Fifteen years ago, while reporting for Fortune on Seagram's troubled acquisition of Universal, I was told by a studio executive about number crunchers flying in from New York, calculating that mega-budget movies generally produce the highest returns, and wanting to know why the studio didn't make just six giant pictures a year. "Literally," he declared in disbelief, "that was a conversation."

Today the disbelief is long gone. Variants of this conversation have been held all over Hollywood – and the result is what you see at the multiplex: a demolition derby of sequels, franchises, tent poles and mega-flops, with the occasional mid-range picture tossed in for Oscar bait.

Aesthetically, it's widely agreed, the result is a disaster – but commercially? Depending on which end of the spectrum you're looking at, 2013 was either the best of years or the worst of years for Hollywood.

The best in the sense that a dozen movies, almost all of them costing well over $100 million to produce and another $100 million or more to market, brought in a total of nearly $10 billion at the box office worldwide – roughly half of which went back to the studios. The worst in that another dozen pictures, made by the same studios for roughly the same amount of money, yielded ticket sales of only $3 billion, a dismal tally that in some cases led to spectacular write-downs.

George Lucas and Steven Spielberg had predicted as much in a public appearance at the University of Southern California last June. Surveying the current landscape, the two filmmakers – the very people who less than 40 years ago invented the blockbuster as we know it with Star Wars and Jaws – saw little to be sanguine about. "There's going to be an implosion," Spielberg said, "where three or four or maybe even a half-dozen mega-budget movies are going to go crashing into the ground, and that's going to change the paradigm."

In the following weeks, the implosion played out as forecast: White House Down on June 28, The Lone Ranger on July 3, Pacific Rim on July 12, Turbo on July 17, R.I.P.D. on July 19, Elysium on Aug, 9. All told, close to $1 billion in production costs alone went up in smoke, with more disasters to come at Christmas. But the paradigm didn't change at all. This year, the studios are releasing so many would-be blockbusters that they had to start in February, in the midst of Oscar distractions, with Sony's $100 million remake of RoboCop.

What explains such apparent madness? One thing: as big as the losses from Hollywood's blockbuster strategy can be, the wins have proved even bigger. Gone are the days, apparently, when a single outsize disaster might sink a studio into the Pacific, as Heaven's Gate did to United Artists in 1980. Universal Studios lost millions last year on R.I.P.D. and the holiday stinker 47 Ronin, but that barely mattered next to the nearly $1.8 billion taken in by Despicable Me 2 and Fast & Furious 6. The $190 million write-down that Disney took on The Lone Ranger was made up over just a couple of weekends by the gushers of cash from Iron Man 3 and the animated feature Frozen. And so on, all across Tinseltown.

It's conceivable that, so far, the studios have simply been lucky. But in her recently published book, Blockbusters, Anita Elberse of the Harvard Business School offers intellectual justification for Hollywood's bet-the-farm strategy. Though the safer course might appear to involve making a larger number of pictures at smaller budgets, she writes, "content producers can't afford to walk away from big bets – doing so would actually increase their chance of failure in the long run."

The "laws of consumer behavior" that explain this, Elberse argues, apply not just to Hollywood, but to entertainment of every stripe. What's more, she asserts, the conventional wisdom that the digital era will prove inhospitable to the mega-hit is wrong. From Justin Bieber to Lady Gaga, "blockbusters will become more – not less – relevant to popular culture, and blockbuster strategies will thrive."

This is a message that embattled mass-media entertainment executives have been longing to hear. But does the evidence back up the claim? Obviously, neither hits nor stars have gone away. Yet that is hardly the same as saying that mass-appeal blockbusters will be the signature mode of entertainment in the digital age. Indeed, on closer examination, the pattern now emerging is actually quite different. Not fewer hits but more, and many of them not undifferentiated world beaters but smaller data-driven hits tailored to the tastes of specific audiences. Not M*A*S*H, whose 1983 TV finale was watched in three out of every five American households, but Netflix's House of Cards. Not mega-hits, but smart hits.

A Formula for the Digital Age?

The claim that blockbusters will prevail in the digital era is certainly a provocative one. Ever since 2006, when Chris Anderson published The Long Tail: Why the Future of Business Is Selling Less of More, many have considered it axiomatic that once the world is liberated from the physical constraints of shelf space, infinite choice will flourish. Anderson – who wrote the book when he was editor-in-chief of Wired, where I was a regular contributor – did not in fact predict the end of hits. His model of demand, like that employed by Elberse and others who study this phenomenon, was a "power law" curve, with a small number of hits clustered at one end of a graph and a large number of far less popular items – the "long tail" – trailing off at the other.

