The Long Tail

by Chris Anderson

  It may sound like metaphysics, but the best Long Tail markets transcend time and space. They’re not constrained by any geographic boundaries, nor do they make any assumptions about when people want what they want. ITunes’ advantage comes primarily from its huge variety and convenient downloads, but being open 24/7 doesn’t hurt either.

  Today, you can get CSI on broadcast TV, video-on-demand, iTunes download, DVD (purchase or rent), and TiVo season pass, and watch it on any device from a plasma screen to a Sony PSP. Likewise for some NPR radio shows, which you can get via terrestrial broadcast (real time and delayed), satellite broadcast, Web streaming, podcast, and, if you like, an emailed transcript. Multiple distribution channels are the only way to reach the biggest potential market.

  Rule 4: One product doesn’t fit all.

  Once upon a time, there was one way to buy music: the CD album (so few CD singles were sold that many artists didn’t bother offering them). Now consider the choice you have online: album, individual track, ringtone, free thirty-second sample, music video, remix, sample of somebody else’s remix, streamed or downloaded, all in any number of formats and sampling rates.

  Umair Haque calls this “microchunking.” Increasingly, the winning strategy is to separate content into its component parts (“microchunks”), so that people can consume it the way they want, as well as remix it with other content to create something new. Newspapers are microchunked into individual articles, which are in turn linked to by more specialist sites that create a different, often more focused, product out of content from multiple sources—the blogger as DJ, remixing the news to create something new.

  We’ve seen this before in the form of product and brand segmentation—a dozen different kinds of spaghetti sauce for a dozen distinct palates. Now that trend is being extended to everything from individual video-game characters and levels (mix your own game) to selling cookbook chunks one recipe at a time. Each recombination taps a different distribution network and reaches a different audience. One size fits one; many sizes fit many.

  Rule 5: One price doesn’t fit all.

  One of the best understood principles of microeconomics is the power of elastic pricing. Different people are willing to pay different prices for any number of reasons, from how much money they have to how much time they have. But just as there’s often room for just one version of a product in traditional markets, there’s often room for only one price, or at least one price at a time. In markets with room for abundant variety, however, variable pricing can be a powerful technique to maximize the value of a product and the size of the market.

  EBay, for instance, offers auctions (typically lower prices, but at the cost of greater hassle and uncertainty) or “Buy It Now” (higher prices). Even iTunes, which has stuck with a single price of $0.99 per track for simplicity’s sake, will give you a lower price if you buy the tracks as part of an album. Rhapsody has gone even farther, experimenting with track prices from $0.79 to $0.49, and finding that cutting the price in half roughly triples sales.

  The natural model for music and anything else where the marginal costs of manufacturing and distribution are close to zero is variable pricing. Retailers can charge more for the most popular items and less for the less popular. Why hasn’t that happened already? Because the labels typically charge a fixed wholesale price of around $0.70 per track, largely to avoid “channel conflict” with CDs, which still produce the bulk of the music business revenues. However, in 2007, acts like Radio-head and Prince experimented with distributing music for free or, in the case of the former, actually let customers set their own price. Someday the labels will see the light and pricing will become more fluid, allowing retailers to draw consumers down the Tail with lower prices.

  LOSE CONTROL

  Rule 6: Share information.

  The difference between an overwhelming shelf of look-alike products and the bliss of “rank by best-selling” is information. In one case, the store knows what sells best but doesn’t tell its customers. In the other, it does. So too for “rank by price,” “rank by review,” and “sort by manufacturer.” All that data already exists; the question is how best to share it with customers. More information is better, but only when it’s presented in a way that helps order choice, not confuse it further.

  Likewise, information about buying patterns, when transformed into recommendations, can be a powerful marketing tool. Deep information about products, from reviews to specifications, can answer questions that would have otherwise halted a purchase. Explaining why a consumer is getting a certain set of recommendations builds confidence in the system, and helps consumers use it better. Transparency can build trust at no cost.

