening time, O’Reilly’s ebook sales have increased by 104 percent…

Now, obviously, there’s been a big rise in the number of eReaders on the market in the past few years, so the claim that this rise is due to dropping DRM is questionable. What about the publishing industry as a whole? How have their eBook sales been going (which still include DRM)?

Based on a comparison of these charts, I’d say that the publishing industry as a whole has seen bigger increases in eBooks than O’Reilly. It looks to me like the publishing industry saw an 225% increase in eBook revenue over the time that O’Reilly saw a 104% increase.

Now, I’m not arguing that the music and book industry should use DRM. Personally, I think it’s nonsense to sell DRMed music when they’re also selling CDs (which aren’t DRMed) where it’s trivial to turn them into mp3s and upload them. I’m just pointing out the distortion of the facts.

What’s really sad is the way information gets cut-up and edited to support a viewpoint that someone wants to advance — to create an alternate reality where “everyone knows X” even though “X” isn’t supported by the data. At each step along the way, information was removed to make the story more simple and convincing, but less accurate. Notably, this isn’t corporations distorting data to serve their own financial interests, but it’s “consumer advocates” who are doing the distorting to reinforce and promote their own views about the world.

Random Thought on DRM

I was thinking about DRM lately, and one idea that crossed my mind was this: what about a system where every executable contains a serial number. The serial number would be long (and therefore resistant to brute-force attacks). It would be used whenever connecting to the (company) servers to get updates, play multiplayer games, or any other downloadable game content.

You could then install the software wherever you wanted, without any kind of server validation, but if too many computers with the same serial number started asking for the exact same update or multiplayer access, then we could assume the game had been pirated. That serial number could be automatically blacklisted from server access. In other words, if someone uploads their copy to a pirate website, they would end up harming their own copy and all the pirates could also be blocked from accessing the server. Optionally, the application could either be disabled or could start show a nag message about buying the software.

While it doesn’t actually stop anyone from using the application (unless the application is disabled, as indicated above), this system is resistant to a variety of hacks. Using a long serial number means people can’t just guess a different serial number for their copy. If they rewrite the serial number to something else (like a bad serial number), they still can’t access the server.

This system would also accomplish a couple of good things. It would allow the application to continue working even if there were no DRM servers running (e.g. if the company goes bankrupt). It allows users to install the application where they want without needing to remember a registration code (the registration code is built in). Users would never have to worry about unregistering their copies on old computers (if you computer suddenly dies or your hard drive goes bad, no problem because the only thing that matters is whether it’s asking for updates or server access). It allows the company to recognize and selectively ban copies that have been spread on the internet by pirates.

While I was thinking about this, I realized that it only works very well for software. And it works better for multiplayer games than single-player games. (If someone pirates a single-player game, doesn’t care about updates, and doesn’t care about downloadable content, it doesn’t really restrict them.) Theoretically, this DRM system could be open-sourced, too, because it doesn’t get harmed by the fact that people can see what’s going on inside the code. Plus, it would be nice to see people’s reaction to the phrase “open source DRM”. I don’t think I’d want to do that, though, because I think open-source DRM would attract too many confederates who want to either destroy the system or build-in backdoors. One other possible problem with the system is that a pirate could hack the application to point it to another server, then setup their own (open source) copy of the server elsewhere, serving up copies of the updates and other downloadable content. This would a somewhat dangerous strategy, though, because it means setting up a website. Websites can get shutdown, they cost money, and it means that their identity might be revealed. Still, I’ve seen websites that were clones of other websites, in an attempt to get some ad-traffic based on someone elses’s content.

Unfortunately, it’s not a system that can work for things like movies or music, because they don’t benefit from updates, server-access, or multiplayer access. Well, the other day, I stumbled on one company seems to be trying this strategy with music:

Is the World Ready for the Successor of the MP3?

A leading technology company is set to launch a new digital music file format that will embed additional content for fans including lyrics, news updates and images in what could be a successor to the ubiquitous MP3 file.

