Blockbusters and the long tail (part 4)

If consumers derive an enormous amount of value from being able to find obscure products that match their tastes, as we have found that they do, that opens up many business opportunities for firms that can create these matches. But in order to capture that value, firms must first identify the specific business processes that are creating the economic value. What are the characteristics of information-technology-enabled markets that allow consumers to discover and enjoy products that weren’t available in the scarce shelf space of brick-and-mortar retailers?

To find out, we teamed up with Alejandro Zentner and Cuneyd Kaya who had obtained data from a major video-rental chain’s physical and online stores. These data showed that rentals of the 100 most popular DVDs made up 85 percent of in-store transactions but only 35 percent of online transactions. But why? Is the online customers’ shift toward obscure titles caused by the increased variety and ease of search offered online, or is it merely correlated with the types of consumers who choose to shop online rather than in physical stores?

To answer that question, we needed to find an event that would cause consumers to shift from physical to online channels in a way that wasn’t correlated with consumer preferences for obscure products. We found just such an event when our retailer began to close many of its local stores.

Because the decision about which stores to close wasn’t driven by the preferences local consumers had for product variety, we were able to isolate how an individual’s consumption patterns changed when a local video-rental store closed and the person was forced to shift from the limited in-store selection to the expansive selection of the online channel. The data showed that giving consumers access to an expansive selection of products made them much less likely to rent blockbuster titles and much more likely to rent obscure titles that wouldn’t have been available on the physical store’s shelves.

We recognized, however, that this shift might be attributable either to supply (because consumers can access products that weren’t available in physical stores) or to demand (because online search and discovery tools make it easier for consumers to discover new products). Separating these two effects requires fixing either the supply or the demand and varying the other – something we weren’t able to do with our data. Fortunately, Brynjolfsson, Hu, and Simester managed to do just that in studying a different dataset. They analyzed differences in the behaviors of online and catalog consumers of a women’s clothing retailer that maintained the same product assortment in its online and catalog stores (thus fixing the supply side). They found that a significant part of the increased consumption of niche products comes from the demand side – that the technological characteristics of online markets can drive consumers toward niche products even when the supply side doesn’t change.

Subsequent studies have examined in more detail the specific technological characteristics of online markets that might increase consumption of niche products. Consider the role of peer reviews in allowing consumers to evaluate obscure products. Some have argued that peer recommendations will result in more concentrated sales, because early tastemakers influence the market toward winner-take-all products.

However, as we discussed previously, peer recommendations could also allow consumers to discover new perspectives and ultimately buy more niche products, as Gal Oestreicher-Singer and Arun Sundararajan discovered when they collected and studied data from Amazon’s product recommendation networks. Their data allowed them to analyze the relative popularity of products in more than 200 categories of books on Amazon. They found that the categories that were more heavily influenced by peer recommendations exhibited much more diverse consumption patterns than other categories. Specifically, doubling the level of peer influence increased the relative revenue of the least popular 20 percent of products by about 15 percent.

Another factor that might shift consumption away from winner-take-all outcomes is the amount of product information available to consumers in online markets. When consumers have little independent information about products, they often follow the crowd and choose what other people are consuming. This behavior, which social scientists refer to as “herding,” is well documented in the academic literature.