Derek Sivers had a blog post about this as well.

http://sivers.org/livecd

His post is about offering CDs after a live performance and asking people to pay what they want. They band in question went from $300/night in CD sales to $1200/night.

They were selling more CDs, but the average price was lower. It's still amazing that they got so much in total sales, but maybe they could have achieved something similar by just lowering the price.

Terry said that the band did this for a while, and soon they were selling about $1200 per night on average, even including those people who took it for free! I think the average selling price was about $10.

I would think that people would pay more because they are involved in a face-to-face transaction. There is peer pressure when you are buying to pay a fair price. In the online world, the human element is missing so I would guess that more people would pay much less to buy the CD than when buying it in person.

And this model does not scale. Derek knows this. I remember him from when I worked at mp3.com -- he is insanely smart and was exceptionally nice. But many things which worked for him do not work as your audience approaches the norm. Not everyone is charitable.

I wonder what the price distribution curve is for this (and for Rainbows)? In particular, were there many outliers who paid an abnormally high amount for a CD, like $200+?

There's more than one variable at work with the Radiohead album.

1. Over time, you would expect to make more for an album. Expanding fan base, inflation, ...

2. Insane word of mouth advertising because of the pricing plan.

3. It was probably a better album. As the article mentions, pay what you want is a great way to inspire better work.

It's great that it went well, but it's a stretch to say it proves anything. It was the first (mainstream) attempt at this model, and so many people (may have) paid to ensure that it worked.

On the flipside, take a look at Creatly, which is closer to what most folks here will be selling.

http://creately.com/blog/experience/how-much-to-charge-for-y...

They told their customers "Pay what you want!" and got a median of $3 and an average of $4 ~ $5, both of which are well under the prices of the cheapest plans of the vast majority of SAAS products and insulting next to the prices their competitors charge.

It proves that it is possible for a "pay-what-you-want" pricing scheme to generate more money for producers than the traditional model.

I am sure more money was spent in total on their previous album. What Radiohead did in this case was shrink the total market for their album yet capture more revenue for themselves. From an economist point of view Radiohead caused the GDP to shrink - less $ of goods were created - yet Radiohead got paid more and more people were able to listen to music they liked. So the world is better off yet the GDP is smaller.

Coupons and rebates are another example of "pay-what-you-want-pricing." The manufacturer sets a high list price then offers a discount to people motivated enough to jump through the hoops needed to get the discount. You decide if it matters enough to you.

My point is that there is room for innovation in economics - trotting out a 200 year old theory about grazing animals may not apply in a world where most of the things people want can be provider for almost zero cost.

Were they with a label in 2003? If so, it is certainly reasonable that 100% of a small number (their take from "in Rainbows") could trump 1% of a large number (their take from "Hail to the Thief"). HTTT sold well upwards of a million copies (gold in the US, platinum in the UK, which is at least 1.1 million). If the label's cut were not so excessive, Radiohead probably would have made more from their previous albums than from "in Rainbows."

But if the label's cut weren't so excessive, they wouldn't have gotten the same level of sales, due to a lack of budget for advertising and such.