Oftentimes the evolution of the blockchain industry is compared to the evolution of the Internet. This comparison is apt, and provides a guideline for the adoption curve of a transformational technology. However, the Internet is not the only industry that can provide a guideline for the pace and shape of development in the blockchain industry.
The below post highlights a wide range of industries and the ways in which they developed to draw parallels to the blockchain industry that extend beyond the often cited comparison with the Internet.
The Pace of Industry Development
Many people believe that the blockchain industry is at a stage similar to that of the Internet in the early-to-mid nineties. This causes them to stand waiting for the development and release of a “killer app” akin to the Mosiac Browser. Some are frustrated that after ten years of existence this “killer app” for blockchain technology hasn’t yet been developed. This frustration may be mitigated by looking beyond the traditional comparison with the Internet, to the development of other transformational technologies.
Let’s look at Artificial Intelligence. As explained in the book Machine, Platform, Crowd, the industry has been evolving since the early 1960’s. Early on, the AI community split into two camps, those that believed “symbolic” or rules based AI would prevail and those that aimed to build statistical pattern recognition systems. However, repeated failures of both approaches caused an “AI winter” to all but stop the development of the technology in the 1970’s. Even still, some researchers continued to work on the problem throughout the 1980’s, building systems with back-propagation capabilities. Despite these advancements, wide spread applicability was still constrained by limited computing power and limited data.
In other words, AI wasn’t held back by lack of a “killer app”, its progression was constrained by the pace of development of complementary technologies. The popularity of video games resulted in wide spread usage / production of GPU units, which are well suited to the parallel processing required by machine learning, in the late 1990’s. The cost of computing was then driven down by Moore’s Law and the proliferation of cloud computing (AWS launched in 2006.) The proliferation of Big Data (Hadoop was created in 2005) was equally important in the development of the AI industry. It was only after these developments that AI applications became truly viable on a large scale.
Avoiding a debate about the directionality of the infrastructure — app cycle, let’s consider the possibility that, scalability and other technical issues aside, the constraints on the blockchain industry are more akin to the constraints on the AI industry prior to the progression of complimentary, emerging technologies. Many of the specific use cases blockchain is well suited to address rely on integration with other technologies. Supply chain use cases, for example, require IoT integration, M2M communications, and in many cases 3D-printing technologies. Stable coins and tokenized assets require technologies that allow us to better bridge the physical and digital worlds without relying on a third party. The use cases of the future may require technologies or inputs that we haven’t yet heard of or capabilities that have not yet been developed. AI lay in wait for decades before complementary technologies were able to support it.
Increasingly, as the pace of technological innovation accelerates, what seems to be needed in order for any / all of these technologies to succeed is their convergence.
While some lament the pace at which the industry is evolving, whether you consider it to be progressing quickly or slowly is directly related to your frame (industry) of reference. However, there are elements of the blockchain industry that are not comparable with the industries that have developed thus far. The instances in which blockchain technology makes sense are by definition the use cases in which there is a low level of trust and coordination and a high level of fragmentation.
Therefore, unlike other technologies which merely require an MVP (minimum viable product), blockchain technology requires an MVE (minimum viable ecosystem), which is clearly a much higher hurdle to clear.
Blockchain technology also enables interaction to occur in such a fundamentally different way that enterprise adoption may require a mindset shift. This could mean that enterprise adoption occurs more “generationally” than other technologies, constrained until those in decision making positions are ready to embrace an entirely new way of thinking. In other words, this transition will take time. The industry will develop gradually. The development of similarly transformational technologies indicates that it will be worth it.
The Shape of Industry Development
Now let’s look across industries to assess whether or not the industry will develop “vertically”, where value and usage accrue to a select few blockchain platforms, or “horizontally”, where value and usage accrue to many unique blockchains which are used for specific use cases. Many prominent industry figures have opined on this topic. Let’s look beyond the blockchain — specific arguments that can be made for one path versus another and instead draw insights from parallels with other industries.
We will look at specific industry examples in more detail below, but at a high level, the evolution of many industries can be divided into three stages: platform fragmentation, product proliferation, and aggregation.
