Food Startup Ecosystems

Food Startup Ecosystems
June / September 2014

Food Startups Ecosystem from Raffaele Mauro

Ethereum: the decentralised platform that might displace today’s institutions

Originally published on Internet Policy Review - Written with Primavera De Filippi

Cryptocurrencies have now become commonplace in the online world. Although most of the media attention is focused on Bitcoin, Litecoin, Dogecoin, or other decentralised payment systems, the true revolution is happening at a much deeper level, one that does not involve only money. Bitcoin’s underlying technology - the “block chain” - has been adopted by many other applications with projects such as Maidsafe for distributed file storage, Twister or Bit-messaging for decentralised online communications, etc.

On 22 July 2014, a new cryptocurrency has become available on the market. After many months of preparation, Ethereum finally launched the pre-sale of its very own cryptocurrency - Ether - raising over 25.000 Bitcoin (approximatively $15.000.000) in less than two weeks. But what distinguishes Ethereum from other (more traditional) cryptocurrencies is that it provides a platform for the deployment of decentralised applications which have the potential to disrupt some of the most powerful organisations in advanced societies: those who instantiate financial and governmental institutions.

THE ROLE OF INSTITUTIONS

An institution refers to any social structure in charge of governing the behaviour of individuals within a given community - such as law, money, religion, education, etc (Durkheim, 1985). This is generally achieved by means of formalised mechanisms of social order known as organisations - such as the government, the bank, the church, and so forth.

Institutions are needed to coordinate actions and stabilise expectations amongst a disparate set of individuals - two objectives that could, historically, only be achieved through hierarchical organisations and centralised forms of control.

The former (coordination) is achieved when there are efficient and effective interactions amongst a non-coordinated group of individuals - just like banks coordinate the savings and investments of multiple individuals and organisations.

The latter (trust) requires an organisation to be both accountable and sustainable over time. For instance, we expect banks to be responsible, and to operate with relatively predictable patterns over a long period of time.

The problem with institutions is not their function, but rather the centralised structure of the organisations that subtend them. Centralisation is costly because it relies on the aggregation of information for decision-making, which reduces the ability of such organisations to react promptly to their changing environment. Moreover, centralisation encourages the accumulation of resources and power in the hands of a few individuals, at the expense of less privileged groups.

Can the core functions of institutions - coordination and trust - be achieved by means of decentralised applications, thus avoiding the costs of centralised control?

BITCOIN - DISRUPTING THE FINANCIAL INSTITUTION

Let us take a look at how the financial system is affected by modern decentralised technologies. There has been a long history of pre-digital complementary currencies, such as the Bavarian “wära” and other Gesellian currencies, and multiple experiments were made during the 1990s and the early 2000s with new digital currencies, such as E-gold and the Liberty Dollar. Most of them failed because of scams, instability and scalability problems, some were even involved with massive money laundering regimes and eventually were shut down and/or seized.

Learning from previous failures, Bitcoin, as a decentralised cryptocurrency, represents a true discontinuity from the rest.

Created by the fictional character Satoshi Nakamoto, the Bitcoin network relies on basic cryptographic tools (such as public/private key encryption and digital signatures) to produce and maintain a decentralised public ledger (or “blockchain”) recording all transactions that have been made (and will be made) on the network. The validity and legitimacy of these transactions is verified through the process of “mining” - a process that relies on full transparency and peer-to-peer collaboration to overcome the coordination problems that are typical of decentralised networks (Nakamoto, 2008).

Contrarily to other virtual currencies, Bitcoin overcame several scams and attacks. In spite of the various incidents of theft due to the so-called transaction malleability (allowing for the unique ID of a Bitcoin transaction to be modified before it is confirmed on the Bitcoin network) and the recent BGP hack (exploiting the Internet Border Gateway Protocol to redirect mining traffic to a malicious server), Bitcoin is still strong and alive. Indeed, most of these attacks are due not to a flaw in the Bitcoin protocol, but rather to a lack of understanding and poor security measures taken by Bitcoin users or exchanges.

Facing dramatic price swings and hostile regulatory environments (De Filippi, 2014a), the Bitcoin network coordinates today over tens of thousands of transactions per day, in a relatively efficient manner. The network reflects within its own system the qualities of coordination and trust: two features which are key to the success of many financial organisations and monetary systems.

