♦ BLOG: Read a response from Chris Anderson, author of The Long Tail, and participate in this online discussion. ♦ BLOG: Read Anita Elberse’s response to Chris Anderson and participate in this online discussion. In a typical year, Grand Central Publishing (formerly Warner Books) goes to market with 275 to 300 book titles spread across two catalogs—its fall and winter lists. For each list the company identifies the handful of books it believes have the greatest sales potential and gives them the full benefit of its marketing capabilities. Of those, it spotlights just two “make” books, one fiction and one nonfiction, for which the company’s publisher is willing, in her words, to “pull out all the stops.” In the fall of 2007 those books were David Baldacci’s Stone Cold and Stephen Colbert’s I Am America (and So Can You!). The effects of this strategy show up in sales figures and profits. Whereas the 61 hardcover titles Grand Central put on its 2006 front list, on average, incurred costs of $650,000 and earned gross profits of just under $100,000, a wide range of numbers contributed to those averages. Grand Central’s most heavily marketed title incurred costs of $7 million and achieved net sales of just under $12 million, for a gross profit of nearly $5 million—50 times the average. Grand Central is pursuing what is known as a blockbuster strategy—a time-honored approach, particularly in the media and entertainment sector. With limited space on store shelves and in traditional distribution channels, and with retailers and distributors seeking to maximize their returns, producers have tended to focus their marketing resources on a small number of likely best sellers. Although such an approach involves substantial risk, they expect that the occasional hit’s huge pay-off will cover the losses of many misses, and that a few big sellers will bring in the lion’s share of revenues and profits. In 2006 just 20% of Grand Central’s titles accounted for roughly 80% of its sales and an even larger share of its profits. Much has changed in commerce, however, in the decades since the blockbuster strategy first took hold. Today we live in a world of ubiquitous information and communication technology, where retailers have virtually infinite shelf space and consumers can search through innumerable options. When books, movies, and music are digitized and therefore cheap to replicate, the question arises: Is a blockbuster strategy still effective? One school of thought says yes. Well represented by the economists Robert Frank and Philip Cook, in their 1995 book The Winner-Take-All Society, that school argues that broad, fast communication and easy replication create dynamics whereby popular products become disproportionately profitable for suppliers, and customers become even likelier to converge in their tastes and buying habits. The authors offer three reasons for their view: First and foremost, lesser talent is a poor substitute for greater talent. Why, for example, would people listen to the world’s second-best recording of Carmen when the best is readily available? Thus even a tiny advantage over competitors can be rewarded by an avalanche of market share. Second, people are inherently social, and therefore find value in listening to the same music and watching the same movies that others do. Third, when the marginal cost of reproducing and distributing products is low—as it certainly is with goods that can be digitized—the cost advantage of a brisk seller is huge. Frank and Cook were elaborating on the economist Sherwin Rosen’s earlier work describing the “superstars” effect, in which a field’s few top performers pull ever further away from the pack. According to this line of thought, hits will keep coming—to the increasing detriment of also-rans. Although that thesis continues to hold sway, another idea has emerged in recent years—presented just as persuasively, and proposing the opposite. The “long tail” theory took shape in an article by Chris Anderson, editor of Wired magazine, which grew into the 2006 book The Long Tail: Why the Future of Business Is Selling Less of More. The book’s subtitle puts the strategic implications in a nutshell. Now that consumers can find and afford products more closely tailored to their individual tastes, Anderson believes, they will migrate away from homogenized hits. The wise company, therefore, will stop relying on blockbusters and focus on the profits to be made from the long tail—niche offerings that cannot profitably be provided through brick-and-mortar channels. (See the sidebar “The Long-Tail Theory in Short.”) The Long-Tail Theory in Short In The Long Tail: Why the Future of Business Is Selling Less of More (2006), Chris Anderson puts forth two distinct but related ideas. The first is that merchandise assortments are growing because when goods don’t have to be displayed on store shelves, physical and cost constraints on selection disappear. Search and recommendation tools can keep a selection’s vastness from overwhelming customers. In the diagram below, all possible offerings in an imagined product sector are ranked by their sales volume, with the gray part representing products that are unprofitable through brick-and-mortar channels. The long tail, in other words, reveals a previously untapped demand. For goods like music, video, and information, which can be digitized, distribution costs approach zero, so the tail can be extremely long. Apple’s iTunes Store lists millions of albums and songs; Amazon offers more than 250,000 albums, whereas even the largest off-line music stores typically stock only about 15,000. The goods in the long