A recent editorial in the Wall Street Journal argues that blockchain proponents must make their business case at the network level, not the level of a single enterprise. This is because the unit of competition for organizations is now the network, and because most of the benefits from blockchain come when multiple organizations connect their businesses through blockchain. This means that blockchain suppliers and their customers must design and test minimally viable ecosystems rather than minimally viable products.
The idea that some technologies provide value at the networked ecosystem level is not a new argument, nor is blockchain the first technology for which this argument has been made. In their 1999 book Information Rules, Carl Shapiro and Hal Varian made similar arguments for transportation technologies such as railroads, ports, and container shipping and early communication technologies such as the telegraph and the telephone. Building from their lifetime research on information economics, their book discusss the economics of networks, often using different words such as network effects, critical mass, and switching costs. But the essence of their message was similar to the editorial on blockchain. Networked ecosystems have existed for decades and past ones can tell us a great deal about new ones.
But what about the early diffusion of network technologies, something that Information Rules did not address? In my historical analysis of six communication technologies – telephones, wireless systems, fax machines, Internet mail, and video conferencing – I found that the initial diffusion of them were driven by small networks of users, and thus it took decades to build large networks. The small sizes of these networks also meant that many different suppliers were able to co-exist, each serving different sets of users. Reported in a 2009 Telecommunications Policy paper, tens, hundreds, and sometimes thousands of networks appeared in the U.S. for each of these technologies during the early years of their diffusion.
Initially the different networks served different sets of users and thus few users were connected to many other users. Over time, the number of networks increased with some growing faster than others. The big growth occurred when different networks became connected, often after costs had significantly fallen, thus leading to progressively larger networked eco-systems. However, while the last half of this diffusion receives most of the attention, it is the early part of the diffusion that is relevant for blockchain suppliers.
Consider the telephone. Hundreds of telephone systems emerged in the U.S. in the late 1800s, first through the leasing of systems to individual enterprises in 1877 for internal communications and later through public systems. Enterprises purchased telephone systems to enable communication within the enterprise, often connecting managers with factory floors in order to transmit changes in production schedules. The real growth in these fragmented networks, however, began after the expiration of Bell Telephone’s patents in 1894, enabling thousands of new systems to be introduced most of which served different sets of users.
These systems can be roughly placed in three categories: commercial services, mutual systems and ‘farmer lines. For the commercial services, the new entrants targeted lower income groups in cities many of which displayed little interest in communicating with higher income groups, thus enabling the new entrants to compete with Bell Telephone. Merchants implemented non-profit mutual companies and farmers implemented so-called farmer lines. By 1902 there were more than 6000 mutual systems in the U.S. and 222 mutual companies were started between 1900 and 1917 just in Southeastern Iowa, or more than eight times the number of commercial companies. The farmer lines often used barbed wire fences as wires and did not include switchboards, thus making the systems extremely inexpensive and fast to implement. Their downside was poor voice quality and no privacy; anyone with a connection could hear the conversation.
The growth in the number of these small fragmented networks enabled diffusion rates to accelerate for telephones in the late 1800s and early 1900s, with penetration rates exceeding 20% by 1900. The extremely fragmented nature of these networks can be seen in the small number of average subscriber households for the farmer lines (11), mutual systems (90), and even non-Bell commercial systems (700) in 1902. Apparently, these small networks provided large amounts of value in spite of the small number of participating households. Many argue that the large number of these commercial systems, mutual companies and farmer lines, which did not emerge in most of the ‘‘centrally’’ controlled European countries, was a major reason why the penetration rate of telephones in the U.S. had exceeded that of most European countries by 1920.
Similar stories can be told for wireless systems, fax machines, Internet mail, and video conferencing, technologies that took many decades to diffuse and involved many fragmented networks. Like the telephone, each of these communication technologies initially diffused through enterprise networks. For instance, wireless systems began with networked eco-systems of police cars in the 1920s while fax machines began even earlier with networked eco-systems of reporters, newspapers, and newspaper agencies. This early diffusion provided value to users and set the stages for subsequent diffusion by providing success stories. The benefits from these technologies were transmitted to others through the media and word-of mouth.
Unfortunately, these stories and the lessons embedded in them are largely forgotten now because most attention is directed to the last stages of diffusion when the winners are determined. Many people think wireless systems began with GSM or fax machines with Japanese manufacturers, when these cases were preceded by decades of earlier systems. There is also too much emphasis placed on capturing value from a new technology and not enough on how to create the initial value, particularly through the creation of small fragmented networks.
What should we learn from these examples? First, the editorial on blockchain is correct. The emphasis should be on the network and not the individual enterprise. Second, these networked ecosystems can take a long time to construct. Not only are the technologies initially too expensive or have insufficient performance, the real benefits come as a network grows in size. Third, blockchain promoters should look for small ecosystems that can provide value even if the participants are not major players in the global marketplace. Farmer lines provided value even though the average number of households in a network was only eleven. Fourth, building small systems can be faster than building large systems and thus enable success stories to be created more quickly. In summary, blockchain suppliers should be looking for small installations that can quickly provide a lot of value and that will demonstrate the economic benefits of blockchain.