In a typically thoughtful post, Ed Felten challenges the gloomy conclusions some people have drawn from the Radiohead ‘choose your own price’ strategy for their latest album.

Comscore released data claiming that 62% of customers set their price at zero, with the remaining 38% setting an average price of $6, which comes to an average price of $2.28 per customer.

Ed points to a basic economic point missing in this debate: Lower average price does not imply lower profit. *Radiohead may well be making more money because the price is lower.*

To see why this might be true, imagine that there are 10 customers willing to pay $10 for your album, 100 customers willing to pay only $2, and 1000 customers who will only listen if the price is zero. (For simplicity assume the cost of producing an extra copy is zero.) If you price the album at $10, you get ten buyers and make $100. If you price it at $2, you get 110 buyers and make $220. Lowering the price makes you more money.

Or you can ask each customer to name their own price, with a minimum of $2. If all customers pay their own valuation, then you get $10 from 10 customers and $2 from 100 customers, for a total of $300. You get perfect price discrimination — each customer pays his own valuation — which extracts the maximum possible revenue from these 110 customers.

Of course, in real life some customers who value the album at $10 will name a price of $2, so your revenue won’t reach the full $300. But if even one customer pays more than $2, you’re still better off than you’d be with any fixed price. Your price discrimination is imperfect, but it’s still better than not discriminating at all.

Now imagine that you can extract some nonzero amount of revenue from the customers who aren’t willing to pay at all, perhaps because because listening will make them more likely to buy your next album or recommend it to their friends. If you keep the name-your-own-price deal, and remove the $2 minimum, then you’ll capture this value because customers can name a price of zero. Some of the $10-value or $2-value people might also name a price of zero, but if not too many do so you might be better off removing the minimum and capturing some value from every customer….

What the discussion about this misses is the change in the cost-structure of music. Some time ago, William Fisher published an analysis of the cost structure of a CD. It turns out that distribution and retailing accounts for nearly half the cost, with artists getting only 12 per cent. Take out the costs of physical distribution and you have a very different picture.