Yet, Bitcoin constitutes a major change in comparison with previous payment systems to the extent that it enables true independence from centralised forms of control. By creating a trustless system (where strangers can interact without having to trust each other), Bitcoin shifted the focus of trust away from the financial institution, towards the technology underlying the network. In this way, Bitcoin has proven that it is possible to implement a working decentralised currency system that remains independent from governments and corporations.

But Bitcoin’s real innovation is not the currency itself. The Bitcoin blockchain can extend beyond the monetary realm, to support all forms of social, legal and political transactions (De Filippi, 2014b).

ETHEREUM - DISRUPTING SOCIAL AND POLITICAL INSTITUTIONS

Thus far the most powerful example of blockchain-based application is Ethereum, an innovative platform that implements a turing complete scripting language (i.e., one that can solve any possible computational problem) on top of a decentralised cryptocurrency.

Ethereum builds upon the technology of Bitcoin to manage and coordinate different kinds of transactions in a trustless and distributed fashion (Buterin, 2014). While Bitcoin is limited to financial transactions, Ethereum can cover different types of transactions, provided that these can be ‘encoded’ into the blockchain. Financial instruments - such as insurance contracts or derivatives - can be translated into code so as to be understood and automatically enforced by the platform. Physical assets - such as smart phones or smart cars - can be linked to one or more cryptographic tokens that will determine both who owns them and who is entitled to use them. More generally, and perhaps most importantly, Ethereum can be used to regulate social interactions between individuals - such as an employer and its employees, or a licensor and its licensees - through a series of electronic agreements (i.e., smart-contracts) whose provisions can be automatically enforced by the underlying code of the platform; the mechanism by which the contract is defined (i.e., the code) is the same mechanism through which the contract is enforced.

As a result, Ethereum eliminates the need for trust between parties as well as the need for a centralised entity coordinating these parties. People can thus coordinate themselves, in a trustless (since the trust has been shifted onto the technology) and decentralised manner, without having to rely on the services of any third party institution - be it a corporate body or public institution.

Hence, what Bitcoin did to the financial system, Ethereum could do to the political system as a whole. In other words, if Bitcoin was designed as a decentralised alternative to counteract the corruption and inefficiency of the monetary system, Ethereum constitutes a decentralised alternative to the notion of the organisation per se.

Through its decentralised application platform, Ethereum eliminates the need for people to rely on centralised authorities and traditional, top-down governance models, to experiment instead with novel forms of distributed governance where decision-making occurs at the edges of the network. In this sense, Ethereum could contribute to supplanting centralised and hierarchical organisations with more decentralised (autonomous) organisations relying on contract-based coordination.

Today, Ethereum already appears to be a promising technology, at least considering the hype that has built up around it. Following the first two weeks of pre-sale, over 54.000.000 Ether have already been sold (worth almost $16.000.000 as of 1 August 2014). But the most interesting part has yet to come, as Ethereum’s official release will only happen during the last quarter of 2014. We are just witnessing today the emergence of new opportunities for individual emancipation and self-coordination.

Ethereum facilitates a new form of distributed private ordering between a decentralised network of peers, which significantly differs from the traditional regulatory mechanisms employed by centralised organisations and public authorities. In the future, we might be able to build decentralised organisations with distributed models of governance, independent legal systems, or perhaps even autonomously governed communities that would compete with both governments and corporations.

References

Buterin, V. (2014). White Paper: A Next-Generation Smart Contract and Decentralized Application Platform. (Self-published). Available online at https://github.com/ethereum/wiki/wiki/%5BEnglish%5D-White-Paper

De Filippi, P. (2014). Bitcoin: a regulatory nightmare to a libertarian dream. Internet Policy Review, 3(2). DOI: 10.14763/2014.2.286

De Filippi, P. (2014b). Tomorrow’s Apps Will Come From Brilliant (And Risky) Bitcoin Code. Wired, athttp://www.wired.com/2014/03/decentralized-applications-built-bitcoin-great-except-whos-responsible-outcomes/

Durkheim, E. (1895). The Rules of Sociological Method. Durkheim: The Rules of Sociological Method and Selected Texts on Sociology and Its Method, 31-163. (1982).

