November 14, 2009

By GLENN PEOPLES

The great hope for digital music was that it would make the recording industry more egalitarian—that up-and-coming bands with pluck and a knack for promotion would be able to get their work to the masses without the backing of record labels. According to "The Long Tail: Why the Future of Business Is Selling Less of More"—a 2006 book by Wired magazine editor in chief Chris Anderson—hits dominated the market mostly because shelf space in stores was limited. Digital retail and online media would exponentially increase the choices available to consumers, who would then use online tools to discover products that appealed to them more than the biggest hits.Anderson's "Long Tail" idea comes from a sales graph that looks like the letter "L" with a curve instead of a corner. On the left are the hits, the 5,000 best-selling titles that would typically be carried by a national chain; on the right, further down the curve, are less popular titles that sell fewer copies. In the physical world, few stores have space for these niche titles, which don't sell well. But in the digital world, where space hardly matters, Anderson suggested, these titles would collectively account for a far greater percentage of music sales—and of movies, books and other consumer products. The ways we think about popular taste, he writes, "are actually artifacts of poor supply-and-demand matching—a market response to inefficient distribution."

For an industry that coined the term "hit parade," this would amount to nothing short of a revolution.

So far, at least according to Nielsen SoundScan data on U.S. music sales from January 2004 through October 2009, that revolution hasn't arrived—although the demand for albums has changed. Sales of albums, especially digital ones, became significantly less concentrated around hit releases since 2004. But sales of digital tracks—which this year account for 56% of digital sales by track volume—have grown more concentrated in hits during the same time period.

Essentially, hit songs are becoming more important while hit albums are becoming less so.

Although "The Long Tail" discusses the waning prominence of hits, Anderson would prefer to look past the top of the head of hit singles at a larger group of tracks to gauge broad shifts in demand. "In short, this isn't enough data to draw any proper 'Long Tail' conclusions about," he wrote in an e-mail, "since it doesn't use Head and Tail the way the theory does."

Since the publication of "The Long Tail," some studies have confirmed the book's thesis, while others have cast doubt on it. In a 2008 paper, Harvard Business School associate professor Anita Elberse found that hit titles still dominated sales even though some consumers were venturing further down the tail. This year, two researchers at the Wharton School of the University of Pennsylvania, Tom F. Tan and Serguei Netessine, examined Netflix user data from 2000 to 2005 and found that new titles are appearing faster than customers can discover them. Perhaps more surprisingly, a study by PRS for Music chief economist Will Page and BigChampagne CEO Eric Garland found that the demand for songs on file-sharing services—which offer users almost unlimited choice—closely mirrors that of purchased tracks. Only 5% of songs accounted for 80% of downloads, resulting in what the authors called a "hit-heavy, skinny-tail distribution."

One thing that hasn't changed since the publication of "The Long Tail" is how hard it is for artists to sell a meaningful amount of music—whether or not they're signed to a label. From 2004 to 2008, the number of new albums released per year has more than doubled. And although digital retail is taking market share from the most popular titles, the sheer number of unpopular albums available means that each of those titles doesn't benefit much from their collective increase in market share. The millions of units that are shifting from a few titles at the head of the tail are migrating to a few hundred thousand at the end of it—each of which doesn't sell much more.

THE POWER OF HIT SONGS

So far digital retail is less about albums than individual songs, which account for 56% of all purchased tracks so far in 2009. And in the last five years, track sales have become increasingly more concentrated, so that hits matter more each year. This trend was first noticed by Elberse, who pointed out that it was happening even as the number of tracks available continues to increase.

The change is significant. From 2004 through October 2009, the most popular tracks have steadily and consistently grabbed market share—and tens of millions in unit sales—from less popular songs. The growth is slight at the top of the chart and more noticeable further down. The top 10 increased to 3.1% from 2.1%. The top 40 increased to 8.3% from 5.9%. And so on. The top 200 tracks—that's just 0.002% of the nearly 9 million currently listed at Amazon—have a market share of 18.7%. In 2004, their share was 14.5%.

At a time when more music is available than ever before, why do so many consumers buy the same few songs? It may be because popular taste tends to reinforce itself, especially in an online world. Or it may be because buyers of single tracks tend to be casual fans who are more inclined to buy songs they hear on the radio and TV.

"One aspect of word-of-mouth online is that it can be an effective discovery technique, driving demand to titles that don't have traditional marketing," Anderson said. "But the other side of it is that it can lead to herd behavior, with 'winner take all' effects. It's possible for both to work at the same time, with some word-of-mouth boosting niche acts, while other word-of-mouth creates bigger hits at the very top of the curve."





