How Blockchain Networks Will Disrupt the Businesses of Today, and Signal the End of M & A
I’ve been thinking about a stark bet to take, which is that the major tech companies of today: Google, Facebook, Microsoft, and Apple, will be far less significant, or even inconsequential players 10 years from now. I believe blockchain networks will change the fundamental paradigm of how businesses operate and that they will undermine the current mechanisms that the dominant tech companies of today use to sustain their competitive advantages.
Consider that in 2000, most people wouldn’t have believed that within 10 years Blockbuster would be completely bankrupt. Blockbuster, along with many other businesses of its generation, was blindsided by internet-first companies like Netflix. They didn’t understand how fundamental a shift the world wide web enabled. Blockbuster’s distribution channel advantage of owning a massive physical retail store distribution network was completely undermined by the ability to digitally distribute content over the web, with infinite-shelf space and selection, regardless of location.
If we believe technology’s adoption generally accelerates, then the shift in business operating principles caused by the blockchain may undermine today’s companies faster than the web undermined the previous generation of dominant retail-based companies. Blockchains also have built-in incentive structures that will accelerate the transition, in so far as they encourage speculative behavior, and use mechanisms that resemble pyramid-schemes to kickstart their adoption.
My hope in this post is to outline how I think this transition may all go down, and what the interesting aspects of blockchain networks are that pose a threat to today’s way of doing business:
- Blockchain Networks Undermine Monopolies based on Control of Data, Social Networks or Compute Resources. The major companies of today’s era are based on monopolies and restricting access to some resource: whether that’s some dataset, compute power, or physical logistics networks. Blockchain networks decentralize and democratize ownership of data, compute and marketplace resources, undermining this competitive moat.
- You can’t acquire a blockchain network / protocol, only participate in it by purchasing tokens, so M&A is not a viable strategy for removing competitors in this new era. Historically, the model has been to acquire your competitors in order to preserve your monopoly and neutralize disruptive trends. Amazon, Microsoft, Google, Facebook, Salesforce and Apple have become masters of M&A and investing in their competitors. They’ve been able to sustain their edge by nipping any threats in the bud, but this strategy may not be a viable defense against new blockchain based competitors.
Lessons from the Web
When the World Wide Web started to really take off in the 1990s and early 2000s, the pre-web businesses of that era were very ill equipped to deal with the new operating paradigm of business.
There were several new aspects to doing business as a web-first company:
- Low up-front capital startup requirements. Typically web businesses had far smaller capital costs to get going. You don’t need store-fronts, and sometimes you don’t even need warehouses.
- Infinite shelf-space. There was the ability to hold infinite shelf-space and inventory: think of Amazon’s selection of less common books vs your local bookstore, or Google and Wikipedia’s ability to find you very specific content vs. your local library. The internet enabled immediate access to the long-tail of demands.
- Digital distribution. The web enabled businesses to instantly reach countless markets by going “online” — i.e. people could visit your website from anywhere
- Personalization. The web gave businesses the ability to personalize storefronts, product offerings and experiences to the individual user due to the digital nature of products.
- Many-to-many networks & user-generated content. Before the web, newspaper classifieds were the primary many-to-many distribution channels: they were the marketplace — e.g. if you wanted to sell a house, or try to find someone to date you put up a listing in the classifieds section of the newspaper. The world-wide web created a new set of many-to-many channels, and for businesses that meant the ability to create monetizeable user-generated content and data at very low cost (this was largely the focus of Web 2.0)
The incumbents from the previous era couldn’t fully grasp the tectonic shift that was occurring as a result of these qualitative changes in how you could run a business. They couldn’t comprehend just how quickly they might lose their foothold as economic juggernauts in light of these new principles. They had no idea how to become “web businesses”.
There’s a book I enjoyed reading a few years back called eBoys: The First Inside Account of Venture Capitalists at Work. The book tells the story of the early days of Benchmark capital, which is a highly acclaimed VC firm founded in 1995 that got its foothold during the early days of the web. Benchmark’s huge early success was that they backed eBay with an investment of $6.7 million in 1997, which became a $5 billion stake by 1999.
In that book you get a glimpse of how Benchmark tried to partner with traditional businesses like Toys R’ Us to help them take advantage of the web’s new distribution model — to no avail. These partnerships were a complete disaster as the old generation of businesses had an operational and organization style that was fundamentally antithetical to the new way of doing business on the web.
