If the Hollywood studios want to thrive in the era of Google, Amazon, and Netflix, they are going to have to think differently about communicating with their customers. In order to do that, they are going to have to make gathering and analyzing data on customers a priority. Consider the instructive story of how Steve Jobs revived Apple and transformed it into one of the most successful businesses in the world. The story in general is well known, but we’d like to focus on one aspect of it that gets less attention than it should: how Apple used connections with customers and data on customers to turn itself around.
In 1997, when Jobs returned to Apple, the company was struggling. It had barely four percent of the computer market, its stock price had just hit a twelve-year low, and many industry experts were predicting it would soon have to fold. On October 6, Michael Dell said at a Gartner Symposium that if he were in charge of Apple he would “shut it down and give the money back to the shareholders.”
One of Apple’s big problems was that, with a small share of the market, it couldn’t reach its customers directly. It relied on third-party retailers, including Sears, Best Buy, Circuit City, and OfficeMax, to sell its computers to customers, and those retailers had no incentive to create customer loyalty to Apple’s brand. Their salespeople knew little about Apple products and, in fact, would often steer customers away from Apple products and toward cheaper Windows computers. Apple’s products were relegated to poorly trafficked and poorly maintained parts of the stores. Many people simply weren’t aware of the benefits that Apple computers had to offer. Jobs recognized this as a major obstacle. His plan for turning Apple around involved delighting customers, but it isn’t easy to delight customers if you have no way of reaching them or even knowing who they might be.
Jobs came up with a crazy solution: Apple would build its own stores. Why was that crazy? Well, for one thing, retail-store space was expensive. It seemed almost ludicrous to propose that the way to compete with Dell in a slim-margin business was to make a huge investment in retail stores. That approach certainly hadn’t worked for Gateway. In January of 2001, four months before Apple would open its first retail store, Gateway, suffering in competition with Dell, had been forced to close 27 of its retail stores.
Not surprisingly, the business press openly mocked Apple’s plan. In an article titled “Sorry Steve, Here’s Why Apple Stores Won’t Work,” Business Week suggested, “Maybe it’s time Steve Jobs stopped thinking quite so differently.” “Apple’s problem,” the company’s former chief financial officer Joseph Graziano said, “is it still believes the way to grow is serving caviar in a world that seems pretty content with cheese and crackers.” David Goldstein, a retail consultant for Channel Marketing Corporation, summed up the conventional wisdom: “I give them two years before they’re turning out the lights on a very painful and expensive mistake.”
Today, Apple has 453 retail stores in 16 countries. In its earnings announcement for the first quarter of 2015, Apple reported that 500 million consumers visited its retail and online stores, and that the retail stores generated nearly $4,800 in annualized sales per square foot—more than those of any other retailer in the United States. 5 According to Forbes, Apple has 50,000 retail employees worldwide, who serve, on average, a million customers per day. Apple’s retail operation is now worth more than all of Apple was worth in 2001.
Apple’s retail stores succeeded largely because, instead of just pushing products as most computer retailers did, Apple focused on the consumer experience. In particular, it designed its stores not around product lines but around the needs of its consumers, showing them how they could use Apple products to listen to music, take pictures and videos, and watch movies. Significantly, Apple also staffed its stores with friendly people who could guide customers through their purchases and teach them how to use what they were buying. That part of the story is well known. What gets less attention is that Apple used its retail stores to take control of how its products were displayed and marketed to its customers—and it did that by using data.
Experiments and data influenced every aspect of Apple’s stores. The company spent lavishly on mock-ups of various store layouts and used feedback from customers to refine their design. It interviewed customers about their best customer-service experiences and used what it learned to design the Genius Bar. It studied market data and demographic data in order to locate its stores conveniently so that new users to the Mac platform (most of them Microsoft users) had to gamble not with “20 minutes of their time,” as Steve Jobs put it, but only “20 footsteps of their time.”
