Marketing in an Unpredictable World

It's time for producers of entertainment - movie studios, broadcast and cable TV networks, video game makers, publishers, music labels - to change the way they launch and market their products. In entertainment markets, a sizable portion of revenue is typically generated by a small number of blockbuster movies, best-selling books, and hit songs. But even talented, experienced executives acknowledge that predicting these hits is effectively a crapshoot. How else to explain why Miramax paid ten times as much for Happy, Texas - which grossed $2 million at the box office - as Warner Independent paid for March of the Penguins, which grossed close to $80 million?

What should entertainment companies do to improve their odds of success? The key is to understand that the outsize performance of hits is not driven solely, or perhaps even primarily, by intrinsic attributes such as sound, plot, style, or even star power. Rather, new research shows, much of the success of entertainment products derives from social influence - the effect that consumers have on one another's decisions. So in addition to anticipating which features individual consumers might find desirable, executives should adopt strategies that take social influence into account.

A study conducted at Columbia University by Matthew Salganik, Peter Sheridan Dodds, and Duncan J. Watts, and published in the February 10, 2006, issue of Science, sheds light on the role that social influence plays in driving aggregate consumer demand. More than 14,000 participants were recruited through the teen networking site Bolt, and the impact of social influence on their choice of songs to download was tested. After seeing a selection of 48 digital songs by unknown bands displayed on a Web page, participants were asked to choose songs to listen to and then allowed to download the ones they liked. As they arrived at the site, they were randomly allocated to one of two experimental conditions: "independent," in which they saw only the names of the bands and songs; or "social influence," in which they were further divided into eight distinct "worlds," and could see, in addition to the bands and songs, how many times each song had been downloaded by previous participants in their respective worlds.

There were three main findings. First, social influence increased the inequality of outcomes in all eight worlds, meaning that popular songs were more popular and unpopular songs were less popular than when participants made decisions independently. Second, however, which particular songs would turn out to be successful in any given world was more difficult to predict. And third, both inequality and unpredictability increased as the strength of social influence was experimentally increased. Overall, the "best" songs rarely did very poorly, and the "worst" songs rarely did very well, but any other outcome was possible.

These results suggest that the success of a particular entertainment product cannot be explained by any measure of intrinsic quality or even by "appeal" - the fit between the product's attributes and consumers' preferences. Rather, when people are influenced by what others think or do or buy, their individual choices interact in complicated and inherently unpredictable ways. In other words, experts fail to predict hits not because they are uninformed or incompetent but because hits are driven by complex networks of social influences that render accurate prediction of specific outcomes impossible.

The implication for marketing executives is that they should de-emphasize designing, making, and selling would-be hits and focus instead on creating portfolios of products that can be marketed using real-time measurement of and rapid response to consumer feedback.

To move in this direction, we recommend five strategies:

1. Increase the number of bets, and decrease their size. Acknowledging that hits can't be predicted would lead movie studios, for example, to plan for several relatively modest films costing, say, $30 million each rather than a few big-budget ones costing $80 million or more apiece.

2. Focus on detection, measurement, and feedback. E-mail and chat rooms, search engines, blogs, and online communities can accurately measure individual and group reactions to new products in real time. By tracking demand and satisfaction indicators as they emerge, and combining them with separately available sales data, marketers can tailor their campaigns to a rapidly evolving and unpredictable market.

3. Follow through with flexible marketing budgets. Marketing resources should quickly be reallocated from unsuccessful to successful bets as consumer demand materializes. Initial outlays should continue to be guided by prelaunch market research, but marketers should aim at a broader target population than that suggested by their data and intuition. More important, they should direct postlaunch resources at consumers who are reacting positively to the product, whether or not they correspond to marketers' initial expectations. Instead of unlocking the door to consumer demand, marketers should focus on finding and then pushing on doors that are already ajar.

4. Exploit naturally emerging social influence. Once a product has gained a following, marketers can amplify the corresponding social influence signal by directing the attention of a much wider audience toward the individuals or groups who are already enthusiastic about it. This strategy differs subtly but importantly from word-of-mouth or viral marketing strategies that seek to identify so-called influentials in order to solicit their endorsements. Instead, we suggest that marketers can, in effect, create influentials by selectively modifying social influence patterns as they emerge.

5. Build flexibility into supply chains and contracts. Supply chains should be designed to respond rapidly to a growth in demand for some products, artists, or services and a drop in demand for others. Firms can also expend less on the majority of flops, but still capture a share of occasional hits, by building flexibility into contracts with creative artists. For example, more generous royalties and offers of support that are pegged to an artist's success could be exchanged for less up-front investment in production, promotion, and distribution along with an option on any derivative revenues of the kind that superstars typically generate - from endorsements, concerts, and follow-up products.

Rapid changes in the technology of media production, distribution, and consumption are driving a proliferation of choices for consumers - the so-called long tail. Some believe that this trend will reduce the importance of hit songs, blockbusters, and best sellers, as sophisticated search algorithms enable audiences to find and consume increasingly niche-oriented forms of entertainment.

We believe, however, that precisely this proliferation of choice will further challenge consumers' limited capacity to discover and digest content, thus strengthening their tendency to like - or at least preferentially consider - what they think other people like. Meanwhile, social networking sites such as MySpace.com and Facebook, tagging sites such as Flickr and Del.icio.us, and user-generated content sites such as YouTube are increasingly exposing ordinary individuals to one another's decisions about what they watch, listen to, and buy.

Together, these trends point to a world in which successes will be more dramatic - and also harder to predict - than ever. Marketers should therefore abandon the notion that they can either anticipate or determine specific outcomes and instead develop their ability to measure and exploit consumer demand as it arises.

Source: Harvard Business Review, Sept. 2006