Within a short span of time, Amazon crushed countless businesses like Barnes & Nobles, Toys R’ Us and RadioShack, while Netflix eviscerated Blockbuster, once a staple of many North American street corners. In a similar vein, Craigslist, Yelp, and others destroyed the cash cow of newspapers: the classifieds and listings business.
This past generation of businesses were entirely incapable of understanding and adapting to the principles of how to operate on the web. Their way of doing businesses was incompatible with that new paradigm. The lesson was that you can’t just graft a new ways of doing business on to companies that were born and based around an entirely antithetical set of principles.
Today, we are on the verge of another fundamental business paradigm transition (and really societal organization shift): blockchain networks. The major companies of this era will be equally ill-equipped to navigate this new economic terrain. All the major tech companies of today are under threat from blockchain protocols and new principles of operating, whether they realize it or not. About a decade after Netflix’s launch, Blockbuster was already largely irrelevant. Though it’s hard to imagine, a decade or two from now we may see Google, Amazon, Facebook and these other companies as irrelevant derelicts from the pre-blockchain era. Though they might want to adapt, it may be impossible for them to do so.
Distributed Data & Compute Networks. Protocols Not Companies
The major companies of today’s era are based on monopolies and restricting access to some “centralized” resource.
The types of centralized resources they control vary, and some of these monopolies may be easier for blockchains to displace than others.
Data & Attention Monopolies. One type of monopoly is a data and attention monopoly. Facebook and LinkedIn fall into this category as they own valuable datasets about people and their behaviors. Google controls the largest index of the web and an algorithm for ranking content. All of these companies have an immense amount of data and user attention as people spend large amounts of time on their sites, creating a data exhaust as a byproduct of their participation.
Compute & Infrastructure Power Monopolies. Another type of monopolistic advantage is one based on compute power. Amazon Web Services and the other cloud vendors operate massive data centers filled with computers that they manage and rent access to. As a startup it’s pretty hard to imagine spinning up your own data centers to compete with Amazon or Google Cloud directly. Today, if you want to start a web-business it’s likely you’ll run your company on top of their centralized data centers (public clouds).
Physical Resource Marketplace Monopolies. In other cases, the monopoly consists of having some sort of other large scale physical infrastructure network, often times coordinated by a marketplace. Uber, Lyft and the ride-sharing companies manage marketplaces that consists of fleets of vehicles, drivers, and passengers. AirBnB is a marketplace of rental properties and travelers. Amazon, while not operating a marketplace, has setup a large scale physical logistics and distribution network, and there are also marketplaces that have spun up around trucking and delivering goods. Uber, Lyft and AirBnB’s business are based on not-owning the inventory of vehicles and properties themselves, but instead acting as an interface and coordination mechanism for the marketplaces.
Let’s consider these three types of monopolies in turn, and then also look at how blockchain protocols might unseat them and provide viable alternatives to these centralized resources.
Centralized Compute Monopolies
In 2007 Amazon started Amazon Web Services (AWS) when they realized that by virtue of running their e-commerce business they had obtained massive amount of computing infrastructure and operational knowledge of how to manage clusters of physical machines. They realized that given their scale, they could manage and rent out time on their computers to other businesses who want to run software without having to buy and maintain actual physical machines. Before Amazon Web Services, if you wanted to start an internet / web based company, you had to buy and administer physical servers, often renting out space in buildings or colocated server centers in order to do so. This was a substantial fixed startup cost, which required considerable expertise and overhead to sustain. AWS made it an order of magnitude easier to start an internet company by letting developers be able to spin up servers without having to know anything about the physical machines their servers run on, or having to worry about issues like physical servers losing power or catching on fire.
Amazon Web Services and other cloud services such as Google Cloud and Microsoft Azure, have continued to scale these large consolidated data centers filled with computers, all based on that very compelling product proposition: write and run software without having to think about managing the physical hardware layer underneath.
Sidenote & Plug: I’m grateful for this amazing lift in the abstraction level of compute, and it also enables new types of businesses to be built like Clay (where I’m a cofounder), that are focused on creating interfaces to make it easier for an order of magnitude more people to build software.
The growth of AWS has been astounding, and they continue to scale as more and more traditional industries undergo digital transformations.
To illustrate how services like AWS are under threat from protocols, let me describe one distributed blockchain compute protocol that some of my friends started called Livepeer.
Livepeer describes itself as “Open-source video infrastructure services, built on the Ethereum blockchain”. Every time I talk to the LivePeer team, I get this feeling that if what they’re doing works, it’s going to completely unseat the existing players in their space, and cause a shift from centralized data centers to decentralized compute networks for this use case. And they’re not alone in their approach to building new compute networks, the same strategy they’re taking is being taken by other companies across many other interesting compute verticals.