tail include former hits as well as true niche content. Most of the latter was never released through traditional distribution channels or consists of orphans of unbundling activity (as with individual tracks in the music industry). Anderson’s second idea is that online channels actually change the shape of the demand curve, because consumers value niche products geared to their particular interests more than they value products designed for mass appeal. As internet retailing enables them to find more of the former, their purchasing will change accordingly. In other words, the tail will steadily grow not only longer, as more obscure products are made available, but also fatter, as consumers discover products better suited to their tastes. Anderson believes that obscure products will erode the immense share traditionally enjoyed by a relatively small number of hits. He predicts that “fickle customers” will “scatter to the winds as markets fragment into countless niches.” A lot of small sales put together, however, can add up to something big. In fact, Anderson boldly forecasts that the many small markets in goods that don’t individually sell well enough for traditional retail and broadcast distribution will together exceed the size of the existing market in goods that do cross that economic bar. In other words, the shaded area under the curve will become bigger than the white area over time. Which phenomenon is actually playing out in today’s markets? To find out, I investigated sales patterns in the music and home-video industries—two markets that Anderson and others frequently hold up as examples of the long-tail theory in action. Specifically, I reviewed sales data obtained from Nielsen VideoScan and Nielsen SoundScan, which monitor weekly purchases of videos and music through online and off-line retailers; from Quickflix, an Australian DVD-by-mail rental service; and from Rhapsody, an online music service that allows subscribers to choose from a large database of songs for a fixed monthly fee (and which Anderson cites often in The Long Tail). What I discovered may be of intellectual interest to readers who can relate both theories to their own consumption experience and appreciate the tension between them. But for managers whose job it is to navigate the digital landscape, the interest will be far more than academic. If you are a producer, you have pressing decisions to make about product development and marketing investments. If you are a retailer, you must decide how broad an assortment to stock and whether to guide customers toward obscure selections that may yield higher margins. In either case, your choices will vary dramatically depending on which theory you subscribe to. You won’t make the right calls unless you understand how online channels are actually changing markets. The Shape of Consumption When selection is vast and search easy, how do sales volumes stack up? Do they skew toward the head of the distribution curve or toward the tail? Rhapsody’s transaction record is a good place to find out. The first chart below depicts the aggregate selections of more than 60,000 subscribers who had more than 1 million tracks to choose from. In the three-month period of 2006 portrayed here, those customers engaged in more than 32 million transactions, or “plays.” And what do we see? Clearly, a high level of concentration. The data underlying the graph reveal that the top 10% of titles accounted for 78% of all plays, and the top 1% of titles for 32% of all plays. Pause for a moment, though, to reflect on those numbers. One percent of a million is still 10,000—far more than the number of titles most U.S. radio stations play in a given year, and when translated into album terms, equal to the entire music inventory of a typical Wal-Mart store. A Snapshot of Niche Products The second chart shows the sales distribution during a six-month period in 2006 at Quickflix, which offers just under 16,000 titles. Here the top 10% of DVDs accounted for 48% of all rentals, and the top 1% for 18% of all rentals. In other words, some 150 titles (roughly the number of movies released annually to theaters by major Hollywood studios) accounted for nearly a fifth of all rentals. The concentration is not as strong as with Rhapsody, but it’s still substantial. The charts provide a snapshot of the value of niche products. Strategists, however, need to understand how the picture is changing. As demand shifts from off-line retailers with limited shelf space to online channels with much larger assortments, is the tail of the sales distribution getting longer and fatter? My colleague Felix Oberholzer-Gee and I studied this question. In particular, we looked at weekly sales of home videos as reported by Nielsen VideoScan from January 2000 to August 2005, focusing on a random sample of nearly 5,500 titles. Using econometric models that control for a number of possible concomitant trends, we found that sales did shift measurably into the tail: The number of titles that sold only a few copies almost doubled for any given week from 2000 to 2005. In the same period, however, the number of titles with no sales at all in a given week quadrupled. Thus the tail represents a rapidly increasing number of titles that sell very rarely or never. Rather than bulking up, the tail is becoming much longer and flatter. Moreover, we determined that this is not simply a function of the sharp increase in the number of titles that have come onto the market in recent years, or of the transition from VHS to DVD; it is the truth of the long tail. Rather than bulking up, the tail is becoming much longer and