Music labels, bands or retailers could then also send updates to the music file every time they have something new to announce such as the dates of future tours, new interviews or updates to social network pages.
(Source: Wired Magazine)

What they seem to be doing in this case is embedding a serial number in the MP3 metadata. They then update the file with new data (images, lyrics, etc) if you have a valid serial number. I don’t think their system works very well, though. First, once someone gets the update, they can pass it around to everyone else, giving them access to the images and lyrics. Also, I’m unclear on what happens if you have 40 songs by one artist. Are all 40 mp3s going to get updated with tour dates? That seems inefficient. And, what if someone had one legitimate mp3 on my system, and 39 pirated mp3s? Would that mean that the one legitimate mp3 be enough to get all the tour date and news information? While it’s generally a step in the right direction, I don’t think it’s going to be terribly beneficial to the music industry. Maybe it will help pull-in some of the music fanatics that absolutely need the best.

Free Coldplay Download

Coldplay is giving away a free copy of their album (9 songs). This is apparently a way to drive up concert sales — they released the free album on the same day their North American tour started; free downloads will end on the last day of their tour in August. I’ve never been a big fan of Coldplay, but I figure there’s no reason not to check it out. If you like free music, get it here (FYI: they ask for your email address, but your email address is not verified by their website).

Norwegian Study: Music Pirates Buy More Music?

I recently stumbled on a recent news article claiming that, according to a Norwegian study of music-buying habits, people who pirate music buy ten times as much music as people who don’t.

I’ve also seen a similar study done in Canada. I haven’t had a chance to delve into them too deeply (though the Norwegian one doesn’t seem to be available in English, anyway). A quick google search reveals that a lot of pro-filesharing websites are praising the results, but I’m skeptical.

The first questions I think when I read about these types of studies are:
– “Did researchers rely on pirates self-reported music purchases?” and
– “Did pirates feel increased pressure to purchase music when they knew their actions were being recorded by researchers?”

According to the write-up, “the study did not rely on music pirates’ honesty. Researchers asked music buyers to prove that they had proof of purchase.” Sounds good, although I’d wonder if all of them had proof of purchase, and how the researchers treated the case where they didn’t have a receipt.

Another possibility is that a portion of pirates are pretending to be non-pirates – either for legal reasons or because they’re saying what is most socially acceptable. (Many social scientists are very skeptical about self-reported numbers when a behavior is not socially acceptable.) If a pirate never buys anything, but masquerades as a non-pirate, it lowers the apparent music-purchasing numbers of non-pirates relative to pirates. On the flip-side, the pirates who “try before they buy” feel more justified in admitting their piracy since they also buy a great deal of music. This would create the appearance that the average pirate is more likely to buy music than is actually true. There is good reason to suggest that this is going on. According to a recent Dutch study, 93% of Dutch people between the ages of 12-24 admit to downloading music, movies, or software. And another study says “81% of Spanish Internet users under 24 download copyrighted content via P2P files.” On contrast, the Norwegian study says only 40% of Norwegians between the ages of 15-20 admit to “free downloading”. Why is there a 53% gap between the Dutch and Norwegians? Maybe this is explained by Norwegians claiming not to pirate when they actually do. If we were to mix-together the Dutch and Norwegian numbers (questionable, I know, but this scenario is theoretically possible), all this study is comparing is music purchases by people who admit to piracy (40%) compared to music purchases by people who don’t admit to piracy – which is mostly pirates (53%), but also a small section of people who aren’t pirates (7%). (Even worse, the Norwegian study didn’t ask about “piracy” specifically, but they asked whether or not people had taken “free downloads” – meaning that 40% might include non-pirates.)

And a fourth possibility is that piracy and music purchases are both correlated with wealth. In other words, if you have wealth, you are more likely to have a computer (which you need for piracy) and you are more willing to buy music (because you have plenty of extra money). Less-wealthy Norwegians, on the other hand, might be making cassette tape/CD copies of music, but aren’t pirating off the internet, and they aren’t buying much music. This situation would create a correlation between piracy and purchases, but would invalidate claims that ‘piracy drives purchases’.

I did notice one odd thing in the news article. Early in the article they say:

A report from the BI Norwegian School of Management has found that those who download music illegally are also 10 times more likely to pay for songs than those who don’t.

But, a few paragraphs later, they say:

Researchers found that those who downloaded “free” music – whether from lawful or seedy sources – were also 10 times more likely to pay for music.