Weekly sales figures show just how important hits have become. Since iTunes launched variable pricing in early April, the top 200 tracks have retained their market share even as the number of tracks purchased each week has fallen by about 6%. (Some of this decline may be due to a midyear weakness in digital track sales.) From April to July, the top 200 averaged a 24% share of each week's total track sales. In the same period in 2008, the top 200 averaged just 22.2% of each week's track sales, even though most of the top 200 songs were less expensive.

Although higher prices have depressed sales of hits—in terms of units, not revenue—consumers haven't been spending their money on other songs. "The Long Tail" suggests that consumers will use increasingly sophisticated digital tools to discover, sample and buy music that appeals to them more than the biggest hits. But in the case of digital tracks, that hasn't happened. Consumers who are turned off by a $1.29 price point for the track they came to buy don't seem to seek out less popular alternatives. In other words, many music fans aren't shunning hits because they don't like them but because the price rose by 30 cents. And if they don't find the hits they want, they forgo a music purchase altogether.

THE LONG TAIL OF ALBUMS

Overall, album sales don't look that different from five years ago, at least in terms of the demand curve. The most significant change has been the overall decline in album sales: 32% from 2004 to 2008. As everyone in the industry knows, it's tough all over.

As "The Long Tail" predicts, the most popular albums fared the worst, losing market share to less popular titles. From 2004 to 2008, sales of the 5,000 albums that make up the head of the demand curve dropped 40.5% while sales of the million-plus albums that make up the tail declined 27.4% And not only did sales of popular albums decline more than those of others, the most popular ones declined the most. Unit sales of the top 1,000 albums of 2008 dropped 41.7% from their 2004 levels. The second thousand most popular albums dropped 36%, the third thousand fell 33.2%, the fourth 31.2% and the fifth 30.9%. Five years ago, the top 5,000 albums represented 74.4% of total sales; in 2008 they accounted for 70.2%. Some of this comes from the sheer number of albums that now make up the end of the tail.





The marketplace for digital albums is also taking shape according to the theories in "The Long Tail." At a time when the big-box retailers that now account for so much of the CD market have cut the shelf space they devote to music, the number of tracks available from online services keeps rising. And digital retailers make it easier for consumers to sample music and use various other filters and discovery tools that "The Long Tail" predicted would distract them from the hits.

Just as "The Long Tail" indicated, demand for digital albums is moving further down the tail than that for albums overall. In 2008, the top 5,000 albums accounted for 64.7% of digital album sales, as opposed to 70.2% of album sales overall. And the market share of the top 5,000 digital albums is shrinking as niche products take away sales from more popular titles. But the rate of change is slowing—the head lost three percentage points in 2007 and 2008 after shedding more than six points in 2006, which means that the demand curve could settle in something close to its current shape.





Within the head of the demand curve, which represents a wide range of popularity, albums have been affected very differently. In 2006 and 2007, the most popular 100 albums lost the most market share in absolute terms. The three percentage points of market share they lost represented about 1 million units. In 2008, albums from No. 101 to No. 200 lost the most share in absolute terms. But in relative terms, the albums in the middle of the head fared the worst in terms of losing share. From 2004 to 2008, albums from No. 301 to No. 400 lost the greatest percent of their market share—34%. From 2005 to 2008, Nos. 401-500 suffered the most—22%. From 2006 to 2008, albums as far down as No. 4,000 lost a greater percent of their market share (7%) then the top 100 ranks (5%).

As "The Long Tail" predicted, sales will disperse across a wider range of titles as consumer choice increases. Within the tail of digital albums, the truly obscure albums seem to be pulling sales away from those that are merely unpopular. In 2008, even as the head of the tail shrunk more slowly than in previous years, albums as unpopular as those around No. 8,000 gave up market share to titles that were even less popular.

LIFE IN THE TAIL

Life in the long tail can be difficult for any individual artist. One such album—former Afghan Whigs frontman Greg Dulli's "Live at the Triple Door"—sold 1,400 digital copies in 2008 and ranked at No. 6,736. As "The Long Tail" would have predicted, an album with that sales rank benefitted from the effects of widespread digital distribution. But during the past three years the gain for an album at No. 6,736 was nil: around 75 additional copies.

Dulli is the type of artist who might be expected to benefit from the economics of the long tail. Signed to a major label in the early '90s, the Afghan Whigs released some moderately successful albums, and Dulli has enjoyed similar success as a solo artist. But now Dulli and artists like him now face more competition, simply because so many more albums come out each year—"Live at the Triple Door" was one of 50,000 digital-only albums released in 2008. Even if Dulli keeps cracking the top 10,000 albums, his market share is likely to be smaller than what it is today. And market share is important because it influences other revenue streams, such as touring and merchandise sales.