flatter. That is the truth of the long tail. Meanwhile, our research also showed that success is concentrated in ever fewer best-selling titles at the head of the distribution curve. From 2000 to 2005 the number of titles in the top 10% of weekly sales dropped by more than 50%—an increase in concentration that is common in winner-take-all markets. The importance of individual best sellers is not diminishing over time. It is growing. Similar trends are obvious in the recorded-music industry. Here I have done research on physical and digital music sales from January 2005 to April 2007 for a random sample of 3,300 artists, including the pop sensations Justin Timberlake and Maroon 5 but also the far less widely known jazz saxophonist Kirk Whalum and the indie rock band The Dears. The data, collected by Nielsen SoundScan, show a period of rapid change, as digital units jumped from one-third to nearly two-thirds of the total sold. My research reveals a shift toward the tail of the sales distribution, and not surprisingly, the change is more pronounced for digital tracks and albums than for physical ones. However, the concentration in digital-track sales is significantly stronger than that in physical-album sales, as we see in the chart below; and as the share of digital units grows month by month, so does the degree of concentration in sales. The tail again lengthens but flattens, and although today’s hits may no longer reach the sales volumes typical of the pre-piracy era, an ever smaller set of top titles continues to account for a large chunk of the overall demand for music. Recorded-Music Sales When I differentiate between artists on smaller, independent labels and those on major labels, I find that the former gain some market share at the tail end of the curve as a result of the shift to digital markets. However, that advantage quickly disappears as we move up the curve: A more significant trend is that independent artists have actually lost share among the more popular titles to superstar artists on the major labels. (These results hold when I control for the number and type of titles that artists brought to market.) Thus digital channels may be further strengthening the position of a select group of winners. A Taste for Obscurity? When we look at customer-transactions data, the trends noted above take on greater meaning. It’s vital for marketers to understand who is responsible for the growing volume of business we see in the tail. Is just a small group of fanatics driving the demand for obscure products? If so, it is unlikely that a truly significant shift in media consumption will take place. Or are large numbers of consumers regularly venturing into the long tail? If so, it’s important to gauge the size of their appetite for those products and the degree of their satisfaction with them. The patterns that emerge in my research suggest that we won’t soon leave what Anderson calls “the water cooler era.” These patterns are far from new: They were described by William McPhee in the early 1960s, in Formal Theories of Mass Behavior. McPhee’s “theory of exposure” (see the sidebar “Consumers in the Head and Tail”) offers two relevant empirical generalizations: First, that a disproportionately large share of the audience for popular products consists of relatively light consumers, whereas a disproportionately large share of the audience for obscure products consists of relatively heavy consumers; and second, that consumers of obscure products generally appreciate them less than they do popular products. McPhee explored his theories in settings that typically provided fewer than a dozen alternatives. But my research reveals that his findings also hold true for the enormous assortments found online, even when sophisticated recommendation engines aim to stimulate demand for long-tail products. Consumers in the Head and Tail: McPhee’s Theory of Exposure Chris Anderson’s thinking about the appeal of the long tail is in many ways directly at odds with the sociologist William McPhee’s groundbreaking theory of exposure. In his 1963 book Formal Theories of Mass Behavior, McPhee describes two phenomena of distribution: natural monopoly and double jeopardy. Natural monopoly. McPhee observed that not only does the most popular product “[get] more raw numbers of people of otherwise marginal participation in the field,” but “a disproportionate share of its audience (a larger fraction of its already larger audience) consists of just such marginal people.” In other words, light users of a product category are a relatively large proportion of those customers interested in the popular products. Because it seems that hit products “monopolize” light consumers, he called the phenomenon a natural monopoly. Double jeopardy. McPhee noted that “the larger the proportion of the people [unfamiliar] with a given alternative,… the less likely are those who are familiar with it to like it especially.” Thus, he said, although we might believe that “the out-of-the-way book is at least a delight to those who find it,” in reality, the more obscure a title, the less likely it is to be appreciated. The people who choose obscure products tend to be familiar with many alternatives; those who know of few alternatives tend to stick with popular products. McPhee described this concept as double jeopardy because niche products have a double disadvantage: First, they are not well known; second, when they become known, it is by people who “know better” and prefer the popular products. McPhee’s theory of exposure finds support in my analysis of customer transactions for the online movie-rental service Quickflix. The