Nakamoto, S. (2008). Bitcoin: A Peer-to-Peer Electronic Cash System. (Self-published). Available online athttp://nakamotoinstitute.org/static/docs/bitcoin.pdf

The Y-Combinator model: High-potential accelerators and innovation ecosystems

Raffaele Mauro ^*

December 2013 - Beta version

Abstract

Business accelerators are playing an increasing role in the creation and growth of innovative companies. Y Combinator, since its creation in 2005, emerged as the most successful accelerator and launched billion-dollar valuation companies such as Dropbox or Airbnb.  The business history of Y Combinator highlights the role of the surrounding ecosystem and the marked need that was addressed by this typology of investor. Different accelerator models are tailored to the maturity of the surrounding environment and the network effects between local venture capital funds, angel investors, startups, public and private R&D programs, universities, incubators and corporate development policies. The paper explores Y Combinator’s history, specific strategies and, from the evolutionary point of view, what components of its model can be replicated in other innovation ecosystems.

JEL Classification: L26, M13, G24, O31

Keywords: business accelerators, venture capital, innovation ecosystems.

Contents

1 - Introduction

2 - Business case

3 - Y Combinator and the Silicon Valley ecosystem

3.1 History and Structure

3.2 The role of the ecosystem and the success factors

4 – Entrepreneurship Policy, Incubators and Ecosystems

5- Conclusion

References

Appendix

-Databases

- Tables and Figures

* Empedocle Maffia Fellow and MPA candidate at the Harvard Kennedy School, 79 JFK St., Cambridge MA 02139. Email: raffaele_mauro@hks14.harvard.edu. I thank Professor Calestous Juma for advice and support. 

1- Introduction

The aim of this work is to show Y Combinator's business case and its relation with the surrounding ecosystem. This is intended as policy paper, with the objective of informing public decision makers about the options to improve entrepreneurship and innovation in their areas of operation. The methodological lens are the ones of business history, structuring  facts and narratives to make them understandable with the contemporary theoretical debate. Y Combinator was chosen because it is the most successful among the existing accelerators, the model is therefore analyzed in its structure and evolution. Then, the findings are applied to a broader context of other acceleration programs and policy experiments.

After the burst of the dot-com bubble, and especially after the following years of recovery, there was a new wave of accelerator programs in both advanced economies and emerging markets. Business accelerators took advantage of the reduced costs to create technology companies and approached the “lean startup” model, with a focus on fast learning cycles and the quick delivery of minimum viable products[1]. The most effective programs addressed the specific difficulties of their economic environments and innovation ecosystems.

According to the literature (Miller and Bound, 2011), accelerator programs are now a fundamental part of developed innovation ecosystems. This is occurring because the declining costs for building a startup, especially in the internet industry, pressures investors to grasp market cycles that are faster in comparison with the past.  Most of the incubators that were developed after 2005 have some key element in common with the Y Combinator model (Gilani and Dettori, 2011), with average deals of $ 35k to $ 50k, 6-10% equity stakes, 3 to 6 months programs, demo days and hands-on support. It is now estimated the existence of 200 seed accelerators program worldwide (Gilani and Dettori, 2011), including some of the traditional incubation programs, but only 38 true and relevant acceleration programs.

Y Combinator qualifies as a "Model 2" incubator (Grimaldi and Grandi, 2005), with a strong emphasis on mentorship and networking: this is an approach that it is different in comparison with traditional incubations benefits related to real estate and business services. “Model 1” incubators are aimed to strengthen the perspectives for businesses focused on the local economy and with moderate growth potential. This typology of incubators provide many forms of cost reduction and the solution of logistical problems. On the contrary, “Model 2” incubators are focused on high growth companies and are aimed at attracting significant venture investments in a fast way. The latter typology, as Y Combinator does, prepares the companies to deal with significant capital injections and to solve scalability issues. More generally, there was a debate related to the significance of the distinction between “incubators” and “accelerators”, and the findings tend to go in the direction of a real differentiation (Van Huijgevoort, 2012).