Platform Fragmentation: Once broad product-market fit is established and core market adoption is achieved, industries tend to fragment as they target more users via specialization. Example: Cable Networks — MTV.
Product Proliferation: As the industry fragments, barriers to consumption are typically reduced which causes supply to increase to accommodate this incremental demand. Example: OTT Content — Netflix.
Aggregation: This oftentimes generates a need for an industry player that can provide cohesion to the fragmented system as user search costs and or / friction increases along with fragmentation. Example: Different OTT Content Platforms Become “Apps” on Entertainment Operating Systems — X fininty1.
In the past, these players have been “aggregators”, aggregating the fragmented ecosystem onto one platform through which users can compare / contrast / access all the disparate offerings within an industry. In our current world, where the trend is towards decentralization (proliferation of edge / mesh networks, popularity of crowdfunding, visions of Web 3.0, etc.), the “aggregation” stage would be better described as the creation of an interoperability or connectivity mechanism.
An “aggregator”establishes relationships between fragmented product offerings while allowing them to exist independently and to thrive. This is in stark contrast to “consolidators,” which by definition centralize power. Whereas industry value tends to accrue to “aggregators,” consolidation usually results in a “conglomerate discount” to valuation.
In the Media industry, fragmentation occurs down to the smallest unit that can be consumed in a process referred to as “unbundling.” Generally, as content “unbundles”, it is enabled to reach a wider audience as monetary barriers to consumption are reduced. Supply increases to meet the increased demand as more participants enter the market. Let’s look at some specific examples:
News: Newspapers began as a bundle of print content. Digital news platforms such as Yahoo or CNN.com increased distribution and also allowed for the consumption of smaller units of content (a single article, for example.) This in turn led to a proliferation of specialized digital news sources (TechCrunch, Bloomberg, etc.) This became overwhelming to consumers, which turned to aggregation platforms such as Google or Twitter to organize all this content, lower search costs, and, in some cases, to create a more personalized “newspaper” (ie. newsfeed)
Music: Albums began as a bundle of songs. As music became digitized, platforms began to emerge to digitally distribute songs (Napster, LimeWire, iTunes, etc.) which reduced the minimum unit size of consumption. Increased distribution channels also lowered the barriers to entry for musical artists. This caused a proliferation of content and overwhelmed consumers. Aggregation platforms such as Spotify emerged to organize all this content, lower search costs, and create personalized playlists (re-bundling.)
TV: Media began with broadcast networks which aimed to reach large swaths of viewers. Networks naturally developed to allow for more segmented targeting by appealing to sub-segments of viewers. The production of digital content (OTT) began to proliferate, which lowered minimum consumption costs to ~$10 / month. As demand increased with lower barriers to consumption, an explosion in content creation followed. Platforms such as Netflix and Hulu, themselves aggregators in a way, began to provide personalized viewing experiences. As the number of these OTT content platforms also proliferated, it required an “aggregator of aggregators” (Amazon, Xfinity1) to allow viewers to organize, search, and access all of this digital content from one place.
While the main parallel relates to fragmentation and re-aggregation, another important component of the media industry’s evolution is the impact that “unbundling” has on consumer economics. “Unbundling” increases accessibility for consumers by lowering monetary barriers to consumption (i.e torrenting is free, the price of an OTT subscription is less than a cable subscription, etc.)
The consumer economics of “unbundling” are not dissimilar from those that can be achieved by applying blockchain technology to inefficient or rigid processes.
Blockchain technology can reduce the cost of entry and enable broader participation in activities such as fundraising, access to credit and /or social services, and international payments. If the above model holds, the first step will be that the blockchain industry fragments (more on this below.) Then as it “unbundles” more services, enabling broader participation in these activities, supply should also increase in order to meet this incremental demand.
The parallel does not hold on the economics of production, however, as one of the main enablers of the unbundling / fragmentation of the above mentioned media industries is the zero-marginal cost of digital distribution. This property does not extend to the blockchain industry where the initial costs of developing a new protocol are currently high.