In the physical world, the need to show movies in theaters or display products in stores puts an obvious limit on how much of this long tail will be available to any given consumer. But in a digital environment, Anderson pointed out, such limits don't exist. This, he argued, would spell the end of "the water cooler era," with its emphasis on a common culture of hits, and the advent of "the micro-culture era, when we're all into different things." A result: "The companies that will prosper will be those that switch out of lowest-common-denominator mode and figure out how to address niches."

Eight years on, we can see that it hasn't quite worked out that way. Even two years after the book came out, Google's chairman, Eric Schmidt, an early supporter of Anderson's long-tail theory, admitted to second thoughts. In an interview published in McKinsey Quarterly and quoted by Elberse, he explained, "I would like to tell you that the Internet has created such a level playing field that the long tail is absolutely the place to be – that there's so much differentiation, there's so much diversity, so many new voices. Unfortunately, that's not the case. … So, while the tail is very interesting, the vast majority of revenue remains in the head," the part of the curve where the hits reside. "And this is a lesson that businesses have to learn. While you can have a long tail strategy, you better have a head, because that's where all the revenue is."

In Schmidt's revised view, the Internet will accelerate the transition to the kind of world described by the economists Robert Frank of Cornell and Philip Cook of Duke in their 1995 book, The Winner-Take-All Society. This is a world in which "the value of what gets produced," as they put it, "often depends on the efforts of only a small number of top performers, who get paid accordingly." It is a world dominated by superstars – some of them household names, others all but unheard-of outside their chosen fields, be it law or banking or academia.

Globalization encourages the development of winner-take-all markets because it broadens the talent pool enormously. It puts opera singers in New York in competition with divas in Moscow, software engineers from Silicon Valley in competition with their counterparts in Beijing, automobile designers in Detroit in competition with those in Stuttgart and Turin. With the triumph of the Internet, Schmidt predicted "larger blockbusters and more concentration of brands" as well as the rise of global superstars – "global brands, global businesses, global sports figures, global celebrities, global scandals, global politicians."

Around the time that the Schmidt interview came out, Elberse addressed the issue in a Harvard Business Review article titled "Should You Invest in the Long Tail?" Producers will find little profit there, she concluded after a bout of number crunching, so they should limit their exposure and focus on potential hits. Retailers might want to adopt a long-tail strategy in order to keep their biggest customers happy, since people who buy a lot of music or books or whatever tends to be the heaviest consumers of niche offerings in these mediums. But there won't be much profit in this, so costs will need to be kept to a minimum. And most people will never go beyond the hits.

All of which seemed quite reasonable. But as she developed and extended this argument for the book, Elberse ventured into dubious territory. Figures that appear to reveal an increasing concentration of sales at the hit end of the spectrum in fact mask a quite different pattern – one of an expanding number of hits, most of them on a considerably smaller scale than hits in the past. At the same time, Elberse conflates blockbusters – aberrant mega-hits of the type currently beloved by Hollywood – with ordinary hits and the stars who typically power them. She even writes about "blockbuster" operas (La Bohème or La Traviata as opposed to, say, Reiner Bredemeyer's rarely performed Der Neinsager) and the Metropolitan Opera's Live in HD program, which simulcasts popular productions in movie theaters from Albania to Uruguay.

Opera is certainly a winner-take-all enterprise, and the Met is unquestionably a winner. But if simulcasts of La Traviata and its ilk can be considered blockbusters when they run once in the same multiplex that plays Despicable Me 2 ad infinitum, you have to wonder: does this term still mean anything?

Jaws and the Hollywood Blockbuster

Hits and stars are nothing new. We've had them at least since the era of Charles Dickens and Jenny Lind – which is to say, since the advent of mass media in the mid-19th century. Blockbusters are a much more recent phenomenon. The movie industry has always delivered a smattering of outsized hits, Gone with the Wind being one of the earliest examples. But until 1975, when Universal released Jaws, high-quality pictures generally opened in a few dozen theaters and built slowly over time.

After Jaws, this all changed. The Hollywood studios and their corporate parents started going for a fast payback, which meant wide releases backed by pricey television advertising. By the end of the 1990s, studio executives like Alan Horn of Warner Bros. had evolved this into a "blockbuster strategy," allocating a disproportionate share of resources to a handful of pictures on the theory that, on average, they would generate wildly disproportionate sales. Book publishers and record labels have employed similar game plans.