exhibit below tells the tale. The bars of the chart show that the higher the decile, of course, the more customers rent titles within it. Note that the average number of titles shipped is much higher for customers of titles in the lowest decile than for customers of titles in the highest. This corresponds with McPhee’s natural monopoly: Hit products “naturally monopolize” light consumers. Heavy users are more likely to venture into the long tail, but they choose a mix of hit and obscure products. Observe that the average customer rating is higher in the higher deciles: In line with double jeopardy, obscure titles, on average, are appreciated less than popular titles as well as chosen less often. Is most of the business in the long tail being generated by a bunch of iconoclasts determined to march to different drummers? The answer is a definite no. My results show that a large number of customers occasionally select obscure offerings that, given their consumption rank and the average assortment size of off-line retailers, are probably not available in brick-and-mortar stores. Meanwhile, consumers of the most obscure content are also buying the hits. Although they choose products of widely varying popularity, top titles generally form the largest share of their choices. (The wide appeal of these top titles is, of course, what makes them popular in the first place.) Is most of the business in the long tail generated by a bunch of iconoclasts determined to march to different drummers? No. The exhibit “The Shopping Carts of Quickflix Video Customers” depicts this finding in greater detail. First, look at the bar farthest to the right. This is the breakdown of selections made by the large group of Quickflix customers who each rented at least one of the company’s most popular (top decile) DVD titles over a six-month period. On average, 61% of these customers’ rentals came from the highest decile, and another 13% from the second highest. Less than 1% of their choices came from the lowest decile—the most obscure titles. The Shopping Carts of Quickflix Video Customers Each vertical bar represents a decile of DVD popularity; DVDs in decile 10 are the most popular. Each bar is subdivided to show how, on average, customers who rent at least one DVD from within that decile distribute their rentals over all the deciles. Customers who shop in the bottom decile, for example, choose only 8% of their rentals from among its titles—and 34% from among top-decile titles. Now look at the bar farthest to the left, which represents the much smaller group of customers who rented at least one of the most obscure titles. On average, 8% of this group’s rentals came from this decile. But still, the biggest share—more than a third—came from the top decile. Another finding from the transaction data is that customers who rent obscure movies are in general the heaviest users of the service. On average, customers who chose a popular title rented 20 videos in the six-month period under review. However, customers who dipped into the most obscure offerings rented an average of 50. The implication is that there is no segment with a particular taste for the obscure; rather, customers with a large capacity for content venture into the tail. Meanwhile, also in line with McPhee’s theory, light consumers concentrate largely on the hit products. My research also answers the question, How much enjoyment is derived from obscure versus blockbuster products? We can all easily imagine the extreme delight that comes from discovering a rare gem, perfectly tailored to our interests and ours to bestow on likeminded friends. This is perhaps the most romanticized aspect of long-tail thinking. Many of us have experienced just such moments; they are what give Chris Anderson’s claims such resonance. The problem is that for every industrial designer who blissfully stumbles across the films of Charles and Ray Eames, untold numbers of families are subjecting themselves to the likes of Sherlock: Undercover Dog. Ratings posted by Quickflix customers show that obscure titles, on average, are appreciated less than popular titles. Skeptical readers might say this is only to be expected; heavier consumers, having seen it all, would be more critical across the board. It is true that these consumers give disproportionately low ratings to obscure products, but they also give disproportionately high ratings to hits; they have a somewhat bigger range in their scores than lighter consumers do. It could be that they are connoisseurs of a category and are better at distinguishing superior products from middle-of-the-road content. Consumption of long-tail offerings is more prevalent among people who tend to stick to a genre—classic rock and roll, for example, or romantic comedies. Their greater familiarity with alternatives may elevate ratings for superior popular products and lower those for inferior obscure ones. Other explanations can be theorized, but the fact remains: No matter how I slice and dice the customer base, customers give lower ratings to obscure titles. A balanced picture emerges of the impact of online channels on market demand: Hit products remain dominant, even among consumers who venture deep into the tail. Hit products are also liked better than obscure products. It is a myth that obscure books, films, and songs are treasured. What consumers buy in internet channels is much the same as what they have always bought. Implications for Strategy Soon after The Long Tail was published, BusinessWeek declared that Chris Anderson’s theory was the biggest idea of the year. The book was widely read, and its title entered the