  Rule 7: Think “and,” not “or.”

  One of the symptoms of scarcity thinking is assuming that markets are zero-sum. In other words, the mistake of assuming that everything is an either/or choice. Release this version or that version. Sell this color or that color. On shelves or broadcast channels, that’s natural enough: There really is room for only one product in any single slot. But in markets with infinite capacity, the right strategy is almost always to offer it all.

  The problem with choosing is that it requires discrimination, and the process of discrimination requires time, resources, and guesswork. Someone must decide, on the basis of some criteria, that one thing is likely to be more successful than another. They may be right at the macrolevel, but such a decision almost always gets it wrong at the microlevel. Consider the phenomenon of the “alternative ending” on DVD movies. Even if most people like the standard ending best, there are always some who prefer the alternative one. Now they can have both. Extend that to the foreign language option, the choice of standard or widescreen, and even the variety of cuts for various ratings (PG, PG-13, R, and uncensored)—each option has an audience, even if it’s not as big as the primary audience.

  All those “extras” are enabled by the abundant capacity of the DVD, allowing directors to “waste” storage with content that they could not have included in a more scarce medium, such as a theatrical screen or a videotape. So, too, for any digital market online, where the falling price and rising capacity of storage ensure that whatever capacity you want, it’s only a matter of time before it’s virtually free. The more abundant the storage and distribution, the less discriminating you have to be in how you use it. “And” is a far easier decision than “or.”

  Rule 8: Trust the market to do your job.

  In scarce markets, you’ve got to guess at what will sell. In abundant markets, you can simply throw everything out there and see what happens, letting the market sort it out. The difference between “pre-filtering” and “post-filtering” is predicting versus measuring, and the latter is invariably more accurate. Online markets are nothing if not highly efficient measures of the wisdom of crowds. Because they’re information-rich, it’s relatively easy for people to compare goods, and spread the word about what they like.

  Collaborative filters, for instance, are a market-based way to do product promotion. Popularity rankings are another voice of the market, amplified by the positive-feedback loop of word-of-mouth. And ratings are collective opinion, quantified in ways that make it easy to compare across products and sort them. All of these tools can order variety in ways that make sense to consumers, without some retailer having to guess at what will work. The lesson: Don’t predict; measure and respond.

  Rule 9: Understand the power of free.

  Free gets a bad rap, evoking piracy and other such evaporations of value. But one of the most powerful features of digital markets is that they put free within reach; because their costs are near zero, their prices can be, too. Already, one of the most common business models on the Internet, from Skype to Yahoo! Mail, is to attract lots of users with a free service and convince some of them to upgrade to a subscription-based “premium” one that adds higher quality or better features. Because digital services are so cheap to offer, the free customers cost the company so little that it can affo
rd to convert only a tiny fraction of them to paying customers.

  Samples, from thirty-second music clips to video previews, are possible because the cost of delivering bits on broadband pipes is so low. Video-game makers routinely distribute demos with a few free levels; if you like them, you can pay to unlock the others. In 2005, Universal released the first nine minutes of Serenity (a sci-fi film) online, free and uncut. Why? Because it could. The cost of delivering nearly 10 percent of a movie to anyone who wanted to watch it online was trivial compared to the marketing value of pulling an audience into the plot and leaving them with a cliffhanger, an itch that only a paid trip to the movie theater could scratch.

  Most television is already free and advertising-supported. Yet the networks still want to find ways to charge for it online, even though the production costs are paid for by broadcast, and online distribution costs are low. Why not give it away online as well, bookending (rather than interrupting) it with ads or just finding a bigger audience for the product placement, which can be neither stripped out nor fast-forwarded past? Ultimately, in abundant markets with loads of competition, prices tend to follow costs. And thanks to the power of digital economics, costs just get lower.