*** Many of the ideas expressed above rely on and expand upon the ideas of my former colleagues at Barclays Capital, particularly Kannan Venkateshwar, Head of U.S. Media, Cable, and Satellite Equity Research.***
Parallels to the Blockchain Industry
While none of these industries are directly comparable, if the development of these industries serves as any indication of the way in which the blockchain industry will develop, the industry will fragment into use case specific blockchains. As explained in greater detail below, meaningful value will accrue to the companies that can create a way for fragmented blockchains to communicate and connect (ie. “aggregators”.)
Phase 0: The blockchain industry began with two dominant platforms, Bitcoin and then Ethereum.
Phase 1: We are now seeing a fragmentation of the ecosystem to meet the needs of specialized use cases, which should continue at least until the trade-off between decentralization, privacy, and scalability is no longer required. Given the constraints imposed by this trilemma, it makes sense that different protocols would be needed to accommodate different use cases. It also probably doesn’t make much sense to expect the same platform to support a country’s digital identity system while also supporting applications such as CryptoKitties. The reasoning extends beyond scalability issues. As highlighted by Cosmos, “ Much like communities, companies and nation-states, each existing cryptocurrency is born with the seed of some cultural ideal.” In order for these companies to express those ideals, they will create protocols that grant them the flexibility to do so. From all lenses, it looks unlikely that there will be “one blockchain to rule them all.”
Phase 2: As rigid processes and closed systems are “unbundled”, wider spread usage and adoption will occur across these multiple blockchains. This will increase friction and transaction costs for consumers. As the number of blockchains increases, so does the complexity of managing a number of different tokens and assets siloed within disparate ecosystems that have no way to communicate or connect.
Phase 3: Companies that can create common standards or mechanisms for interoperability and / or connectivity between blockchains will occupy a central role in industry development (more below.) There are many projects currently working on becoming this “aggregator” or “Internet of Blockchains” including Cosmos , Polkadot, and FourthState Labs.
The idea of interoperability in the blockchain industry is not new. However, looking at other industry evolutions makes it seem clear that the blockchain industry will continue to fragment into specialized blockchains and that as this happens, value will accrue to the platforms that facilitate interoperability, communication, and connectivity between chains.
Network Effects in the Industry
As the blockchain industry fragments, new entrants will need to offer defensibly differentiated value propositions if they are to disrupt established network effects. An “aggregator” or connectivity mechanism will be crucial to further development of the industry. Let’s now look at several industries where multiple networks have had to compete and co-exist side by side.
E-commerce and Social Networks
While Amazon and Google are perhaps the two platforms which best exemplify network effects, they have also both failed at incentivizing users to join new networks at one time or another. Amazon was not initially successful in its attempts (Amazon Auctions, zShops) to compete with eBay. Amazon failed at competing with eBay (selling used goods) until it realized that it couldn’t compete by trying to be “ a better eBay.” With the introduction of the Single Detail Page, which gave Amazon customers the option to view new or used versions of a product, the company began targeting Amazon’s own customers instead of eBay’s. This then offered eBay sellers a more compelling value proposition to join Amazon’s network as it opened up a new market to them (Amazon’s customers.) This was enough to start the flywheel needed to develop a network of both buyers and sellers.
Prior attempts failed because they tried to create a marginally better, copy-cat network. This requires all network participants (buyers and sellers) to overcome the “gravitational inertia”of their current network and move to a new network together as the value of an incremental feature will never offset the magnitude of switching costs for an individual user.
Google+ also failed with its attempt at a “me too” social network. Again, this effort failed because it wasn’t defensibly different from other, more established social networks (Facebook.) It’s much easier for an established network to copy a competitor’s new feature than for a new entrant to create network effects from scratch.
While there are valid reasons for creating specialized protocols to accommodate differing use cases, blockchain platforms that focus on making marginal improvements over existing networks are unlikely to succeed. Instead, new networks must offer a value proposition compelling enough to justify switching and unique enough that competitors can’t easily replicate it.