management vernacular. Anderson has spoken to numerous management audiences about its implications. All this has had an impact on practice: The long-tail theory increasingly influences the development and appraisal of business models, particularly in the media and entertainment sector. It is undeniable that online commerce has significantly broadened customers’ access to products of all varieties, including the most obscure. However, my findings suggest that it would be imprudent for companies to upend traditional practice and focus on the demand for obscure products. The data show how difficult it is to profit from the tail. What, then, are the implications of my research for practice? I have four recommendations for producers of media and entertainment goods, and four for online retailers or content aggregators seeking to profit from long-tail demand. Although my research has focused on media content and information goods, these recommendations probably apply to physical goods as well. In fact, their payoff for manufacturers and retailers of physical goods might be bigger, because of the higher production costs involved. Advice to Producers 1. Don’t radically alter blockbuster resource-allocation or product-portfolio management strategies. A few winners will still go a long way—probably even further than before. My research suggests that the tail is long and flat, and therefore that content providers will find it hard to profit much from it. It remains to be seen whether the new media environment will indeed make many previously unprofitable niche products profitable. Online channels lower the barriers to market entry for such products, and thus introduce the possibility of additional sales—but they also lead to a flood of products all competing for consumers’ attention. In my most recent correspondence with managers at Nielsen SoundScan, I learned that of the 3.9 million digital tracks sold in 2007 (the large majority for 99 cents each through Apple iTunes), an astonishing 24% sold only one copy, and 91%—3.6 million tracks—sold fewer than 100 copies. Although increased concentration of sales may make it tougher to turn a focus on blockbusters into a winning strategy, no effective alternative strategy is readily available. 2. When producing niche goods for the tail end of the distribution, keep costs as low as possible. Your odds of success aren’t favorable here either, and they will probably become less so. The extremely low demand for the large array of products in the tail means that simply recovering the costs of producing them will be challenging. Given that obscure products tend to be appreciated less than hits, it will be very difficult to earn any kind of price premium for them. 3. When trying to strengthen your presence in digital channels, focus on marketing your most popular products. By definition, they reach the largest number of customers, and they are also appreciated more by those who consume them. This insight is perhaps particularly relevant for content providers competing in advertising-supported markets. Advertisers hoping to reach a broad cross-section of consumers in a world of proliferating media are better off placing ads around popular products; not only will their messages be seen more often, but, because those products are generally liked better, they will be seen in a favorable context. Hit products may therefore have a disproportionately high value. No wonder, then, that large media companies increasingly insist on more control over pricing and bundling decisions involving their most popular offerings. NBC’s recent spat with iTunes is one example. 4. Leverage your scale to improve online exposure and demand for products across your product portfolio. Again, hit products play a key role here. The long tail consists of a mixture of true niche products (which, by Anderson’s definition, do not meet the bar for traditional distribution) and old hits resulting from blockbuster-focused strategies. Such products can now live forever online, even if they have long been cleared from physical shelves; thus the old hits may present a real opportunity. Larger producers have an advantage in that they can use new releases to trigger demand for old ones—previous movies in which a cast member appeared, for example, or earlier recordings by an up-and-coming artist. Companies can benefit from finding ways to regularly remarket products in their back catalogs and from bundling old with newer products. The caveat here, again, is that the benefits may not outweigh the costs. Music companies, for instance, often decline to make old content available online because clearing the rights is too cumbersome. Similarly, although channel partnerships frequently prevent companies from leveraging their scale (Apple’s iTunes often gives relatively more promotional space to artists from independent record labels than to those from the majors, for instance), companies can use their hit products to negotiate better terms with channel partners. Larger, better-established firms with strong pipelines should therefore benefit more than smaller companies from any increased demand in online channels. Advice to Retailers 1. If the goal is to cater to your heavy customers, broaden your assortment with more niche products. My research shows that even when online assortments of videos and music are enormous, and thus even the most frequent customers could easily satisfy their appetites with products in the top decile, those customers are disproportionately active in the tail. They want a wide assortment, so offering one helps attract and retain them, whether they pay by the product or for a subscription (frequent customers typically opt for more-expensive subscription plans). 