The literature related to incubation programs is vast but the number of studies specifically related to accelerator programs is not yet at this level. After the classical papers (Cooper, 1985; Merrifield, 1987; Mian, 1996) and the work made in the wake of new economy during the late ‘90s (Sherman 1999; Hansen et al, 2000) a new wave of academic interest rose from 2009-10 onward.  On the other hand, the fact that the majority of the programs are focused on high potential startups allowed major news coverage; some of the information is now aggregated in databases such as Crunch Base, Seed-DB and CB Insights.

2 - Business case

Y Combinator is one of the most successful acceleration programs in the world: it has the most robust metrics in terms of funding, exits and company valuations (Table 4 and Table 6 in the Appendix show respectively the YC’s aggregate performance and the compared performance with other accelerators). In 2012  Forbes placed it first in its ranking of business acceleration and incubation programs[2].  The initiative was created in 2005 by Paul Graham, Jessica Livingston, Robert Morris and Trevor Blackwell. Since then Y Combinator invested in 566 companies in the very early stage of their lifecycle (Table 10 in the Appendix shows the complete list of YC companies from 2005 onward). The accelerator invested since 2005 over 9 million dollars in 17 semi-annual rounds; the target companies are mainly internet and mobile companies, with over 40 sub-industries covered, with occasional investments in other areas such as robotics, health and life sciences[3]. The aggregate portfolio reached, according to the information disclosed by the accelerator, a total valuation of $13.7 billion at the end of October 2013. The estimates are calculated for the 306 companies with an assigned valuation, with a $2 billion increase since June 2013. The growth in aggregate valuation is mainly due to late stage fundraising of the most mature companies, with a limited impact from the eight acquisitions that happened in same time range. The $13.7 billion figure is almost doubled in comparison with the 2012 total valuation of $7.78 billion.

The average valuation of Y Combinator startups is close to $45 million per company, but the information is distorted by the “black swans” such as Airbnb. It is estimated that 37 of over 500 start ups have a valuation of over $40 million.  According to the online accelerator database Seed-DB, Y Combinator companies have collected over $1.65 billion in venture capital funding and realized 73 exits, with a total cash value of $1.2 billion. In fact, according to the own words of YC founder Pual Graham, their intention is to scale a “black swan farming”, identifying teams with the potential for hyper-growth when they have no metrics and are in the very early stage.

As shown in Table 4 and Table 6 in the Appendix, this performance is remarkable and uncommon among incubation programs, even considering the fact that the number is distorted by the top two cases, Dropbox and Airbnb. Dropbox raised $250 million in venture capital funding at a $4 billion valuation and Airbnb raised $113 million at a $1.3 billion valuation (Table 8 in the Appendix lists the top 25 Y Combinator companies in term of subsequent funding). In November 2013 the former was close to a new valuation of $8 billion for his late stage round and also Airbnb is now valued in the $2.5 billion range, with potential for future increase over $10 billion in case of IPO. In comparison with other business accelerators and incubators, Y Combinator demonstrated a unique capacity to generate “black swans”, companies that generate extremely high returns in comparison with the industry average, and what in the jargon are called “unicorns”, companies with a valuation higher than $1 billion. Other companies are less popular but successful, such as Stripe, Scribid, Posterious and Reddit, and some of them had interesting exits such as Heroku, acquired by Saleforce for $212 million (Table 9 in the Appendix lists the top 25 Y Combinator companies in term of exit value).

3 - Y Combinator and the Silicon Valley ecosystem

3.1 History and Structure

Y Combinator had different phases in its development and was itself a startup. As shown in Table 1, YC evolved over time and aligned itself with the changes of the surrounding ecosystem.

Table 1: Y Combinator’s growth phases and pivots, 2005-13.

Phase 1: Startup	2005	Creation of the program and first batch in Cambridge, MA.
Phase 2: Proof of concept, first growth curve	2005-2008	Expansion to Silicon Valley and growth, first successful exits.
Phase 3: First pivoting	2009-2010	Partnership with Sequoia Capital and restriction of the program only at the Silicon Valley.
Phase 4: Second growth curve	2011-12	The Start Fund, led by Ron Conway and Yuri Milner, allowed in 2011 all of the admitted companies to access $150,000 in convertible note at favorable conditions.
Phase 5: Second pivoting	2012-onward	Increased selectivity and reduction of the additional funding to $80,000.