Apple used data to create its stores and the experiences people had in them—but then it used the stores to create a data feedback loop from the customers back into the company. The stores—from the Genius Bar and the one-to-one training areas to the sophisticated technology used to identify the physical location of in-store customers—were carefully designed to collect data about customers that Apple could then feed back into how it designed and marketed products and served customers. How were customers using Apple devices? What did they like best about them? What about the devices did customers find frustrating? What devices were breaking, in what way, and after how long? What did people want? What did they need? The retail stores put Apple directly in touch with customers, and the company used the stores expertly to collect information that helped them understand and serve their customers’ needs.
We aren’t suggesting that Paramount Pictures or Universal Music open stores in swanky shopping districts all over the world. However, in a world in which direct connections with customers are becoming increasingly important, we do believe that the major entertainment companies will put themselves at a strategic disadvantage if they rely exclusively on third-party distributors to showcase their products to customers. To compete, they will have to make the kind of transformation that Apple did—a shift that involves communicating value through direct connections with customers rather than through intermediaries, and that involves collecting and using data on customers to understand and serve customers’ individual needs. How might this work in practice?
Let’s consider the motion-picture industry. Some of the data the studios will need to connect with their customers is readily available on various social media sites and can be accessed relatively easily. Legendary Pictures, for example, has invested in an analytics division that aggressively collects consumer data from every source they can find it in—Twitter, Facebook, Google, ticketing data—and then uses it to better reach the right customers with the right promotional message. Without this kind of targeting, Legendary’s CEO Thomas Tull has said, studios are wasting money by spending as much money on marketing The Dark Knight to eighty-year-old women as they spend on marketing it to teenaged boys.
However, as we have discussed previously, the most valuable data for the studios are proprietary and closely held by Apple, Amazon, Google, and Netflix. A simple approach for the studios to adopt would be to prioritize getting access to data on individual customers when they negotiate with their distributors. In fact, many studios have already begun to press their distributors for more detailed data. But the studios will face obstacles as they try to adopt this approach. They will have to make significant concessions to powerful online distributors if they want to be able to gain access to their customers. Even if the studios manage to negotiate that access, they will still only be able to see customers’ behavior for their own content, whereas online distributors will be able to observe customers’ behavior across all content on their platforms. The studios also will not be able to benefit from the sort of direct testing and experimentation that we discussed in the previous chapter. And, maybe most important, if the studios rely on these distributors to maintain access to customers, they will increasingly be relying on their competitors for access to strategic information.
It’s important to realize that Amazon, Netflix, and Google are vertically integrating upstream into the original content business in large part so they will be less reliant on the major studios for access to the studios’ content. For that reason alone, the studios should turn the tables on their new rivals: vertically integrating downstream into direct distribution so they will be less dependent on the distributors for access to customers. The most straightforward way to do that, at least for content with a strong brand, would be to invest in attracting consumers to the studios’ existing online portals. Here the studios could adopt J. K. Rowling’s strategy with Pottermore and create communities in which creators can share additional information with fans. This would enable the studios to keep track of consumers’ behavior across the various communities in their roster, and to use that information to promote content directly to fans. The main obstacle facing that approach is that, as we have discussed, consumers have a strong desire for simplicity. They may be hesitant to learn how to use multiple websites and unwilling to maintain multiple logins across multiple content sites. Likewise, if each studio maintains its own separate platform, it will be able to observe only how consumers interact with the studio’s own content.
A more ambitious approach, and one that we believe holds the most promise for success, involves the studios’ forming a strategic partnership to invest in a common platform that would allow each firm direct, targeted access to its customers while enabling each firm to see and understand customers’ behavior across multiple firms’ content. And such a platform already exists. In March of 2007, three major studios—21st Century Fox, NBC Universal, and Walt Disney Studios/ABC Television—announced that they had teamed up to create “the largest Internet video distribution network ever assembled.” 10 Today that network, called Hulu, is the fourth-most-popular video-streaming platform in the United States, just behind Amazon’s Instant Video service.
Unfortunately, Hulu’s success has created an almost intractable problem. The more popular Hulu became, the more likely it was to reduce the profitability of the studios’ established channels. Disputes arose almost immediately after its launch about which shows would be made available to Hulu, how many minutes of commercials Hulu should include in their streams, how many episodes of content would be available on the site, and how long Hulu would have to wait after the TV broadcast before making content available. These questions were all decided in the context of the current television business model, in which about half of a network’s revenue comes from advertising dollars and the other half from “retransmission fees”—payments from cable companies for the rights to broadcast the networks’ content. As a result, anything that might reduce a show’s Nielsen ratings or retransmission fees was seen as a threat.
