The Blockchain Use Case That Is Ahead of Its Time and How Machine Learning Will Be the Catalyst That Gets Us to the Promised Land
In this piece, I will share lessons that I have learned while advising companies on blockchain strategies through my work at Fraktal Group. Fraktal is a blockchain advisory that I co-founded with Kevin Xu, the former Director of Cryptoeconomics at ConsenSys Token Foundry. We engaged with companies of all sizes, ranging from contractual partnerships with newborn start-ups and successful growth-stage companies to extensive consultations with some of the largest established enterprises, such as Foxconn. My experiences with these companies, each in different industries, has revealed to me that the most valuable applications of blockchain technology are still ahead of their time. In this piece, I will describe why most companies are unprepared to usher in the largest promises of blockchain and why machine learning (ML) will be a primary catalyst that gets us to that promised land.
Please note that this piece will focus on how blockchain technology can help existing businesses coordinate any data, not just financial data. As such, the merit of other blockchain narratives, such as Web 3.0, decentralized finance, and cryptoasset speculation are out of scope.
I have been living and breathing blockchain for years. What started as a curiosity turned into a full-blown career when I quit my stable finance job to navigate the wild world of frontier technology. I used my extra free time to read through every white paper I could find, and to consume news fast enough to keep up with the rapidly moving industry. After exhaustive research, I gave a seminar on blockchain fundamentals and wrote a white paper on blockchain interoperability. This work caught the attention of an ICO-focused venture capital fund that hired me as a venture partner to assess their prospective investments and source deals. Through my work with this fund and their portfolio companies, I came to realize that nearly all the cryptocurrencies that crossed my desk suffered from design flaws, primarily excessive token velocity, that would doom these cryptocurrencies, or “tokens,” to failure even if the underlying networks were successful.
As a result, I co-founded Fraktal Group to identify and repair token design flaws at scale. Our token engineering began with crypto startups, but ultimately helped me realize that blockchain’s potential went far beyond cryptocurrency. We developed a thesis that blockchain is of most value in the creation of marketplaces (see below for more on marketplaces). The trouble with marketplaces is that they face a chicken-and-egg problem; no buyers want to enter a market without sellers and no sellers want to enter a market without buyers.
Our solution to this bootstrapping problem was to find companies that could seed one side of the market themselves so that we could focus our efforts on growing the other side of the market. This led to our work with a growth-stage company that was large enough to seed the marketplace with its existing growing business, but nimble enough to be receptive to using a disruptive technology like blockchain to unseat its incumbent competition. We worked with this company to create exciting new marketplace designs, but the bigger the entity to seed a marketplace the better so the next clients we pursued were enterprises.
This pivot to enterprises led us to conversations with some of the world’s largest companies. The use cases we explored with them centered around establishing data standards and making their data more transparent so that service providers could more effectively cater to their needs. While these enterprises had tremendous potential to create blockchain-enabled marketplaces that could facilitate streamlined services, we found that most were hesitant to make such a large technological leap. We faced the classic innovator’s dilemma. Enterprises did not feel a sense of urgency to invest the time and capital necessary to spearhead the development of new industry infrastructure, even if spearheading that effort would position them well to be industry leaders in the future. Moreover, even if they did have a sense of urgency, they still had major data standard issues to handle internally before they would be ready to discuss data standards across partner organizations.
I was left to wonder, what is it going to take for industries to prioritize blockchain? To answer this question, let’s first understand today’s blockchain headlines and identify what these efforts are missing.
The recent news of Facebook and J.P. Morgan releasing their own cryptocurrencies has breathed new life into the blockchain space. While these announcements do represent major steps in the development of the blockchain industry, these proposed cryptocurrencies represent a 1-to-2 style of innovation rather than a 0-to-1 innovation that more creative uses of blockchain technology can enable. Both Facebook’s Libra and J.P. Morgan’s JPMcoin are stable coins. The purpose of a stablecoin is to track the value of another unit of account, such as the US Dollar (JPMCoin), a different stable currency or asset, or a bundle of stable currencies and/or assets (Libra). The benefit of these stable coins is that they can be used to transfer value without relying on legacy financial infrastructure. This is akin to driving a hovercraft over the traffic of a congested, pothole-ridden highway. It makes sense for J.P. Morgan to be one of the first to implement blockchain because even a small incremental improvement could lead to massive economic benefits when used to move trillions of dollars.
While large institutions are focused on using blockchain to transfer value, other disruptive applications are being overlooked. Blockchain is a tool that is best used to coordinate data between different stakeholders, but this data does not need to be limited to account balances or involve direct transfers of value. This data could be store inventory, shipment status, financing offers, financing requests, or just about any information that entities care to coordinate.
To understand data coordination’s potential, let’s dive into a real-world example and trace its trajectory to see where it might be heading.