  15

  THE LONG TAIL OF MARKETING

  HOW TO SELL WHERE “SELLING” DOESN’T WORK

  In the early spring of 2006 a team of planners at the Campbell-Ewald advertising agency in the Detroit suburb of Warren watched Burger King’s ridiculous (and ridiculously successful) “Subservient Chicken” interactive Web ad and pondered how they might do something as cool as that for cars. Not just any car, but Chevy’s workhorse Tahoe SUV, which had lifted the fortunes of the carmaker’s parent company, GM, for nearly a decade but was now running headlong into $3-per-gallon gas.

  Campbell-Ewald had run Chevy’s advertising ever since Mr. Campbell and Mr. Ewald met Louis Chevrolet in 1914, and the firm was both keen to show that it was not stuck in the past and mindful of the legacy that it must live up to. But the one thing that was clear was that this time the usual solution—another overproduced TV car ad—would not do. For one thing, TV ads just aren’t working the way they used to. McKinsey, the consultancy, projects that by 2010 advertising on broadcast television will be barely one-third as effective as it was in 1990, thanks to rising costs, falling viewership, ever-proliferating ad clutter, and viewers’ TiVo-fueled power to zip through commercials. And then there were the small matters of the Iraq War, climate change, and a wave of green consciousness sweeping the country, all of which made selling a huge rolling display of oil consumerism an even trickier proposition than usual.

  The Campbell-Ewald team was hoping that the answer for the Tahoe lay on the Web. Not just because an online marketing campaign could be interactive, and thus potentially more engaging than a passive TV ad, but also because it could reach consumers who weren’t paying attention to TV ads at all anymore. The days when network television was a good proxy for all of America itself are gone. The sought-after demographic of males between eighteen and thirty-four is increasingly online, playing video games, or otherwise turning away from the broadcast schedule. Females of the same age are just a step behind them. And the more money you have, the more likely you are to be time-shifting your watching with a digital video recorder like TiVo or a cable set-top box—and probably skipping the commercial interruptions.

  Campbell-Ewald’s thinking went something like this: Chevy’s tagline is “An American Revolution.” So let’s do something, well, revolutionary. Basic interactive Web ads—steer the car with your mouse!—aren’t enough. Let’s take this thing up a notch—let’s have an online contest to see who can create the best TV ad for the new Tahoe. Chevy will supply the video clips and music—mostly standard fare of Tahoes sweeping along trafficless vistas, parking at the edge of spectacular cliffs, and otherwise delivering their drivers to landscapes of rugged beauty and presumably bumpy access. Users could then mix and match the material and add their own captions. The wikification of the thirty-second spot—what could be more revolutionary than that?

  Starting in March, the contest ran for four weeks. In spots that aired during a special episode of The Apprentice on TV, viewers were urged to go to a newly built microsite, Chevyapprentice.com, and create ads of their own using the video and simple editing tools posted there. The contest drew more than 30,000 entries, the vast majority of which faithfully touted the vehicle’s many selling points—its fully retractable seats, its power-lift gates, its relative—emphasis on “relative”—fuel economy. But then there were the rogue entries, the ones that subverted the Tahoe message with references to global warming, social irresponsibility, war in Iraq, and the psychosexual connotations of extremely large cars. Annoyingly, they were some of the most clever ones, too. “Like summer? Get used to it.” “Peak Oil is here. Maybe you should walk.” And, of course, “What would Jesus drive?” On its own Web site, the Tahoe now stood accused of everything from terminal squareness to crimes against humanity.

  What’s worse, the most subversive and funniest of these were now circulating on YouTube, where they were racking up hundreds of thousands of views—more than those on Chevy’s own site. Campbell-Ewald had gotten just want they wanted—a viral hit like the Subservient Chicken—but for all the wrong reasons. An online mob had co-opted the campaign and turned it against the product. And a network TV–sized audience was now gleefully dancing around the bonfire of a great American brand.

  It didn’t take long for bloggers and reporters to realize that something weird was going on over at Chevyapprentice.com. At first, everyone assumed it was just another case of a big corporation not “getting it” about the Internet. Then, when the ads weren’t yanked down immediately, they figured Chevy was too clueless even to notice what was happening on its own site. Only gradually did it dawn on people that Chevy had no intention of removing the attack ads.

