On Startups, Platforms, and Innovation
Editor’s Note: This testimony was delivered by a16z managing partner (and former chairman of the board of the National Venture Capital Association) Scott Kupor to the FTC as part of their hearings — with various professors, policymakers, industry representatives, and other panelists — in October 2018 on multi-sided platforms and competition.
Thank you very much for the opportunity to speak with the FTC regarding market power and consolidation in the U.S. technology community and potential policy responses. I applaud the FTC for its efforts to take a closer look at this area to ensure that the U.S. remains a robust participant in the global entrepreneurship-driven economy.
By way of background, I am the Managing Partner for AH Capital Management […] My comments today will of course be informed by my professional role — that is, representing an organization that seeks to identify new startup business opportunities and provide them financial backing in the hopes that they turn into valuable, long-term self-sustaining enterprises.
I understand the FTC to be looking at two related issues.
First is whether the success of companies such as Facebook, Google, Apple, Amazon and potentially others positively impacts economic growth and access to capital for startups. And second, depending on the answer to the first question, whether attempts to regulate these companies by, among other things, scrutinizing their future acquisitions would be a valuable use of the FTC’s regulatory authority.
I would like to make three points today relevant to these topics:
- First, while many of these companies do enjoy very strong market positions, I would argue that their value as potential distribution platforms to the technology community, and in particular to the pace of new startup investments, remains very important and is in fact pro-competitive.
- Second, these large players are significant acquirors of venture-backed startup companies, which plays an important role in maintaining the ongoing health of the venture ecosystem.
- And, third, where startups have concerns about the potential market power of these institutions, the free market works very well in addressing those concerns. I will use the rise of cryptonetworks as an example of such a well-functioning free-market response to incumbency concerns.
Let’s start with the distribution platform issue.
Prior to the existence of these large consumer platforms, it was very difficult for consumer startup companies to successfully and cost-effectively build a customer base. That is, the customer acquisition costs for new startups were high because of the need for each new startup to build its own customer acquisition channel. This is among the reasons why many of the early consumer Internet based startups in the late to mid-1990’s failed — it was simply too expensive for them to acquire customers directly and the economic rents they could ultimately earn from those customers was too small to be able to ultimately recoup these costs. The Internet was simply too nascent and thus the number of customers to whom they could ultimately market their services was severely limited.
Fast forward to the last 7-10 years of consumer-related startup activity and we have had an unprecedented amount of new and successful large company formation. As examples, we have seen the creation of a whole new set of multi-billion companies delivering enormous consumer utility […]
These companies of course benefited from the virtual ubiquity of the Internet (itself supported by the commercialization of smartphones by Apple and Samsung, among other large platform providers), which drove an enormous increase in the available market size. But, they also benefited from the customer acquisition channels available to them from many of the other large platform providers that are today the subject of these discussions — Google, Facebook, Amazon and, of course, Apple.
That is, to acquire customers, these startups did not have to go to the expense of building out massive new sales channels, but rather they leveraged the advertising and SEO platforms of these large incumbents. This means a lower cost of customer acquisition, a pay-as-you-grow model for increasing advertising and customer acquisition costs depending on your stage of development and the most successful platform results, and thus ultimately an attractive model for entrepreneurs and venture capitalists to fund experimentation in the startup world.
It is the existing of these platforms that in many ways explains the significant growth we’ve seen in the last 7-10 years in consumer startup and VC financing activity. Simply put, the math works — companies can experiment with customer acquisition via these channels and fund their marketing campaigns iteratively based on what yield the highest return on capital.
Without these platforms, I would venture that the economics of customer acquisition would be cost prohibitive for most startups and thus that the venture capital community would shift it’s investment into other more cost-effective areas.
To be fair, there are no doubt times where too much platform risk can create problems for companies; that is, where complete dependency on these third-party platforms can impact a business. Zynga, for example, ran into this problem with Facebook in that it became wholly reliant on Facebook as its sole customer acquisition mode and failed to establish a direct relationship with its customers over time.
The successful startups that I mentioned before recognize this and utilize these platforms as mechanisms to jumpstart or bootstrap their initial customer acquisition approaches, understanding that over time they need to build direct relationships with their customers to mitigate the risks of long-term dependence on the platforms. Thus, the platforms play a crucial role in the early-stage development of potential long-term network effect businesses, while giving way to the more mature customer retention and acquisition strategies that the startups can employ as they become more mature.
Let’s turn to the important role of the platform companies in the M&A environment.