Blockchain Beyond Currency and Into Marketplace Creation
One of the most ambitious blockchain projects underway today is IBM’s Food Trust. The Food Trust is novel global supply chain infrastructure that is used to track food across the world. It is a collaborative network of growers, processors, wholesalers, distributors, manufacturers, retailers, and others, enhancing visibility and accountability across the food supply chain. Built on IBM Blockchain, this solution connects participants through a permissioned, immutable, and shared record of food provenance, transaction data, processing details, and more.
Now, instead of relying on manual data entry or bespoke direct connections, these entities have come together to develop a common language, also known as a “protocol,” that their systems can use to speak with one another to share data and streamline processes. Not only does this save time and money, it also opens up new territory in which to innovate. This innovation is made possible by the unique characteristics of a blockchain network, which makes it easy for multiple organizations to create shared data and for third parties to access that data. Providing third parties secure access to the data flowing through the network enables these companies and entrepreneurs to build applications on top of the network to provide improved or novel services to the network’s existing ecosystem.
For example, in addition to simply providing shipment status, this network could also give companies the option to coordinate and provide visibility into each other’s invoice statuses. Instead of dealing with today’s opaque and manually intensive process of sourcing invoice financing, an entrepreneur could build an application that automatically finds and tracks invoices that fit a given risk profile and offer the most competitive rates because of all the costs that the application saves.
While the creation of a consortium requires investment and coordination from multiple stakeholders, once the consortium members begin utilizing the infrastructure, they will want this infrastructure to handle as many of their needs as possible. This will be especially true since the services provided on the infrastructure will likely have structural cost and capability advantages over existing service providers. As the network grows, more service buyers will beget more service providers and vice versa.
Additionally, the more data that the blockchain network shares, the more opportunity there will be for entrepreneurs to find ways to service the network. Corporations have been increasing their investments in their venture investment arms almost every year for the past 9 years. Instead of only investing capital in select companies, these corporations could also share their data. Removing this data moat would lower the barriers to entry for entrepreneurs, which has the potential to usher in a new wave of innovative activity, especially if entrepreneurs can plug into a consortium’s workflows without having to deal with the long enterprise sales cycles that often kill startups. The creative energy of entrepreneurs has historically led to game-changing products and business models. Now, enterprises can have the opportunity to have this creative energy work for them rather than against them.
So, if blockchain networks can offer such a promising future, why aren’t we seeing more companies adopt the technology? While many companies have not yet wrapped their head around this vision for the future, those that have still face hurdles to adoption. The primary hurdle is data standardization.
How Machine Learning Will Accelerate Blockchain Adoption
IBM’s Food Trust is possible only because multiple entities came together and agreed on a protocol they will use to coordinate their supply chains. As a management consultant at Oliver Wyman, I spent years conducting big data analytics for Fortune 500 companies. This work made it painfully apparent how rare it is for companies to have effective data standardization and management within their organizations, which is a prerequisite to establishing a data standard with other organizations. Internal standardization of data has historically been only a moderately high priority for CEOs who increasingly see the need for their business intelligence teams to have such data to generate insights.
The urgency for companies to standardize data has recently dramatically increased as ML moves from a disruptive threat to an actual disruptive force. CEOs now realize that ML is enabling companies to operate smarter, faster, and cheaper. Boards throughout the world are asking themselves what their ML strategy should be and how they should implement it.
While there are four types of ML, supervised ML is the most mature and is causing more disruption each day. It works by training an algorithm with labeled historical data so that it can make predictions. An example ML application in use today is customer service tickets. When most companies receive a customer service ticket, they need humans to read the ticket to route it to the appropriate department for a response. The startup company Forethought is changing this status quo. If a company has tracked the content of each ticket it receives and where each ticket was routed, Forethought can feed that labeled data into its ML algorithm to automatically route tickets with a high degree of accuracy.
While there are techniques to use ML without labeled data, it is usually more effective and easier to implement ML with high-quality labeled data. The easiest way to generate and organize labeled data is to have standardized digital processes that automatically generate this data as employees execute their actions. As companies standardize and digitize more processes, they will by default also define their internal data protocols. As companies develop clearer internal data protocols, those companies will be better prepared and more interested in working with other companies that have also standardized their processes and data. As the activation energy required to form inter-company protocols decreases, and the benefits of creating shared infrastructure becomes more apparent, more industries will develop blockchain-based infrastructure.
Nothing is ever certain when it comes to what frontier technology will create. However, what I have presented here is a perspective that is supported by the unique set of experiences I have had while working with my clients, speaking with enterprises, and collaborating with my blockchain industry peers daily.
Nearly all of the work done in the blockchain space today is happening within the out-of-scope narratives that I mentioned in the introduction. I hope that this piece has helped expand your imagination of what blockchain technology is capable of enabling so that we can begin to take steps toward this promising future.
Please reach out if you are involved with an ML company targeting business process improvement or financing activity, work with an enterprise interested in developing an ML or blockchain strategy, or are simply interested in learning more about this vision. I welcome any feedback on what I have proposed here.