So, what is Livepeer? Livepeer is a decentralized compute protocol focused on video transcoding. “Transcoding” video means turning videos from one format to another (e.g. MP4 to WMV), which is needed under the hood if you’re going to show videos on different devices: for example if you live-stream video from your phone’s camera to lots of other devices that might need to view the video in another format. Transcoding is an essential part of running most video-related business on the web.
Historically, people have turned to services such as AWS’ Elastic Transcoder Service, Mux, Brightcove’s Zencoder or others. These services do the transcoding work within Amazon’s data centers, running transcoding software on computers within the major cloud providers’ data centers.
Livepeer’s protocol aims to essentially offer a competing data center that you can run transcoding software on, except that this data center will not be owned by Livepeer or anyone for that matter. Instead of buying lots of server computers and managing their own data centers, Livepeer’s protocol incentivizes individuals and organizations to contribute their existing, underutilized compute resources to a decentralized network of computers that is managed by the protocol itself (i.e. the transcoding work is fanned out to a distributed set of machines that join the network). Livepeer’s protocol acts as the coordination mechanism, where instead of AWS doing the transcoding work, the same type of transcoding work is accomplished by a distributed set of participants.
Pretty soon, it’ll be just as easy to setup and use Livepeer’s distributed network as it is to use Amazon’s Web Services or alternative transcoding networks. More importantly, it’s conceivable that they could bring down the cost of transcoding to a small fraction of what AWS and the existing players charge, thereby undercutting the economic viability of the existing players. If the price dynamics shift quickly so that the price is a small fraction of what it would cost to transcode on a central network, and the product is easy to setup and reliable, then we’ll see a shift to Livepeer style services at a breakneck pace.
To go into a bit more detail, let’s talk a bit about the way Livepeer’s protocol works in practice, and how the incentive and token dynamics play out. I can participate in the Livepeer protocol by installing some Livepeer software on a server and becoming one of the “nodes” on the network that helps to run transcoding tasks. This “transcoding work” is what is sometimes referred to as “mining” on other protocols: doing some computing work in exchange for compensation. I’m incentivized to join the network because in exchange for contributing compute resources, I receive Livepeer token (or other tokens supported as compensation on the network) which can be used in turn to buy transcoding services from the network. These tokens represent stakes in the network, and as usage of the network increases so does the value of the tokens. These token based incentive structures encourage speculative style behavior. Just the way that eBays shares went from 5 million to 5 billion, LivePeer tokens could eventually skyrocket in price, and so people try to participate in the network early on.
Participation in these networks has become a new type of investment strategy. I can spin up nodes to participate in the network, acquire Livepeer tokens, and then sell the Livepeer token to someone who wants to do transcoding work. As long as there’s demand for transcoding on the Livepeer network, there’s a market for the tokens I get from spinning up nodes. As demand increases, the value of the tokens will also increase. This is why there’s benefits to being early in the blockchain “mining” game across networks, and why I’m super excited about services like Bison Trails that make it easy to spin up nodes and in turn acquire tokens on the various blockchain networks that are emerging. These new types of blockchain infrastructure providers make it easy to participate in networks for people who don’t know the details of spinning up and managing nodes, in a similar way to how AWS made it easy for developers to spin up servers without having to know how to manage physical machines.
(Disclosure: I’m a small angel investor in Bison Trails). The difference is that unlike AWS, spinning up nodes via Bison Trail creates compute infrastructure that does some useful work, but in addition also generates tokens I can hold on these various networks that accrue in value. It’s a new way to get both practical compute and financial participation in these protocols.
If Livepeer succeeds in offering reliable, scalable video transcoding at a fraction of the price of Amazon’s services, then the shift to their services will likely be as fast if not faster than the transition from Blockbuster to Netflix.
While Livepeer is focusing on creating decentralized compute networks for video transcoding, others are focused on creating similar distributed networks for file storage (companies like Sia and Filecoin compete with S3 and Dropbox), machine learning and rendering (Golem), and many more.
You could look at a map of AWS’s services and likely find a large subset that will likely be challenged by decentralized compute protocols.
One possible way things evolve is that specialized protocols emerge tackling each of the services that AWS offers. There could be various companies and protocols that offer some sort of bundling on top of these services (e.g to simplify monitoring and coordination). It’s interesting to consider what the Heroku of the blockchain compute world will eventually look like.