In terms of overall album sales—not just those of digital albums—the greatest changes may be taking place in what might be called the middle class: albums ranked from No. 200 to No. 2,000 in terms of sales. Many of these are catalog titles that benefited from year-round price-and-positioning programs at retailers like Virgin Megastore and fye. Sales of these albums dropped as much as 34% from 2006 through 2008, compared with the 27% decline in overall sales.

Most likely because so many music stores closed, catalog chestnuts like the Phil Collins collection "Hits" have stayed close to their overall sales rank while selling far fewer units. In 2006, "Hits" sold 116,000 copies, enough to rank at No. 699 among the best-selling albums of the year. By 2008, "Hits" sold 82,000—a 29% drop—but ranked at No. 703.

In the digital world, which relies less on merchandising programs, "Hits" is all but absent: It hasn't cracked the list of the top 10,000 digital albums since 2006. Bargain catalog makes an appealing impulse buy at physical stores, and since many retailers can't carry all of Collins' albums, they focused on a hits collection. In the digital world, consumers have many more options for Collins' catalog. In addition to "Hits," shoppers can choose from his studio albums like "No Jacket Required" (No. 3,273) and "But Seriously" (No. 9,652) or buy their favorite tracks individually.

ARE THERE RICHES IN NICHES?

So how can music companies adapt to this new world? "The Long Tail" urges businesses to "think niche." Since the future is "selling less of more," it makes sense to make available every product possible. And since niche titles are rarely discounted, Anderson argues, online retailers like Amazon are wise to use recommendation engines to subtly nudge their consumers toward relatively unpopular items.

At the time "The Long Tail" came out, this was smart advice—and it still is for online retailers like Amazon and Netflix, which sell physical goods. For years big-box retailers like Wal-Mart have used popular CDs as loss leaders to drive sales of more expensive, high-margin products. So far, however, pure digital retailers work differently, and iTunes, which typically has the highest prices of any online music store, still has the highest market share. (iTunes also sells more expensive versions of some albums.) And as long as the most popular titles command the highest prices, as they now do on iTunes, retailers would be wise to steer consumers toward them in order to maximize revenue and, presumably, margins.

There is also evidence that a retailer could alienate consumers by steering them toward niche items that don't appeal as much as hits. Anderson wrote in the book that, as listeners stop buying CDs and explore the tail, they are "typically more satisfied with what they find." But Elberse studied user ratings at the Australian DVD rental service Quickflix and found that the more popular titles also received the most favorable ratings. Users who rented obscure titles tended to rate them less favorable than they did hits.

"No matter how I slice and dice the customer base, customers give lower ratings to obscure titles," she wrote in her article for the Harvard Business Review. "There are signs that if you keep pushing people into the tail because the economics for you are really good, that might actually hurt you in the long run." "That may be true for the specific example of the Australian DVD data," Anderson wrote on his blog, "but it is not clear from the paper why she feels able to extrapolate that to all Internet commerce." In their analysis of Netflix user ratings, Wharton's Tan and Netessine also found that consumers tend to be more satisfied by hits than niches.

While it's easy to see how retailers could adapt to the world of "The Long Tail," what about content creators and the companies that fund and market their work? Any label or artist that stopped trying for a hit in order to focus on a niche is almost certainly doing the wrong thing, at least in economic terms. Although niche titles collectively account for a greater percentage of sales, no individual one accrued any meaningful income—and few have received the attention their creators would need to perform or sell merchandise at a time when those revenue streams are becoming more important.

Major labels and independents that are run as serious businesses should continue to focus on how to reach a mass audience—especially on how they can do so using new digital tools and the advertising and sponsorships that are becoming increasingly important in the music businesses.

Indeed, labels have continued to focus on finding hits for a reason: It's almost impossible for them to make real money any other way. (Even if a company or act decides to give away music in order to play live or sell other goods, they still need to reach a significant audience to make that pay off.) Elberse, for one, doesn't think content companies should focus on hits any less than they do now. "I don't think they need to go about their job any differently now than they did 10 years ago," she told Billboard. "They will still bet on a few projects more than other projects in their portfolio and hope they will become the winners that pay for the majority of things that don't make a profit." ••••

Editor's note: This story was edited by Louis Hau and Robert Levine; Levine several years ago worked with Chris Anderson, the author of "The Long Tail," at Wired magazine.

Subscribe to Billboard magazine today!