Yet much of the success that Hollywood has enjoyed with its blockbuster strategy to date is a function of circumstances unique to the movie business – in particular, the ever-growing sophistication of digital effects (including 3-D) and the extraordinary rise of the international box office. Foreign sales, once an afterthought, accounted for nearly 70 percent of the industry's box-office receipts by 2012. Razzle-dazzle effects and a spate of new theater construction around the globe, combined with a spectacular rise in consumers' discretionary income, have turned what had largely been a domestic business into one dominated by non-English-speaking territories – markets where, much as with American teens, subtleties of characterization and dialogue are hard- pressed to compete with action sequences.

Even so, it's difficult to make a convincing case that Hollywood's blockbuster strategy is a uniquely digital-age formula for success. Compared with music and the print media, the movie business is still at a relatively early stage of disruption. In their early phases, digital technologies often favor established industries that are smart enough to adopt them because digital can lower costs or facilitate price increases, or both. This was the case with CDs and the music industry in the 1980s and 1990s, with DVDs and the movie and television industries in the 1990s and early 2000s, and with digital 3-D and digital distribution to theaters for the movie business in the 2010s.

But once digital starts to bypass physical objects (like CDs and DVDs) and enable direct online distribution via downloads and streaming, these advantages can quickly vanish. When CD sales collapsed, for example, sales of recorded music fell by half – and not just because of piracy.

This is currently happening with home video. A decade ago, home video provided half of Hollywood's profits. But DVD sales peaked in the United States in 2006, and in 2007 the number of households with broadband Internet access passed 50 percent. As people realized that it's cheaper and easier to stream than to own, they began to forsake $15 DVD purchases from Blockbuster Video (now defunct) in favor of $3 rentals from Amazon and iTunes, or all-you-can-watch binges from Netflix.

In response, studio executives have attempted to mitigate their exposure by making huge bets on properties they perceive as low risk. Since they lack any real data about audience tastes, this generally means retreads of whatever worked in the past. The magic word is "pre-awareness" – a quality possessed by Batman and Spider-Man and anything else the audience already knows about, as opposed to an original idea. Sometimes this works: Fast & Furious 6, Iron Man 3, Despicable Me 2. With The Hangover Part III, not so much.

In other respects, however, Hollywood is still on the golden side of digital. The studios are saving on distribution costs because it's far cheaper to send files to movie theaters electronically than it is to reproduce and deliver reels of film. At the same time, theaters have been able to charge more for tickets – in the case of 3-D, significantly more. And while the elaborate digital effects so typical of blockbusters cost plenty of money, they also serve as a powerful draw to audiences and as an even more powerful barrier to entry. As Alan Horn, now chairman of Walt Disney Studios, remarked to Elberse: "Very few entities in this world can afford to spend $200 million on a movie. That is our competitive advantage."

Nonetheless, a deep-seated malaise has gripped the industry. The collapse in DVD sales has everyone spooked. And the relentless focus on blockbusters has demoralized even moviemakers as big as George Lucas and Steven Spielberg. As Lynda Obst, producer of the 1993 hit Sleepless in Seattle and now author of the book Sleepless in Hollywood, said in a recent interview: "All formulas will only work for a while. How many times can you see the same cities destroyed? How many ways are there to destroy them?"

Last year, even the geeks at Comic-Con, the annual comics convention that the studios have turned into a Cape Canaveral for blockbusters, were said to be disappointed by the sameness of the fare at hand. Far from a savvy game plan for the digital future, Hollywood's embrace of the blockbuster appears to be the last-ditch strategy of an industry on the verge.

Justin Timberlake, meet Michael Jackson

At least people are still buying movie tickets. Not quite as many as before – in the United States and Canada, anyway, where annual per-capita sales fell from 4.9 in 2003 to 4.1 in 2012. But ever-rising ticket prices and the still-growing international market are making up for that. The same cannot be said for books, recorded music and most other media. Across the board, "blockbusters" are almost never what they used to be.