Wait – free music from lawful sources? I’ve downloaded free music from lawful sources. I’ve downloaded free albums from band’s websites. I’ve downloaded free songs from MySpace (when they have the little “download” link). I’ve downloaded free songs using those little Starbucks / iTunes promotional cards. I do not pirate. Apparently, I would be included in the group of “those who downloaded free music” – which some people are interpreting as “pirates”. So, my music purchases would be included in that “music purchases by pirates” group. I’m not at all surprised that people who downloaded free, legal music would purchase more music than people who didn’t. Afterall, people who download music (legally or illegally) are more likely to be “big music fans” than people who don’t, so I would expect a correlation.

There’s several reasons I don’t trust these studies:

(1) It doesn’t fit my personal experience of how pirates operate. I know a few people who pirate. If you told them to pay for some digital content, they’d look at you like you bumped your head and forgot that they have this awesome thing called piracy. On different occasions, I’ve seen two of these pirates chide other people for buying things that could be pirated. Clearly, they aren’t the “buying” type of pirates if they complain when other people buy digital media. [June 9, 2009 Update: According to this BBC article, the British Phonographic Industry is quoted as saying “The consensus among independent research is that a third of illegal file-sharers may buy more music and around two thirds buy less.” This pattern would, at least, confirm my observation about many pirates refusing to pay for digital content. I have to wonder about the other one-third. Are they using piracy merely as a “try before you buy” service, but trying to follow the spirit of the copyright laws? I think that’s possible. At the same time, if that interpretation were true, then it would suggest that they might act differently in a world with legalized filesharing – which is something that some anticopyright activists are pushing for. It’s likely an incorrect assumption that pirates today behave similarly to the way people would behave if filesharing were legal.]

(2) One of the common reasons people on the internet give for piracy is that they are poor. If someone is poor, they aren’t likely to pay for the stuff – especially if they’ve already pirated it. Other pirates complain that music companies or musicians make too much money, they hate DRM, or they’re sticking-it to the nasty record companies who are evil because they sue pirates or pay the musician pennies per sale. I even saw one pirate claim that people pirate as an act of “civil disobedience” against unfair copyright laws. People using those explanations for their piracy don’t sound like people who turn-around and buy music. So, when this Norwegian study says the average pirate buys ten times as much music as non-pirates, I’m skeptical. It would be interesting to see the detailed information about this. Maybe half of all pirates never pay for anything, and the other half actually do buy stuff. In this case, the average “buying” pirate would have to purchase 20x as much stuff as the average non-pirate. That seems like a stretch, but at least that explanation would confirm my own observations about the existence of non-buying pirates while also allowing the study to be right.

(3) Music sales have been on a downward trend for the past eight or so years. Now, there may be plenty of factors involved, but they’re down something like 40% compared to the year 2000. In Spain, where filesharing is legal, music sales dropped 56% between 2001 and 2008. Piracy might only be part of the explanation for that decline, but let’s say that piracy increases music sales (i.e. piracy increases sales, rather than piracy is correlated with sales). This means music sales are down 40% (or 56%) despite sales-increases due to piracy? That seems unlikely. Afterall, if we assumed an increase in piracy of only 10% (for example, from 20% to 30% of the population), and we believed that piracy increases sales by 10x, then a 10% increase in piracy should result in nearly doubling total music sales. Instead, we see things going in the exact opposite direction.

So, I’m still doubtful about the result that pirates (on average) buy 10x as much music as non-pirates. I also have to question whether or not these studies are done accurately and honestly (by both research subjects and researchers themselves). Afterall, South Korean scientists have faked their results. Piracy is a contentious issue, and I wouldn’t doubt that people would fabricate results. Another possibility is that some biased news reporter is twisting the results to say something that they didn’t say. And since the original results are in Norwegian, it’s hard to verify the actual results.

Links to the Norwegian sources (translated by google):
BI Norwegian School of Management
Norwegian News

NIN and Business Experiments

I’m always kind of curious about the results when companies try different business strategies. Nine Inch Nails released their last two of their albums for free. Well, free if you wanted them for free. Anyone can download The Slip off of their website for free. You can get the first nine songs of Ghosts I-IV for free. If you want the other 27 songs, you can get them a few different ways: free, $5, $10, $75, or $300 – depending on which package you want. (The packages at $10+ contain various physical media, including DVDs, Blue-Ray with surround sound, etc.)

When they first did this, I wasn’t quite sure about his motivation, but I can imagine a whole bunch of possible motivations – to drive up concert sales, as charity to the fans, faith that fans would pay, etc.