Source: Personal elaboration based on Y Combinator historical data, bibliography and YC partners official communications.

The actual form of Y Combinator is deeply rooted in the resources and constraints of the Silicon Valley. That was not always case: the program started in Cambridge, one of the initial labels was "Cambridge Seed", but then moved permanently to Mountain View, California, with a greater focus on Silicon Valley. 

Y Combinator's ability to enable startups to grow very fast is a mixture of several elements: personal reputation of the founders, serendipity opportunities during the program, deep relations with top venture capital firms such as Sequoia Capital, creation of an "hacker friendly" culture[4], proximity with all the resources of the Silicon Valley such as corporations ready to make acquisitions, VC funds, specialized law firms and early adopters. These elements increased in effectiveness over time and the accelerator accumulated credibility with both entrepreneurs and investors. When there was a manifestation of scalability problems, Y Combinator partners were effective in resetting the direction and maintained the sustainability of the organization.

During Phase 1, in 2005, the first batch consisted in only 8 startups, while in summer 2012 only the summer batch consisted in 80 startups. Up until 2012 Y Combinator was funding 100+ startups per year, with a track record of 395 teams accepted in the program since its inception. To address the growth in scale the original partners, Paul Graham, Trevor Blackwell, Robert Morris and Jessica Livingston, accepted new collaborators and extended the partnership to Geoff Ralston, Harj Taggar and Paul Buchheit. When the accelerator reached a robust maturity stage, in the 2011-12 phase, the biggest issue was the scalability of the program. Then, in Phase 5, it was recognized an over-growth phenomenon, with decreased effectiveness in supporting deals; therefore in 2012 the additional funding level was reduced from $150,000 to $80,000. Moreover, the partners pushed for and increased selectivity rate: 47 companies were accepted in the Summer 2013 batch in comparison with the 75 of the Winter 2013 batch.  The selectivity rate has now settled t at 2% level.

In October 2013 Y Combinator opened a second office in San Francisco and added new part-time partners to its structure, reaching a total of 10 full time partners and 8 part-time partners. The partnership structure differentiates Y Combinator from other institutions and generate a network of “employed mentors”. There was also a change in the composition of the admitted companies: in the summer of 2013 Y Combinator started backing no-profits, with an initial investment in Watsi, a crowdfunding platform for medical are, and opening a specific application for no-profit projects in Winter 2013.  In the same period, the accelerator replaced the Start Fund with a new vehicle, the YC VC, backed by investors such as Yuri Milner and Andreesen Horowitz.

From the point of view of the acceleration package, Y Combinator offers to the founders strong networking and mentoring resources, optimizing the program with standard contractual terms, acceleration lifecycle and "graduation" with the Demo Day. The initiative is structured as a three months program, with a seed investment of $5,000 plus an average additional $5,000 per founder, generally with an equity request from 2% to 10%. The total investment is typically from $14,000 to $20,000 in exchange for 6% of equity.

The admission process at Y Combinator it is structured in two phases. The first one is an application form with traditional questions about the startup project, such as the description of the underlying customer pain and the existing solutions, and some non-traditional ones focused on understanding team member’s personalities and cultural fit. Emphasis is given to the "hacking" mentality of the founders and their strength as a group. The second phase is an interview: only 4% of the applicants reach this stage and then only half of the respondents are admitted to the accelerator program. The interviews are short in time, only 15 minutes, and are organized as a test of "fluid intelligence", with the aim to understand if the entrepreneurial team will be resilient when facing the frequent pivots that are likely in the first phases of a startup life.

After the admission, the Y Combinator experience is organized around weekly meetings with entrepreneurs, office hours with the partners and a constant pressure to deliver the best presentations and products, with strong user growth metrics, at the Demo Day[5]. The latter occurrence is now one of the most important events in the Silicon Valley and all major investors and observers of the Internet startup scene attend it.

3.2 The role of the ecosystem and the success factors

Y Combinator is part of the Silicon Valley's ecosystem, a "rain forest" that provides fertile ground for company formation and scaling[6]. The concept of ecosystem is fundamental to understand the positive externalities that are available to investment firms and incubators. Several operations of the venture investment cycle are easier within an highly effective ecosystem: from due diligence to fundraising, from knowledge gathering to exiting investments. For instance, research shows (Hochberg et al., 2007) that there is a significant network effect among investors and that better networked firms perform syndicate investments with a superior performance.

