Wednesday, November 7, 2007

Reading into Radiohead's Results

There has been what looks like some rigorous analysis of the Radiohead move (which I blogged about here, here, and here) to make their album available online at a pick-your-price, no-DRM format. This comes to us from

The results are astonishing. An estimated 1.2M copies of the album were downloaded, with 38% of downloaders chosing to pay, and of those, the average was $6 per copy paid (Americans paid a more generous average of $8).

Back of the envelope math says 1.2M * 38% *$6 = a little under $3M. I have no idea of the distribution arrangement, but lets say they give up 10 points for billing, bandwidth, etc, andanother 10 points to someone to manage all this for them. That still leaves $2.5M.

Now the article linked to above comments about the number of copies circulating via bit torrent, etc, but that's beside the point, isn't it? That'd be happening if they were on iTunes. The real question is how many copies they'd have to sell on $10 - $15 plastic where they are making, say, 10 points. (I have no idea what rev shares are in the music biz, so I'm really blowing smoke here). At that point, they'd have to move 2M copies in order to reach the same level.

That's totally doable, but not *certain*. Their first four albums went platinum in both UK and US, so at least 2M units per album, but their last couple albums were between 500k and 1.5M-ish each.

So in this case, they (a) hit at least the minimum threshold they'd have done via the traditional channels, (b) now have a direct relationship with their customers, and (c) created a marketing vehicle for the album and the band that money couldn't have bought.

Sounds like a win to me!


Patrick said...

Free to play knows no bounds. I bet you can apply this to your work. Casual games/worlds with a stronger server component (running the gamut from instanced with leaderboard to full-on MMO) are starting to become as prevelant as downloads and Flash games were four years ago.

Chris 'Zeke' said...


In the olden days of the music biz, 15% of suggested retail price was generous.

Roughly $3/unit. Some bands would make less than $1/unit.

Ben Sizer said...

The problem with this model is that it doesn't extend to the middle-sized artists. Discussion tends to continue the industry-perpetuated fallacy that the physical medium and distribution are the only costly aspects here. After all, economists like to disregard the fixed costs, because after enough units, they effectively disappear.

However, recording an album is not cheap to most people. The likes of Radiohead have plenty of cash up front, so they can easily afford to go into the studio and do a recording. Then they can take advantage of this to use a model such as the one we've seen, which may well produce a better yield than the usual one. But that doesn't extend to smaller acts, who don't have the capital up front. Record labels are often effectively investors, who make a call on which bands are worth the investment. Take them out of the equation, and that investment dries up. Where would the next Radiohead come from? Many medium-sized labels are not just corporate machines churning out a commodity, but are actually experts specialising in certain areas and who can pick out the diamonds in the rough.

Radiohead also have plenty of 'investor interest' in terms of their existing fan base. They're not just an established band, they're one of the most popular in the world. They could take out a loan to cover recording costs based on that prior interest because it would be obvious they'd be able to repay it. Beginning artists would not have that luxury. Even artists on their 2nd or 3rd album might not be able to do that. And Radiohead's existing fan base provides massive amounts of traffic to their site - where is that traffic going to come from, for the smaller bands?

It's also important to remember that Radiohead benefitted massively from the publicity surrounding this public experiment. Once everybody's doing it, that excitement is going to die down. Who's going to go from site to site and use these mechanisms to donate and download when they can just use a single P2P system? I expect the success rate will drop dramatically.

However, this isn't to say I'm in favour of DRM or anything like it. I just don't think turning music into donationware is the answer, and I don't think pointing to one extreme outlier and using it as an example of success is very accurate.