One way to deal with these conflicts would be to allow Hulu to operate completely independently, giving it the freedom to pursue new business models that might ultimately cannibalize the old ones. That was the idea behind a 2010 proposal for an initial public offering for Hulu, but the networks quickly killed it. “I don’t know if any amount of money would have been enough to get them to give up control,” the Wells Fargo analyst Marci Ryvicker observed in a Fortune interview.
In February of 2011, in an effort to convince the networks to stop impeding Hulu’s success, Hulu’s CEO, Jason Kilar, posted a 2,000-word entry on Hulu’s blog in which he lectured his bosses about the future of their business. Kilar argued that there were too many advertisements on television, that consumers should be allowed to watch content on their own schedule, and that the business of forcing customers to buy huge cable packages of channels unrelated to their interests was dying, as was the business of charging cable companies huge retransmission fees to fill those channels with content. Kilar closed with a warning to network executives: “History has shown that incumbents tend to fight trends that challenge established ways and, in the process, lose focus on what matters most: customers.”
The networks were not immediately grateful for Kilar’s feedback. “Eighty to 90 percent of what he says is right,” one unnamed network executive told the Wall Street Journal. “But why print that? Does he think we’re going to say, ‘Oh, thank you! You’re right! We’d never thought of that! Let’s give away retrans [retransmission fees]!’?” Another executive, quoted in the Financial Times minimized Kilar’s accomplishments and his ability to work within the existing system: “If I were given billions of dollars worth of programming, I too could probably build a business. But I know that in order to build a long-term, viable business I would have to do so in a way that works for everybody.” A third executive was even more blunt: “These are clearly the musings of an elitist who is obviously disconnected from how the majority of America watches television.”
In the end, many industry observers feel that a vibrant new streaming platform isn’t compatible with the studios’ existing business models. “They don’t want it to succeed,” the media analyst James McQuivey wrote. “It’s media economics—if it succeeds, it will do so by cannibalizing the currency of the television media business today, which is television ratings. So they made a decision to make sure Hulu didn’t get too good or become too successful.”
The gut instinct to disadvantage digital channels to avoid cannibalizing existing revenue streams may seem perfectly reasonable under the old rules of the entertainment business. After all, people who can’t consume digitally will have no other choice but to buy the physical product, right?
That’s a question we examined recently using digital and DVD sales data collected from 2012 and 2013. Before 2012, the conventional wisdom in most of the motion picture industry was that delaying the release of movies on iTunes and other digital channels would protect the studios’ valuable DVD revenue. However, from 2012 to 2013 several studios started to experiment with releasing their titles digitally at the same time as, or in some cases before, the DVD release. This shift in strategy at the studio level allowed us to analyze how sales in both channels changed when consumers could choose between DVDs and iTunes downloads. The data showed that delaying digital availability has a huge downside and almost no upside. When digital movies were released after the DVD release date, digital sales were cut by almost half and there was no statistical increase in DVD sales.
Delaying digital availability is also risky for firms because without strong digital platforms the majors will not be able to tap into the many important benefits of digital distribution. We have identified five benefits of digital distribution that we feel are particularly significant. We have already discussed two of them extensively: the ability to better evaluate the potential market for content and the ability to more efficiently promote this content to consumers. Netflix has exploited both of these advantages, along with a third: the ability to conduct and learn from detailed experiments about how consumers respond to content. In her book Netflixed: The Epic Battle for America's Eyeballs, Gina Keating describes how the company has used its website to understand its customers’ needs:
[Netflix] designed the Web site to double as a market research platform that could display multiple versions of a page or feature to test groups of customers and gather detailed data on their reactions and preferences. A typical A-B test involved measuring the effect of a red logo (choice A) versus a blue logo (choice B) on acquiring a customer, and their lifetime value, retention rate, and usage. … The constant testing, gathering of consumer input, and subsequent adjustments to the site formed an ongoing conversation between Netflix and its customers that would provide a crucial advantage in the coming battle with store-based renters.