Take Justin Timberlake's The 20/20 Experience, which according to Nielsen SoundScan was America's best-selling album of 2013 – even though it sold only about 2.4 million copies, making it the smallest best seller since SoundScan began in 1991. Timberlake's was the only album to crack the two million mark last year. Adele had a truly outsized success in 2011 and 2012 with 21, which sold more than 10 million copies to become the top-selling album for both years running. During that same period, however, only three other albums topped two million. In fact, of the 20 best-selling albums, according to SoundScan, Adele's was the only one to come out after 2004 – and it didn't begin to compare in sales with albums released in earlier decades by Michael Jackson, the Eagles, Led Zeppelin or AC/DC, even though the U.S. population has grown by a quarter since 1991. Globally as well, Adele was far outsold by earlier stars.

Album sales have been hit particularly hard in recent years because people downloading from iTunes and the like tend to cherry-pick individual songs – a prime example of "unbundling," which is one of the signature effects of digital. So perhaps we should look at songs, as Elberse does in Blockbusters. Examining the SoundScan numbers for 2007, 2009 and 2011, she finds a pattern of increasing dominance by a tiny percentage of superhits.

In 2007, of the 3.9 million tracks sold, fewer than 0.001 percent sold more than a million copies, while a remarkable 91 percent sold fewer than 100. Two years later, 6.4 million tracks were sold; this time, a little over 0.001 percent sold more than a million copies – but that tiny percentage made up 12 percent of the total sales volume, even more than the 7 percent in 2007. And of the eight million tracks sold in 2011, again 0.001 percent sold more than a million copies, this time accounting for 15 percent of total sales. Meanwhile, at the other end of the chart that year, 94 percent sold fewer than 100 copies and fully 32 percent sold just a single copy. So much for the long tail. "It is staggering to see how few titles at the top contribute to a significant portion of sales," Elberse declares, "and how many titles at the bottom fail to do the same."

That's one way of looking at it. But in focusing on the lack of promise in the long tail, Elberse misses another pattern entirely. As the total number of single songs on sale grew from 3.9 million to 6.4 million to 8 million, the number that sold more than a million copies went up at an even greater rate, from 36 to 79 to 112. At the same time, the average number of copies that each song sold actually fell slightly. In other words, a growing supply of songs produced a growing number of hits, even as the average size of those hits declined. The tail got a great deal longer and flatter, but the head did not get commensurately taller and thinner. Except for a couple of outliers like Adele, it actually got a bit shorter and wider.

In television, this is even more apparent. The percentage of TV households tuned to the season's top series has been dropping at least since the 1960s. NCIS, the top-rated show of the 2012-13 season, had roughly half the 36 rating that The Beverly Hillbillies enjoyed in 1962-63. It didn't even rise to the level of The Flintstones, a middling hit a half century before. And the shows that have people talking about a third golden age of television, as the medium's current creative renaissance has been dubbed (with cable series like AMC's Mad Men, Showtime's Homeland and HBO's Girls) pull in far fewer viewers than that. Only AMC's The Walking Dead gets ratings big enough to rival even the slimmed-down hits on broadcast television today.

So while the top of the curve is still where the money is, the pattern does not conform to the prediction that the Internet will lead to ever-bigger blockbusters. We do indeed live in a winner-take-all economy, and electronic media certainly encourage the rise of global superstars. But that's as true of broadcast TV as it is of the Internet, as true of the Beatles as of Adele. As Internet usage takes hold, hits – even the superhits – appear to be shrinking in size and growing in number. Which raises an interesting question: if neither the blockbuster hypothesis nor the long-tail theory holds, then what?

Data is the New Gut

Elberse argues in Blockbusters that Netflix, the Silicon Valley outfit that once "took pride in calling itself a long-tail company," has started "acting more like an old-school television network" by producing expensive, star-laden series like House of Cards, the Kevin Spacey-Robin Wright vehicle that won three Emmys in its first season. True, there's nothing small about House of Cards, but there's nothing old-school television about it either. Neither long tail nor blockbuster, House of Cards is, in fact, the media property that most plausibly suggests a blueprint for the digital future.

In broadcasting, as in Hollywood, the key decisions have always been made by people acting more on instinct than information. In the old days, when the entertainment industry was run by showmen (women didn't get to run things back then), that meant they went with their gut – a phrase that gained resonance with the discovery that the intestine is lined with about 100 million neurons. But the MBAs who run movies and television today are expected to base their decisions on something more reliable, more quantifiable or at least more professional-sounding than their digestive tracts.