The VC business is a high-risk one — we fund companies at their earliest inception (often when they are merely founders with an idea of what they intend to build but no demonstrable product or service), with the idea that most of these companies will not yield significant economic value. In fact, about 50% of what we invest in fails, meaning that the companies have little to no remaining economic value. About 25-30% of what we fund turns into small returns — the companies build something of value but can’t ultimately grow to their potential. And finally, a small minority of what we fund turns into large opportunities that either exist as stand-alone public companies or are sold for significant economic returns to larger players. We know and understand this risk, but nonetheless require a small number of companies to yield very high returns on capital to make the ultimate venture business succeed.
To that end, about 15-20 years ago, the venture business enjoyed what we call “liquidity” events (the ability to convert an investment into a real return and therefore return capital to our underlying investors) in the form of about 50% IPOs and 50% M&A events. Today that math is closer to 80-90% M&A and 10-20% IPOs.
The reasons for that are beyond the scope of this hearing, but this trend plays a very important role in the potential actions that the Commission is considering with respect to some of the large platform players in the industry.
In full disclosure, we have been the beneficiary of acquisitions in our portfolio from some of these players […]
But, it’s important for the Commission to understand the important role that M&A transactions play in the venture ecosystem. We raise money from institutional investors — University endowments, Public foundations, Pension funds, etc. — who are seeking a return on their investment and want to be able to recycle these investment returns into a long-term portfolio of venture investments. To do this, we need to provide them liquidity in the form of cash returns from acquisitions or IPOs.
As the number of IPOs has shrunk considerably — down about 50-60% in the last 15-20 years relative to the long-term median volumes — and as the time to IPO has significantly elongated — now about 10-13 years from founding vs a long-term median of 6.5-7 years, the role that IPOs play in providing liquidity continues to shrink.
As a result, M&A continues to grow in importance to the well-functioning of the ecosystem — it provides much need liquidity to the institutional investors, who then use that liquidity to re-up their venture capital investments. Policy that could impact the timing or availability of M&A could do meaningful damage to the capital flows in the venture business, particularly at a time where the capital markets remain a much delayed and smaller exit opportunity for venture-backed companies.
Finally, I’d like to turn to the role of the free market in addressing any potential competitive concerns that the platform players may provide.
We believe that in many ways the growth of what we call cryptonetworks is in part a response to the developer community’s concerns about too much reliance on large platforms for customer acquisition. We don’t think it’s an accident that, at the same time U.S. policymakers are holding hearings to discuss potential concentration in the technology industry, we see the increasing growth of venture capital investment and new company formation in the cryptonetworks industry.
Why is that?
Recall I mentioned earlier that successful companies leverage the platforms to bootstrap their customer acquisition efforts, but over time develop ways to control their own destinies. They don’t want to be overly reliant on the platforms for the long-term as the platforms may change their business practices — by the way, in perfectly fair and legal ways — that could favor or dis-favor certain other participants in the network. This potential risk is one that many companies seek to mitigate.
We define the term “cryptonetworks” as:
- a new way to build “digital services”;
- where those services are “owned and operated” by a “community of network participants,” rather than by a centralized corporation; and
- where the repository of activity on the network (i.e., the database) is decentralized and maintained by the community.
What are “digital services”? They are simply internet-based applications, such as many of the ones we enjoy today — ridesharing, messaging, grocery delivery, enterprise applications, to name a few. We believe that developers will create a whole new set of digital services utilizing the principles of cryptonetworks, many of which are likely beyond our imagination today but will also yield enormous consumer utility.
And what are those principles of cryptonetworks? That they are both owned and managed by the community that develops, maintains and utilizes the networks. This is distinct from the large platforms today, where the ownership and management of those services are governed by a centralized corporation.
At first blush, this may sound crazy — that there may be value in community ownership and management of an asset that exceeds that of centralized corporate control? But, in fact, there is well-established precedent for this in the history of the technology industry.
First, is the open source software movement. This started in 1983 as a movement to create free software, led by an MIT researcher named Richard Stallman. Understandably, this was a radical concept at the time — that a community of developers would publish their software freely for others to modify and incorporate into various other open projects. But over time, this work morphed into the mainstream development of open source software, which today is the predominant method by which software is developed. Examples of open source software that have experienced widespread adoption include Linux, an operating system that governs most data center servers today and is a major component in virtually all smartphones and tablets, and Git, an open source software development system used by millions of software engineers globally.