Of course, competitors can simply copy those design decisions, and at the time Blockbuster.com did just that, quickly incorporating elements of Netflix’s design. But although Blockbuster.com copied the look of the Netflix site, it obviously couldn’t copy the algorithms underpinning it. Without ongoing optimization of costs, of the matching algorithm, and of the market-research platforms, Blockbuster didn’t have anything close to the whole picture.
Platform companies can do a very good job of customizing their marketing by using direct experiments, as we noted in the previous chapter. But they can also generate insights that wouldn’t be possible without data on individual customers—our fourth benefit. For a long time, demographic data of the sort sold by Nielsen and other market-research firms were the only data on customers that were available, because demographic data were all that marketers could measure efficiently. But demographic data tell you almost nothing about who people are or what they want to consume. In a world of digital interactions with customers, and of tremendous computing power, demographic data have become almost worthless for the purpose of making marketing decisions.
How valuable is it to know a customer’s purchase history rather than only his or her demographic characteristics? Peter Rossi, Robert McCulloch, and Greg Allenby asked that question in a paper published in 1996. 20 At the time of their study—not long after grocery stores had adopted bar-code scanners and customer loyalty cards—marketers were just beginning to appreciate the importance of data on customers. Loyalty cards enabled grocery stores to see their customers as individuals for the first time. Rossi, McCulloch, and Allenby used data from a store’s scanner data to compare the effectiveness of untargeted “blanket” coupon drops, coupons targeted on the basis of demographic data, and coupons targeted on the basis of a customer’s purchase behavior. They found that knowing a person’s demographic data increased the profitability of coupons by 12 percent relative to “blanket” couponing, and that adding an individual’s purchase history increased the profitability of coupons by 155 percent relative to the blanket strategy.
A tenfold increase in marketing effectiveness is incredibly powerful at the scale at which Amazon works. But Amazon and other platform companies didn’t stop there. Increasingly, they are basing their marketing decisions on their ability to observe a customer’s behavior in real time. Amazon wants to customize its marketing on the basis of what you’re currently searching for, what you’re currently looking at, how often you click, and so on, because data on such behaviors help to answer the most salient marketing question of all: Why are you here right now?
The fifth benefit from observing detailed data on customers has to do with products rather than with customers. To clarify what we mean, let’s go back to the grocery industry in the mid 1990s. Before grocery stores adopted customer loyalty cards, the conventional wisdom in the industry, based on a study conducted by the Food Marketing Institute, was that to improve their inventory management grocery stores should reduce the number of niche products they carried. The H-E-B grocery chain, however, discovered that the Food Marketing Institute’s study had ignored something fundamental to a store’s profitability: that the most profitable customers were the ones most likely to purchase “niche” products. H-E-B recognized that if it eliminated these slow-selling niche products it might lose its most profitable customers, so its managers decided to stock more niche products.
Might something similar be true in online purchasing—that the most profitable customers might be those who are interested in the least popular products? Seeking an answer to that question, we partnered with a major movie studio to analyze its customers’ online purchases. The answer turned out to be Yes. On the whole, we found, sales skewed heavily toward blockbuster movies. That was no surprise. But it was a surprise that sales from the most profitable customers skewed heavily toward purchases of obscure movies. The most profitable customers were between 50 and 200 percent more likely than other customers to buy movies from the long tail.
The overarching lesson for the entertainment industries is that, to succeed in the future, companies are going to have to control the interface with their customers (and the resulting data about their customers’ needs) in addition to controlling the production of content. That’s what we have been arguing throughout this book.
Of course this transition will be difficult for the entertainment industries. But we’re optimistic about their future. That’s because the steps we have called for above are the same steps that have always defined success in the entertainment industries: a willingness to take big risks on emerging opportunities, a desire to invest in new talent, a passion for finding creative ways of connecting artists with audiences, and the skill necessary to take a grand concept and make it a reality. One way or another, the show must—and will—go on.
This article was excerpted from the new book Streaming, Sharing, Stealing: Big Data and the Future of Entertainment from The MIT Press by Michael D. Smith and Rahul Telang.