Centralized Data & Attention Monopolies
Companies like Facebook, LinkedIn and Google are valuable because they have the attention of billions of users, and the underlying data sets related to those users’ interests and behaviors.
Their business model is personalized advertising, and is underpinned by them keeping a monopoly on attention, along with granular targeted data about the interests of their users. They tend to give away their products for free because their real customers are the advertisers, and the real product they are selling is people’s data and attention.
In the case of Google, it’s hard to build a competitor to Google for several reasons:
- They have an enormous index of websites, and trying to index that much of the web and storing the index would be quite expensive for any startup at this point
- They have a world-class front-end interface UI for accessing this index, and the ability to load results at incredible speeds by virtue of operating a global Content Delivery Network (CDN).
- They’ve gotten really good at hiring some of the smartest people in the world, and paying them really well in order to reap the benefits of their work.
I’ve been thinking for some time about what a blockchain competitor to Google looks like, and I think it’s likely that eventually there will be blockchain based networks that give access to the pieces needed to build a competitor, but more on that in another post.
At a high level, I believe that we’ll see a decentralized index of the web built, and that blockchain protocols will incentivize engineers to participate in donating compute and storage for indexing, as well as incentivizing the creation of new types of ranking algorithms and user interfaces, leveraging the scalable blockchain compute layers that get spun up in order to run these new types of services. In this new world even the fourth advantage I mentioned, the ability to attract and retain talent, may be undermined as talented individuals find ways to participate in decentralized networks and receive substantial monetary compensation for building on top of the network without having to work for a centralized company like Google.
Another interesting headwind towards decentralizing these datasets is that companies are increasingly under pressure from government intervention to store less and less data about users in centralized data stores. GDPR in Europe is one example of this type of additional pressure that could encourage adoption of decentralized application approaches.
Physical Resource Marketplace Monopolies
Centralized marketplace companies like Uber, Lyft, AirBnB, Craigslist, and others are already largely protocol companies, though they just happen to not run on the blockchain. At the end of the day, blockchains are just protocols for building decentralized databases that no one entity owns. They are shared, ownerless databases which give participants permissions to access, and coordinate through them using the protocols that specify how data can be written and read. What is the significance of a shared database which no one owns?
A large part of what marketplace sites do is create shared databases (e.g. Uber is a database of the pool of drivers willing to pick someone up, and passengers looking for a ride, AirBnB is a database of properties for rent, and travelers looking to book them).
Uber, AirBnB and the other marketplace companies have a business model premised on taking a % fee on every transaction that happens in their marketplace. This % fee presumably goes to cover the cost of operating the marketplace and generating a profit for the company that manages it. These companies also have a monopoly that is a data monopoly, but it’s premised on them owning all of the supply-side listings. There’s no way for me to tap into AirBnB or Uber’s supply listing dataset and build my own interface for travelers or riders to book services.
New blockchain based marketplaces can deconstruct and modularize the different parts of these centralized marketplace based businesses. It’s likely we’ll see competitors to these companies spring up, where a protocol will encourage people to list their properties, and there will be little or no-fee involved with transacting on the protocol. This will also create a shared database of supply-side listings, and individuals and other organizations will be able to build their own interfaces for booking, and managing transactions on the protocol. In this way, the marketplace dataset and the front-ends and services for interacting with that dataset are decoupled, which means that there’s the opportunity to have companies that focus on just building new applications, interfaces and services on that shared resource.
Competition Without Control: You Don’t Control the Data. You Don’t Control the Compute.
The common basis for these companies that hopefully stands out, is that their businesses are very much predicated on control of some central resource, whether that be data or compute. Blockchain protocols challenge these business dynamics fundamentally. Organizations like Livepeer are focused on setting into motion the right incentives and underlying primitives for their protocols. Once in motion, the role of these organizations becomes somewhat more passive: they nurture and participate in the protocol, but they do not control it or own it. There is no CEO of the Livepeer protocol, there is no switch to turn it on or off. These protocols are distributed, and feel much more like biological organisms — they are loosely coupled, and not coordinated by a top-down, hierarchical corporate structure. When the networks are small enough, you might be able to acquire enough of a token stake to disrupt the network, or modify the operating principles in some malicious way. But as the networks scale, its very difficult to take control of a sufficient portion to affect these protocol based networks.
This introduces its own new fascinating set of interaction principles. In many ways, when I think about these protocols, I think of water flowing fluidly. Engaging feels like standing in a river. Trying to establish control in this new paradigm is like reaching into the river and trying to grab a fist of water — there’s nothing to hold on to.