Second, is the development of the very internet protocols that have given rise to the tremendous job and economic growth and consumer utility that we all currently enjoy from existing digital services. These protocols — which include, for example, SMTP (the protocol for email transmission), HTTP (the protocol to exchange structured text on the internet) and TCP/IP (the protocol for end-to-end data communication) — derived largely from academic or government-funded efforts and have been maintained in most cases by communities of academics and developers. They are “open” protocols in the sense that they are the well-established foundations on which many very exciting for-profit businesses have been built (e.g., Facebook, Amazon, Google), knowing that the protocols themselves cannot be changed by a centralized corporation.
The value of cryptonetworks thus are manifold.
First, they provide a platform on which new companies can be built, without the attendant concerns that many come from relying wholly on centralized platforms. That is, decentralization and transparency ensures a more transparent governance model and allows the very startups who build on these platforms to be active members of the platform governance.
Second, cryptonetworks provide a symbiotic economic relationship between the value of the network itself and the work that is done by those who govern and manage the network via the introduction of a network-specific token. “Tokens” in the cryptonetworks world perform a series of functions: (i) they are the method of value exchange between network participants — that is, consumers “pay” for services using the token and sellers “receive” tokens in exchange for the services and (ii) they provide the financial incentive to reward developers and other maintainers of the network — that is, people may receive tokens for ensuring the authenticity of the transactions completed on the network.
I bring all of this up not to laud the value of cryptonetworks themselves, but rather to use them as a very real and current articulation of the role that the free markets play in reacting to perceived market challenges. That is, to the extent that developers perceive the platforms to be more powerful than they might otherwise wish, cryptonetworks is a free market solution to that perception.
In other words, it doesn’t require new government regulation or changes to existing regulations to address changes in a very progressive and fast-changing technology market. Rather, the industry has a way of remedying any perceived challenges through the operations of economic self-interest.
Interestingly, there is an exact parallel in the enterprise community that mirrors what we have seen on the consumer platform side of the world.
In the early 2000’s many commentators thought that market power and consolidation among the large enterprise players — e.g., Microsoft, IBM, Cisco, Oracle, etc. — would create anti-competitive behavior and consumer dis-utility. And, in many ways, the concerns were identical to those being expressed today — these large organizations controlled access to customers (akin to the consumer platforms today) in a way that made it difficult for new entrants to offer competitive solutions to those very same customers. The “customers” in those cases were the centralized buyers of enterprise IT solutions (namely the Chief Information Officer), who controlled IT budgets and thus dictated the adoption (or not) of new technologies.
Again, however, the free market solved the perceived concentration problem through the development of new technology. In particular, the development of cloud computing democratized access to technology in the enterprise, enabling point solutions to be adopted by individual departments vs the tops-down adoption of technology from the centralized IT organization. As access to the technology democratized, so too did access to budget dollars to produce technology — thus defeating the centralized budget control of CIO and in turn reducing the power of the incumbent IT providers.
It is this broad shift in the underlying technology that has given rise to a whole new generation of enterprise IT providers, fueled by venture capitalists who recognize the market landscape. Examples of such companies are Salesforce.com, Workday, Splunk, Servicenow, Box, Okta, among others.
It goes without saying that the free market does not always work and the role of government regulation remains an important one to remedy such market functions. But, particularly in a market as fast moving as technology, we should exercise caution in supplanting the ability of the free market to react appropriately by developing new government regulations.
For in many cases, government regulation can have unintended consequences. In particular, regulation can often have the opposite of its intended effect by providing more market power to the incumbents as they are often the only ones who can afford to absorb the costs of the regulations. Startups inevitably are at a disadvantage in such circumstances, as venture capitalists won’t finance new businesses where the incumbents have perceived regulatory advantages in the form of unlimited compliance resources. Barriers to funding new companies that may in fact help produce competition for market incumbents not only has economic losses from the lack of new job formation attendant to startup growth, but will also calcify market control for incumbents.
Such risks are particularly acute today given the global opportunities for startup development and the ability of foreign governments to utilize pro-startup policies to attract new investment. Twenty years ago, the US had 90% market share of venture capital dollars; today that is just barely above 50% as foreign governments continue to seek to attract startups through the development of pro-entrepreneurship policies — e.g., taxes, immigration, regulatory-lite frameworks. Capital is indeed highly fluid globally and thus we do need to be careful that well-intentioned regulatory policies don’t disadvantage US startups to the benefit of foreign-domiciled competitors.
I appreciate very much the opportunity to be a part of today’s discussion and look forward to answering your questions.