Original URL: http://www.theregister.co.uk/2007/11/06/radiohead_comcast_stats/

American Radiohead fans are 73% more irrational - survey

And 99% of downloaders lie to pollsters?

By Andrew Orlowski

Posted in Media, 6th November 2007 11:05 GMT

Radiohead last month let punters "set the price" for the digital download of their new album It's Raining In Rainbows, which is coming out on CD shortly. Since people could download it for free - how much did they really pay?

Estimates to date have relied on self-selecting opinion poll data. This one, ComScore reminds us, is very different (see Update, below).

(Just as people lie to opinion pollsters about tax to make themselves appear more virtuous - they'll tell pollsters they're happy to pay more tax, but then vote otherwise - people lie to opinion pollsters about their voluntary contributions for digital music.)

ComCore based its survey on software monitoring user's behaviour (with their consent) - making it much more accurate than earlier polls.

In this survey, almost two-thirds of downloaders paid nothing for the album, and only 38 per cent made a contribution at all. However, the results show up an interesting quirk.

Overall, the band grossed about a quid per download, reckons ComScore. People who paid contributed an average of $6.00 (£2.89) - but once freeloaders were included, that falls to just $2.26, or £1.09 per album.

Either way, you can't build much of a business off a quid an album - that much we already knew (although digital utopians spend much of the time in denial about this).

Now here's the quirk.

From the US, the average contribution was $8.05, but outside the US it was $4.64. That's quite a disparity.

Since Radiohead made the download legally available for free, and since a consumer acts rationally to find the lowest price possible, one can surmise that US Radiohead fans are 73 per cent more irrational than Radiohead fans outside the United States.

Statistics: doncha just love 'em?

Nice gimmick, shame about the business model

The takeaway point from this will trouble anyone selling sound recordings, whether they're a basement indie or an established label - it's the fact that a top band with a worldwide fanbase which has been waiting four years for a new release, can only expect a quid from each LP in a voluntary model.

Comscore quotes independent A&R man Mike Laskow of Taxi:

"Radiohead has been bankrolled by their former label for the last 15 years. They've built a fanbase in the millions with their label, and now they're able to cash in on that fan base with none of the income or profit going to the label this time around."

(That's not strictly true - Radiohead actually are releasing a physical CD of the album using an old school record company, and the low bitrate MP3 preview is just a marketing gimmick.)

"At some point in the not too distant future, the music industry will run out of artists who have had major label support in helping them build a huge fan base. The question is: how will new artists be able to use this model in the future if they haven't built a fanbase in the millions in the years leading up to the release of their album under the pay what you'd like model?"

The instant response is that you use the internet to scale up to millions. But like so many utopian answers, the case for that at best is "not proven".

We'll know the answer once an artist has been able to build up and sustain a worldwide fanbase in the millions over the course of a few years - without major label support, and purely from sales of digital music. But we might be waiting for that for a very long time.

Clever bands who want a worldwide audience jump onto a major label. Now you know why. ®

Update

ComScore writes:
"Our study was not based on a poll at all, but rather on the actual observed behavior of our panel of Internet users. In other words, we saw exactly what they paid for the album so there is no potential for survey response bias. Our total worldwide panel has 2 million consumers, and we observed a few hundred transactions at the Radiohead site. Based on the size of the sample, the margin of error would be pretty small."