Starting with the topic of Atlantic's "artificial scarcity" experiment with Estelle's album Shine, the LA Times' Jon Healey veers off into a thought-provoking post on pricing, demand and digital downloads. He spoke with Frank Luby, a partner at Simon-Kucher & Partners and co-author of "Manage for Profit, Not for Market Share." It's mandatory reading. Here's a sample.

Luby's central assumption is that when people want to buy a song, they're insensitive to the price. That's not the same as saying the demand for music shows no price elasticity; Luby acknowledges that sales of some titles and types of music may increase disproportionately to a cut in price. But lower prices won't stimulate demand from fans of a band or a particular song, Luby argued, because they've already made up their mind to buy. In fact, he believes that labels should be raising the price of popular singles in order to generate more album sales. The right price might be $1.49 or even $1.99, Luby said. "If I've got to spend $1.99" to get each of two singles from an album, he said, "I might as well just make the leap and buy the album."

As the post later explains, the inflexibility of single track pricing and the lack of experimentation do not allow for any conclusions to be drawn. In theory, though, labels could both enable music discovery and encourage album sales if given the flexibility (and the technology, would would assume). "You can't say `Discovery' and push that on the one side," said Lubby, "and then have a strategy that is forcing people to exclusively buy an album." Agreed. On the surface it may appear that MySpace streams and other free alternatives (imeem, last.fm) could allow consumers to sample the single (which they have already heard since it is Top 10 hit). But downloads allow people to live with a song for a while, give it portability and allow a person to better judge the artist.

If hit singles did cost $1.99, would demand drop enough that labels and artists would collect less revenue? With standard wholesale and mechanical rates, sales of a $1.99 hit single would have to drop more than 46.6% for total revenue to be less than if the single was priced at $0.99. If the higher price encouraged more album sales, as Lubby suggested may happen, an even further decrease would be acceptable. The main problem with that strategy, with maximizing revenue at the expense of unit sales, is that more fans are usually better than fewer fans (I say usually because even a small yet fervent fan base can be lucrative). Ticket sales, merch and other revenue streams could be negatively affected. Maybe those marginal fans wouldn't bring much additional revenue, but again, more fans are better than fewer fans.

By the way, if you're into pricing strategies, I have read parts of "Pricing on Purpose" and thought it was very good and very insightful.