2. Strictly manage the costs of offering products that will rarely sell. If possible, use online networks to construct creative models in which you incur no costs unless the customer actually initiates a transaction. Managing a large number of products that rarely or never sell could easily pose a problem. Long-tail products may offer more-attractive profit margins for retailers than hit products do—in part because the latter are often used as loss leaders. But extremely low demand for long-tail products, coupled with whatever it costs to make them available, presents difficulties in successfully executing a long-tail model. Making “onesies” and “twosies” profitable may require completely eliminating any associated costs. It is therefore worthwhile to explore creative solutions for the very end of the tail. One example is Amazon’s Marketplace, in which third parties pay to communicate a title’s availability, and Amazon incurs costs only when a customer actually places an order. Another is having volunteers create, adapt, and manage information in web businesses. Just imagine if Wikipedia paid authors for every page created. Even if the fee were nominal, Wikipedia would probably lose a substantial sum on its least visited pages. 3. Acquire and manage customers by using your most popular products. Precisely because hit products reach the greatest number of consumers and are appreciated most, their value as loss leaders in traditional channels will carry over into the digital realm. The seventh book in the Harry Potter series, introduced by Scholastic at a suggested retail price of $34.99 in the United States, was a blockbuster loss leader: It was sold at sharply reduced prices by Barnes & Noble ($20.99, a 40% discount) and Amazon ($17.99, a 49% discount) in an effort to stimulate other purchases. Like producers, online retailers can benefit from bundling hit products with obscure or older products that are cheaper to acquire. Another, probably more common approach is to direct customers to the tail with recommendation engines. A third strategy worth considering is designing the flow of web pages so that consumers, even those searching for hit products, are naturally directed into the tail. The list of recommended titles can be manipulated—often instantly and cheaply—to spotlight higher-margin obscure items or to smooth demand for sought-after titles over time. 4. Even though obscure products may have a higher profit margin, resist the temptation to direct customers to the tail too often, or you’ll risk their dissatisfaction. Finding a good marketing balance between obscure and popular products is critical. Online retailers cannot expect their customers to prefer long-tail products to hits—in fact, the opposite is more likely. They should take this into account when managing customer expectations and satisfaction—which, after all, lead to long-term profitability. The continued dominance of hit products and the natural shape of demand suggest that efforts to fatten the tail by spreading consumption more evenly across titles may be fruitless anyway. Who Will Prosper? Without question, today’s consumers have advantages that no prior generation had. Online commerce has done away with the constraints of the physical store; selections are now vast and supported by rich information. A hip-hop fan just discovering the lyrical talents of Jay-Z need not be limited to his recent hits; she can follow him all the way to his first album, Reasonable Doubt (1996), which had only modest sales, and she can easily jump to Talib Kweli and other lesser-known contemporaries, some of whom may be available only in digital format. For Chris Anderson, the strategic implications of the digital environment seem clear. “The companies that will prosper,” he declares, “will be those that switch out of lowest-common-denominator mode and figure out how to address niches.” But my research indicates otherwise. Although no one disputes the lengthening of the tail (clearly, more obscure products are being made available for purchase every day), the tail is likely to be extremely flat and populated by titles that are mostly a diversion for consumers whose appetite for true blockbusters continues to grow. It is therefore highly disputable that much money can be made in the tail. In sales of both videos and recorded music—in many ways the perfect products to test the long-tail theory—we see that hits are and probably will remain dominant. That is the reality that should inform retailers as they struggle to offer their customers a satisfying assortment cost-efficiently. And it’s the unavoidable challenge to producers. The companies that will prosper are the ones most capable of capitalizing on individual best sellers. How appropriate that proof of this can also be found in management literature. Over the course of 2006, Hyperion Books, which publishes adult trade fiction and nonfiction, brought dozens of original hardcovers to market. For a handful of them it spent heavily on acquisition and marketing, hoping for the profits that only blockbusters can provide. One was Mitch Albom’s novel For One More Day, which became the single best-selling hardcover of 2006. Another was a business title that had engendered an intense bidding war. Hyperion was determined to get it; New York magazine quoted an industry insider as saying that “jaws hit the floor over how much they paid.” Everyone recognized it as a high-stakes gamble in a high-risk genre. But ultimately it paid off big. It was, of course, The Long Tail.
