Organizational Behemoths - How Do Amazon, Facebook, Google and Apple Co-exist with Granularity?
Google, Facebook, Amazon and Apple are behemoths in a winner-take-all market due to supply side economies of scale and demand side economies. Information has a unique cost structure, unlike most other goods. There are supply side scale economies generated by high fixed costs of production and near-zero marginal costs of informing additional people. So it costs zero to sell information to additional users. However, high startup costs will naturally give an advantage to early entrants in the market.
Network effects, both direct and indirect, generate demand-side economies of scale, further increasing the early entrant advantage. Direct network effects arise when the value of a product depends upon the size of the network of users as in a telephone network. An important requirement here is implicit coordination among consumers so that all link to the same network. If consumers are spread across different networks, the benefit of network size is dissipated and the direct network effect is lost. The value of joining a social network, say Facebook, increases when more of my friends are also members of this network.
Indirect network effects arise when complimentary goods are produced in proportion with network size as in the case of applications for mobile devices. When more consumers are on the Android network, it pays for more apps to be written for Android devices. Hence, it becomes vital to get to the tipping point or the critical network size in order to collect these network effects. Similarly, when there are more shoppers on Amazon, it pays to be a seller of goods on that site.
Also on the demand side are information cascades. This is the case when behavior conveys information, so if a group of people are making a certain choice then it is may be optimal for subsequent decision makers to ignore private information and follow the crowd. They base their decisions on what others do and ignore their own preferences or private information, leading to a cascade of information. The important feature of information cascades is that people can see what others do but not what they know. Sequential decision-making is also critical to this result as earlier decision makers “share” their information via their behavior. Note that information cascades will increase the network size leading to network effects, which will exacerbate the initial effect. This can result in an outcome where few products enjoy a large number of buyers and many niche products co-exist, an illustration of the Power Law. However, when decisions are made simultaneously and based on information that is independent and identically distributed, information cascades cannot arise because there is no leader to follow - individuals make their choices at the same time.
In the discussion of granularity, we considered the network economy where connectedness mobilized free information flow. However, in some markets ownership of information can be effectively fenced in and managed like a private good. BD is information that can be managed as a private good. Each transaction leaves behind crumbles of personal information, adding to the stock of data owned by the seller. This data set is the resource that sustains behemoths by generating revenue in two ways. First, it promotes efficient, and therefore valuable, advertising by targeting the critical demographic group. Second, it enables superior product development. As more transactions or sales occur, more data are generated, which in turn can be fed into algorithms to target buyers and advertisers. A revenue-generating feedback loop is created and sustained by demand and supply side economies, strengthening the behemoth status.
However, there is granularity at the product level, or product differentiation. Indirect network effects promote links between product sub-groups across these firms, so firms choose compatibility as a strategy to increase their individual market. For example, the iPad and the Kindle Fire are both in the same product category. Having access to a vast library of content via Kindle makes the iPad a more valuable device, which is the indirect network effect. iOS and Android are in the same product category, so that the more smartphone owners that exist, the more valuable is the ownership of smartphones. Amazon Web Services, Google Cloud and iCloud all offer Cloud products to users, encouraging the use of Cloud-based services in the first place. These links across product categories suggest a sort of granularity even while the umbrella organizations are large financially unified firms.
When firms increase compatibility across product categories, they enable users to segment their consumption into functional categories, purchase different categories from separate firms. Functionality becomes the product characteristic, not the business-defined category. For example, the iPad can be a reading device competing with the Kindle Fire, or it can be complementary with the Kindle content library. While the design engineers at the firm choose compatibility, the user chooses functionality. Take the case of the iPhone. Regardless ofthe intent ofthe iPhone designers, the device may be used more as a tool for business information management. Small businesses with dispersed service locations can have their employees send snapshots of their completed work, together with invoices, to central management for more efficient and coordinated billing.
The Amazon phenomenon of winner-take-all can co-exist with the granular market structure. These are markets where a single firm dominates but there is space for a multitude of niche firms that cater to smaller consumer segments, which were previously not served or underserved. The Internet itself can be a driving force of long-tail markets since buyers and sellers meet on the global, digital marketplace. Neither distance nor geography presents barriers to transactions. The term “long-tail” comes from the observation that the distribution of revenues over all firms follows a Power Law. So, for example, the fraction of firms that have k dollars in revenue is distributed as k where c is some positive number. An important market where this phenomenon is demonstrated is the labor market. There are lots of jobs listing mass-market skills and few jobs for specialized skills. The important point to note is that these specialized skills would have zero listings without the Internet, so the fact that this part of the distribution is “tailing” away toward zero but does not hit zero is the “long tail.”