Unfortunately, that doesn't leave them with much. So they rely on formulas and sequels and "pre-awareness," and they give copious notes in an effort to smooth out rough edges and justify their own salaries, and they demand recuts and reshoots and sometimes entirely new endings on the basis of whatever feedback they get from test screenings. "You've got people who don't know movies and don't watch movies for pleasure deciding what movie you're going to be allowed to make," Steven Soderbergh declared in a State of the Cinema address at last year's San Francisco Film Festival, shortly before the debut of his celebrated Liberace biopic, Behind the Candelabra – on HBO.

Netflix doesn't employ such people. And far from using a blockbuster strategy, it relies on a data strategy. Unlike the movie studios and all but a handful of television networks, it has a direct relationship with its customers – 44 million subscribers worldwide who constantly feed it information about their likes, their dislikes and their viewing habits. Where a blockbuster strategy depends on mass appeal and instant success, a data strategy is targeted, specific, patient and "smart." Both aim to generate hits. But there, the similarities end.

With access to billions of data points, Net­flix can build statistical models to produce a risk profile of any show it might want to license or produce based on such variables as genre, subject matter, stars and Emmy and Oscar wins. When the film director David Fincher was pitching House of Cards, a remake of a BBC political drama in which Spacey would play a scheming congressman, Netflix's chief content officer, Ted Sarandos, knew how many subscribers had watched the original, how many had watched The West Wing, how many liked David Fincher movies and how many were Kevin Spacey fans. This enabled Sarandos – a college dropout who got his start running a mom-and-pop video store in a Phoenix strip mall in the 1980s – to do what no television executive ever had: offer a reported $100 million for two 13-episode seasons without even asking for a pilot. Then, in an appeal to audiences' demand for control and a growing appetite for binge viewing, he released each season's episodes all at once.

Nielsen doesn't measure Netflix streams, but Procera, a company that works with broadband providers to enhance network performance, estimates that when the second season was made available last February, as many as five million subscribers watched at least one episode during the first weekend and more than 500,000 of them gulped down the whole thing. Five million people is less than a quarter of the viewership CBS gets for NCIS, but it's only a couple of million below the prime-time averages for ABC, NBC and Fox.

In any case, Netflix gets its revenue from subscribers, not advertising, and the buzz around House of Cards and other original shows like Orange Is the New Black has sent subscriber growth skyward. In 2013, only a couple of years after a misguided restructuring and price increase sent the company spiraling into a near-death experience, annual profit rose by more than sixfold.

Data has been central to the company's success. But it isn't just what Sarandos has done with all the information at his command; it's also what he hasn't done. Having decided that the combination of elements in House of Cards would work, he gave total creative freedom to Fincher and his partners.

"Netflix was the only network that said, 'We believe in you,' " Spacey declared in the MacTaggart Lecture he delivered at last summer's Edinburgh Television Festival, the annual confab for UKTV. He contrasted this with a 1980 NBC memo that's quoted in Difficult Men, a book on the producers who are driving television's current creative renaissance. The memo summed up the reaction in focus groups to the pilot for Hill Street Blues, the landmark series that was about to kick off the second golden age of television: " 'The most prevalent audience reaction indicated that the program was depressing, violent and confusing.' … 'Too much was crammed into the story.' … 'The main characters were perceived as being not capable and having flawed personalities. Professionally, they were never completely successful in doing their jobs and personally their lives were in a mess.' … 'Audiences found the ending unsatisfying. There are too many loose ends.' "

Fortunately for NBC, the showrunner, Steven Bochco, had been guaranteed autonomy. After a slow start, Hill Street Blues went on to win multiple Emmys and become one of the most popular and influential TV dramas of the 1980s. As Spacey declared in his talk, "It's the creatives, stupid."

The Unpredictable Business of Entertaining Humans

Hits are hits, and we all want more of them. But the blockbusters we have today are not smart, in any sense of the word. The problem is not that it's inherently reckless to spend upward of $200 million to make and market a film, or even to do it over and over. Properly executed, in the current environment, this can be a formula for success, as Elberse demonstrates. The problem is that it's reckless to do it the way Hollywood does now – with minimal understanding of the market, little faith in the people who actually create the stuff, a naïve conviction that whatever worked in the past will continue to work in the future, and the false sense that audiences will grow ever larger.

To the frequent consternation of those who try to run it, the entertainment business is inescapably in the business of entertaining humans, a species that craves novelty as much as it craves spectacle – and one prone to sudden and unpredictable shifts in taste. Any theory that fails to take this into account is unlikely to survive the next cycle.