I’ve seen a number of people use the NIN experiments as evidence that giving away music “works for musicians”. In many cases, this is part of a “piracy doesn’t hurt anyone” or “this shows that creative commons works” kind of an argument. For example, RollingStone says: “it proves — along with Radiohead’s chart-topping success of the physical release of In Rainbows — that music fans are willing to support artists even if their music is offered up at no cost.” And Ars Technica: “Nine Inch Nails’ Ghosts I-IV topped Amazon’s digital download charts in 2008, despite being offered to fans free of charge. Creative Commons touts the project as a new model for music… The next time someone tries to convince you that releasing music under CC will cannibalize digital sales, remember that Ghosts I-IV broke that rule, and point them here.”

I’m less certain about the success of these experiments. Here’s the results: NIN Ghosts I-IV was the number one downloaded album on Amazon’s 2008 charts, and brought in $1.6 million in revenue within the first week of sales. Sounds good. But, let’s break down those numbers.

They sold Ghosts I-IV in a few different formats. First, you can download the first 9 tracks for free off their website. All 36 songs are available for free on filesharing networks. Even though it was released it under Creative Commons for free, Trent Reznor didn’t make it too obvious and doesn’t provide a link to download all 36 songs. It appears that most people didn’t know that the full album was available for free. And even if they did know it was free, they wouldn’t know how or where to get it anyway. It’s also available in $5, $10, $75, and $300 Ultra-Deluxe packages. We know that there were 2500 copies sold of the $300 packages. (It pays to have rabid fans.) But we have to guess on the rest. We also know that there were “781,917 transactions“. Obviously, “781,917 transactions” includes people who downloaded the 9 free songs (if it only included actual sales, then 781,917 transactions would be a lot more than $1.6 million).

Here’s what I think is a reasonable guess of sales for the first week:

$300 package x 2,500 copies sold = $750,000
$75 package x 5,000 copies sold (just a guess) = $375,000
$10 package x 10,000 copies sold (just a guess) = $100,000
$5 package x 78,884 copies sold (just a guess) = $394,420
Free download x 685,533 people (88% of the transactions)
Total Transactions = 781,917 (the number given to us from this article)
Total Revenue = $1,619,420 (the number given to us from this article)

Now, under these numbers, about 12% or 96,000 people bought something. (If no one bought the $75 or $10 packages, then $5 sales would be at 174,000 copies.) About 700,000 people downloaded the free 9-song version, and who knows how many people downloaded all 36 songs via peer-to-peer networks.

I think one lesson that can be learned is that tiered pricing a big benefit to bands with rabid fans. Most bands couldn’t get away with a $300 package, but (based on my guesses) NIN made 70% of their money from the $300 and $75 packages.

But, I was also surprised by the low number of purchases. Fewer than 100,000 people bought something. Okay, so this is only the first week of sales, but in 2005 (less than a month after their fifth album was released) NIN claimed to have over 20 million albums sold worldwide, which means an average of 4 to 5 million copies sold per album. Additionally, with 800,000 downloads, it means that 7 out of 8 people downloaded the free version and didn’t buy anything. Of course, this makes me question Ars Technica’s conclusion that “[Creative Commons won’t] cannibalize digital sales”.

As far as being the number one downloaded album on Amazon in 2008 (which is based on the entire year, not just the first week’s sales), it looks like NIN routed all their $5 sales through as mp3 downloads. Since Amazon doesn’t have strong mp3 sales in general and all of the $5 sales went though Amazon, and the fact that competing music was split between physical media and downloads as well as being available at a variety of stores, then maybe it’s not surprising that they placed so high in Amazon’s best-selling mp3 list. Digital music downloads are still dwarfed by physical-music sales, making up only 10-20% of total music sales. Given these facts, “#1 downloaded album on Amazon in 2008” is a little less impressive than it sounds.

So, was Reznor’s experiment a success? Does this vindicate Creative Commons as a viable and reasonable alternative to selling music? I’m doubtful. It seems like Reznor has mixed feelings. In an interview done last year (before releasing Ghosts I-IV) Reznor talked about his collaboration with Saul Williams. They put an album together and released it as both free download and as a $5 purchase:

Reznor produced and helped bankroll the album, which debuted November 1. All the more reason why he was stunned when fewer than one in five people who downloaded the music were willing to pony up $5, roughly the cost of a McDonald’s Quarter Pounder.