  In fact, the Campbell-Ewald team had assumed all along that they’d get some negative responses, and they decided they’d lose all credibility if they pulled any of them down. Ed Peper, Chevy’s general manager, pointed out in a post on GM’s FastLane blog that the Tahoe can run on ethanol and gets better gas mileage than other large SUVs, but as far as Chevy was concerned, that was that.

  Within the advertising industry, consultants gleefully watched the riot and reassured their clients that this proved that you can’t turn over your brand to the great unwashed. But by any objective measure, the Tahoe Apprentice campaign has to be judged a success. The microsite attracted 629,000 visitors by the time the contest winner, Michael Thrams, from nearby Ann Arbor, was announced at the end of April. On average, those visitors spent more than nine minutes on the site, and nearly two-thirds of them went on to visit Chevy.com; for three weeks running, Chevyapprentice.com funneled more people to the Chevy site than either Google or Yahoo did. Once there, many requested info or left a cookie trail to dealers’ sites.

  Sales took off, too, even though it was spring and SUV purchases generally peak in late fall. In the first nine months after its introduction in January 2006, the new Tahoe accounted for more than a quarter of all full-sized SUVs sold, outpacing its nearest competitor, the Ford Expedition, two to one. In March, the month the campaign began, its market share hit nearly 30 percent. By April, according to auto-information service Edmonds, the average Tahoe was selling in only forty-six days—quite a change from the year before, when models languished on dealers’ lots for close to four months. Even the hooting among marketing pros died down after Scott Donaton, the editor of Advertising Age, asked in a column for a show of hands from all those who thought the campaign proved the dangers of user-created content. “Ah, yes,” he wrote, “there’s quite a few arms raised—you’re all free to go, actually; the marketing business doesn’t need your services anymore. We have a toy railroad set as your lovely parting gift.”

  Donaton was taking aim at a central tenet of “golden age” mass-media marketing: that by controlling the ad message, Madison Aven
ue can somehow control perception of the product. “When you do a consumer-generated campaign, you’re going to have some negative reaction,” said Ed Dilworth, a Campbell-Ewald executive. “But what’s the option—to stay closed? That’s not the future. And besides, do you think the consumer wasn’t talking about the Tahoe before?” They were, of course; the difference is that in the YouTube era, the illusion of control is no longer sustainable. “You can either stay in the bunker, or you can jump out there and try to participate,” he says. “And to not participate is criminal.”

  Why did the Tahoe ad work, despite having been hacked by haters? Well, for starters, the notion that SUVs are not a great way to save the environment was not news to their potential buyers. Indeed, it’s not impossible that the fact that huge vehicles drive crunchy green types mad may be part of the attraction for some of the Tahoe’s red-state base. And then there is the factor that Chevy, by starting the contest in the first place and then not taking the negative spots down, actually looked pretty cool. And for a stodgy American car brand, that is close to pulling off the impossible.

  But as the Tahoe example demonstrates, all this comes at a price. Users, not marketers, dictate the dialogue online, and never more so than when they’re invited to contribute their own thoughts. The search for engagement and interactivity online invariably leads to loss of control. For custodians of brands, who have spent most of their careers polishing their messages, the idea that the message is no longer theirs to create and own is somewhere between heretical and terrifying.

  There’s no stopping it, however. More and more, consumers are doing this stuff on their own. YouTube is full of unsolicited Chevy “ads” that are far more sophisticated than anything the Tahoe Apprentice campaign yielded, pro or con. There’s “Chevy Lowrider Commercial,” which shows vintage Chevys hopping down the street, and “Chevy Ridin’,” a slide show of custom Chevys set to the gangsta rap hit “Chevy Ridin’ High.” There’s even “Chevy Tahoe Memorial,” an elegant video that shows a young man jumping, drifting, and ultimately wrecking his much-loved SUV.