One exception to this rule is if a player is willing to add a feature that the incumbent is not willing to add. This is how TaoBao beat the already established eBay in China. TaoBao introduced direct messaging between buyers and sellers, which eBay wanted to avoid since side channel communication between buyers and sellers increased the likelihood that transactions would be conducted outside the eBay platform. eBay would then forego the associated transaction fees.
Blockchain networks are doing exactly this, adding features (trust, transparency, immutability, and direct P2P interaction) that the incumbent platforms are unwilling to add.
While new blockchain networks will have to offer strong value propositions in order to compete with established blockchains, blockchain networks in general should be able to beat out traditional networks.
***This section summarizes and expands upon the ideas of former Amazon and Google engineer, Steve Yegge.***
Once the industry has fragmented, these disparate networks will need to communicate with each other. The telecom industry is a prime example of an industry that requires the bridging of networks.
Due to capital and regulatory constraints, it is not feasible to build a global telecommunications network. Therefore, cross-country communication often requires cooperation among carriers. Sometimes this occurs in the form of roaming agreements. Other times it has required agreement upon global standards. Pre- 4G LTE, Europe operated according to GSM standards while some U.S. carriers (Verizon and Sprint) operated according to non-compatible CDMA standards. Without compatibility, a Verizon customer traveling in Europe wasn’t able to use their phone while overseas. This obviously created a high level of friction for consumers. Telco companies again realized that they had to work together to develop a mechanism to allow consumers to hop from one network to another more seamlessly while moving outside a given carrier’s coverage zone. The 4G-LTE global standards were designed and expected by all major carriers with this in mind.
Issues with compatibility across geographic networks can be compared to managing multiple native tokens across multiple blockchain ecosystems that can’t easily communicate with each other. This currently requires conversion into other currencies via an exchange (sometimes multiple times) before being able to use assets in another ecosystem.
Like the telecom industry, the blockchain industry is a network of networks and will need to continue to set standards aimed at facilitating more seamless communication between these networks.
Telecom is also a highly regulated industry and differences in regional regulations have caused the industry to develop very differently in different geographies. For example, the European Telecom industry is much more competitive (harsher anti-trust enforcement) than its U.S. counterpart, with +10 different major carriers relative to ~3–4 in the U.S. In the U.S., operators tend to lease cell towers from third-parties while a higher percentage of European operators still own their tower infrastructure. In other words, regulatory and regional differences have created a less restrictive operating environment for U.S. telcos relative to their European counterparts. As the blockchain industry fragments, regional regulations and adoption patterns may heavily influence which projects succeed.
The mechanism by which these networks connect to and communicate with each other is of central importance to an industry. Comparison with the networking infrastructure industry illustrates this point clearly.
With the advent of cloud computing, enterprises have shifted from on-premise data centers to third-party colocation facilities. These third-party facilities have traditionally been divided into two models: wholesale, custom built dedicated facilities also called “server farms,” and retail facilities that focus on interconnection via direct fiber cross connections between customers. These retail focused companies have created a neutral, third-party location for companies to directly connect to each other, to peer or exchange traffic and /or data securely and rapidly.
The infrastructure providers that have built their business on being a facilitator of connectivity between disparate enterprises, ISPs, and telcos have accrued the most industry value, historically trading at a ~4x premium to their wholesale peers. Much of this premium is related to the network effects created by facilitating communication between transacting parties within a densely interconnected data center.
The parallel with the networking infrastructure industry perhaps best underscores the need for connectivity between chains and / or side chains and underscores the value of platforms that allow for cross network communication and connectivity. This is the value of an “aggregator.” This is the value of interoperability.
The parallels between the blockchain industry and the early stages of the Internet are clear and well cited. However, limiting ourselves to one frame of reference is just that: limiting. Looking at a broader range of industries indicates that interoperability will play a central role in the development of the industry. Cross-industry comparisons also indicate that the development of the blockchain industry is unlikely to be quick, unlikely to occur “vertically,” and unlikely to take place uniformly across the globe. That doesn’t mean that it will be any less transformational.
***Many thanks to Wesley Graham for the feedback.***