Reznor ended the hoopla last week when he reported on his blog that 154,449 people had downloaded NiggyTardust and 28,322 of them paid the $5 as of January 2. In the blog, Reznor suggested that he was “disheartened” by the results.

[Y]ou have a lot of reasons why you [don’t want to buy music from the record labels]. So I thought if you take all those away and here’s the record in as great a quality as you could ever want, it’s available now and it’s offered for an insulting low price, which I consider $5 to be, I thought that it would appeal to more people than it did. That’s where my sense of disappointment is in general, that the idea was wrong in my head and for once I’ve given people too much credit.

Saul and I went at this thing with the right intentions. We wanted to put out the music that we believe in. We want to do it as unencumbered and as un-revenue-ad-generated and un-corporate-affiliated as possible. We wanted it without a string attached, without the hassle

I’m not saying it was a failure or a success. I think it was both. But it wasn’t 90 percent of the people that showed up paid us what we asked for. Nor did I ever think it would be. I’m not sure what I did expect.

But I’ve found it entertaining reading different people’s perspective on the Web, what they’ve thought of what I’ve said. There’s been a wave of people that said, ‘Oh, that’s depressing. Only 18 percent chose to pay for it.’ Another whole wave of people feel just the opposite. I don’t really know.
CNet News: Trent Reznor: Why won’t people pay $5?

Now, looking at the downloads versus purchases of Ghosts I-IV, we can say that only about 12-13% of the people paid $5 or more for the 36-track album (as opposed to downloading the free 9-tracks from the NIN site). Additionally, a lot of people got the music via peer-to-peer, which means a lot less than 12-13% of people who got the album paid for it. On the positive side, the expensive $300 and $75 packages helped offset the large majority who paid nothing, and selling directly to fans let them skip the record-label middleman – letting them reap more revenue per sale. Although, it would be a false dichotomy to say that NIN only had two choices: the record labels or free/creative commons. A third alternative would be to simply sell the album from their website (without putting it under creative commons, and without putting it on bit-torrent).

It’s true that he did “allow” it to be shared, and that he uploaded it to bit-torrent himself – which seems to suggest that he believes that “giving it away for free” doesn’t hurt sales. On the other hand, NIN didn’t make it easy or obvious to download the entire 36-track album from their website. This made it difficult for anyone without bit-torrent to get it for free. My take on this behavior is that Trent Reznor doesn’t try to stop filesharing because it’s going to happen, but he also doesn’t want to make it easy for non-bit-torrent / non-pirates to download his music for free. I can understand the logic of this. Afterall, the music would be leaked to bit-torrent regardless of NIN’s stance on the issue. And people who torrent wouldn’t be swayed by Reznor’s appeal to not pirate his music, anyway. Since he has zero control over piracy, he might as well try to make some friends and get free publicity by taking a controversial stand.

It would be nice if we could generalize his experience to other bands / companies. But, there’s obviously a number of complications in trying to do that.

I suppose some of the lessons or non-lessons might be:

– Sell physical products with tiered pricing. This way, your biggest fans can pay you $75 or $300. Unfortunately, this only works for bands with a strong following; most bands aren’t going to get very far selling $300 packages. But, there might be other tiered pricing that works – like $10 / $20 / $30.

– Publicity stunts like releasing music under creative commons or making it available on torrents might get you some media attention. Unfortunately, this only works with bands that can already garner news media attention. No one cares if the local band has gone creative-commons.

– Selling directly to fans lets musicians get a bigger cut of each sale. But, this strategy involves forgoing the record-label’s publicity machine. NIN already has enough fame that “more publicity” has limited value.

– Even if you eliminate all the “barriers” that most people cite for pirating music (including: DRM, don’t want to pay the evil record labels, the musicians don’t make money from music sales, lowering prices down to a ridiculously low $5 for 36 songs), most people still aren’t going to pay you if you give them a free option. The percent who pay will probably vary depending on how much people like the music, but if NIN is getting fewer than 1 in 8 to pay anything (and numbers are a lot lower if you include torrents), I think things look rather dismal. On the other hand, we have no idea how many of those “free download” people even liked the album or would’ve bought it.

– Even Trent Reznor seems to agree that making it easy / convenient for non-pirates to download your music for free probably hurts sales. On the other hand, releasing it for free on bit-torrent probably doesn’t hurt you much because it’s going to be there regardless of what you do, and people using bit-torrent are going to pirate music anyway.