These protocols represent new modes of organization that are entirely different from their predecessors. How to structure these protocols, how to make them resilient and resistant to control, and how to model and iterate on their dynamics is an ongoing conceptual exploration. How do you adapt a decentralized agreement? It’s much the same question as how do you modify a document like the US constitution. What are the protocols to modify the protocols?
To date, we’ve primarily achieved large scale projects by coordinating our resources under large top-down organizational entities such as companies and nation states. Amazon is able to build AWS because of the scope of its existing business. Google is able to index the web because they started early, and their monopoly power as a company compounded as they became the primary entry point to the web and custodians of the index. The United States was able to coordinate massive top-down projects to build nuclear capabilities and successful lunar missions. The reason that blockchains will be historic is because they allow for new, semi-anonymous modes of large-scale coordination, based primarily on incentive driven protocols rather than more centralized modes of coordination.
The End of M & A: You Can’t Buy What No One Owns
To date, Google, Amazon, Facebook and Microsoft have successfully kept growing their businesses in large part by acquiring any potential threats, and it’s proven a fairly effective strategy.
As examples, consider that in search Google purchased Metaweb which ran Freebase that nows serves as the foundation for the structured search results you often see as cards on the top of your searches. They also bought Aardvark, which could have offered alternative approaches to search by leveraging human powered responses.
Facebook purchased Instagram, WhatsApp and has tried to purchase Snapchat as well, in order to stem the rise of alternative social networks. Amazon has purchased most budding e-commerce threats such as Zappos, and Diapers.com, both as ways to enter new categories but also strategically to remove any adjacent competitors before they become viable threats.
I’m sure that Blockbuster wishes they had accepted the offer to buy Netflix for 50 million dollars in the early 2000s, but the past generation of companies wasn’t as wise to the potential disruptors as this generation is.
The strategy of M&A is well established, and much of corporate strategy consists of evaluating the right acquisition targets to eliminate competition and enter adjacent growing markets.
One of the interesting by-products that I think will accelerate the challenge to the major tech players is that the strategy of acquiring your competitors might no longer be viable — or will at least have to look very different. Consider the case of Livepeer. Over-time the aim for Livepeer’s governance as I understand it is that there will be no one company that owns the Livepeer protocol, tending towards a decentralized governance that helps to shepherd the development of the protocol over time. If Livepeer successfully decimates the price of video transcoding and starts taking Amazon’s market share, Amazon has no way to purchase Livepeer because it’s a protocol — there’s no one who owns it. At best, they could try to acquire as many Livepeer tokens as possible, and then participate in the network, or possibly try to compete by creating their own distributed protocol which they also would not control in the same sense as they have controlled other resources historically. The era of control is over, and is being replaced by an era of participation. If M&A is no longer a viable strategy, this will be another fundamental shift in how the engine of capitalism works.
I think Facebook’s launch of Libra is one example of a large tech company cluing in to this shift, and trying to find ways to get ahead of the curve by proposing slightly-more controllable, and slightly-more centralized protocols to slow down the transition to a control-less business paradigm in which they are not well suited to operate. The companies that were born as web-first companies had an inherent advantage over the generation of companies before them. Similarly, the protocols and companies born as blockchain-first organizations will have inherent properties that make them likely to quickly outpace their pre-blockchain counterparts in the years to come. I think initiatives like Libra may speed things up in some ways, and slow things down in other ways, but they won’t reverse or co-opt the overall trend to these more fully decentralized, participation based modes of organization.
How these new types of organizations will operate is non-obvious, and there are many open questions. In the early days of the web the same types of questions existed, it wasn’t clearly understood how exactly the new web-first businesses would function in practice, or how exactly they would monetize. That history is probably instructive to what we’ll see with blockchain companies as well — we’re still very much in the figuring it out stage.
It’ll be interesting to watch to what extent the existing companies of today will be able to adapt. You have books like The Innovators Dilemma that explains a lot about how disruption of incumbents occurs, but even when you know that you’re getting disrupted, it’s hard to actually take the types of actions needed to sufficiently change course, rather than just trying to bide your time. Along with all this, humans are bad at understanding things that happen on exponential curves. Today society is organized in a way that causes technological progress to accelerate exponentially which makes it hard to fully wrap our minds around how things might quickly change and mesh together. Blockchains offer an interesting new tool for societal coordination, which we’ll hopefully be able to mindfully employ in beneficial ways in the years to come.