The role of an accelerator within an ecosystem is twofold: first, from the investor side, it provides a pipeline of high quality projects to venture capital funds, reducing their due diligence costs and creating new opportunities for investment. Second, from the entrepreneur side, it is a way to reduce the risk of starting a business and turning an idea into a company. Moreover, the "class experience" of the accelerator reduces both the transaction costs, with the pre-structured deals, and reduces the risk in the case of the startup failure, because the teams have not yet fully committed their careers to the new venture. In the specific case of Y Combinator some of the biggest venture capital funds, such as Sequoia Capital, integrated their dealflow process with the accelerator and invested on a regular basis in the startups presented in each batch. Moreover, Sequoia invested $2 million in 2009 in Y Combinator itself, allowing the accelerator to increase significantly the number of funded companies for each year. In 2010 the accelerator raised a second $ 8.25 million from other partners with Sequoia Capital as lead investor.

There are several reasons that are related to Y.C. outcomes and successes, some of them are schematized in Table 2.

Table 2: Y Combinator’s success factors

Reputation of the partners	Most of the founders and partners of Y Combinator have remarkable backgrounds and are on a recognizable position in their domain of activity. Graham, Blackwell and Morris founded in the past Viaweb, a company that was acquired by Yahoo in 1998. Graham was a reputed scholar in the programming language Lisp, Morris crated the first Internet "worm" and was the founder of the low cost robotics firm Anybot.
Selection focused on high-scalability startups	Most of the project accepted on Y Combinator have in common a strong scalability potential, with the aim to reach big markets and high growth rates. This qualifies Y Combinator as an extreme version of "Model 2" incubators, according to the classification of Grimaldi and Grandi[7].
Proximity to Silicon Valley	Y Combinator cannotbe conceived without the Silicon Valley and the close relationship with universities, angel investors, venture capitalists and specialized professional service firms. All the enablers are in close proximity, for instance many Y.C. startups were acquired by Google, which has the headquarter only 3.5 miles away.
Power of the alumni network	Former Y Combinator attendees created a powerful network that acts as a mentorship hub and, in some cases, supports new companies with angel investments. The entrepreneurship cycle is therefore accelerated with both knowledge transfer and a reduction of both cost and time constraints to raise seed capital.
Engineered serendipity	One of the key elements of the innovation processes is the random exchange and recombination of perspectives that occurs in special places. The Y Combinator program is set to maximize the productivity of the three months, creating an environment where the team are encouraged to exchange ideas and change radically their projects during the work in progress.
Hacker friendly environment	The focus of Y Combinator is placed in building products and there is a culture that is appealing to engineers, where part of the company creation has external help and it is heavily coached. The average age of the founders is lower in comparison with other programs and the interaction style is direct and informal.

One relevant element from the ecosystem is that big players such as Google and Apple are acquiring an increasing number of companies both for technology and talent acquisition. They have now an open innovation strategy and are aware that most of the technical talent and potential for disruption is located outside the walls of their organizations. Some of them are pursuing aggressively these opportunities: for instance Google acquired 131 companies from February 2001 to November 2013. The process significantly increased during the last years and 73 out of 131 companies were acquired from 2010 onward. 

Y Combinator responded to a clear marked need from both venture capital firms and IT companies involved in corporate development trough acquisitions.  As shown in Table 5 in the Appendix, between the companies funded by Y Combinator, 9 startups were acquired by Google, 7 by Facebook, 4 by Twitter, 2 by Motorola, 2 by Apple and several other by the other big players of the information technology industry such as Amazon, Likedin, Zynga, Groupon and Yahoo. In fact, often the high growth layer of Y Combinator-backed companies served as acquirer for the younger ones in the deal pipeline, with companies such as Dropbox and Reddit beginning to acquire new startups such as Snapjoy and TapEngage, in the case of the former, or Infogami, in the case of the latter. Sometimes big companies outside the tech industry acquired Y Combinator companies in order to be better positioned in the digital industry, one example is the Reddit acquisition from the media group Condé Nast.  The large number of acquisitions and the limited role for IPO’s is a long standing inheritance of the dotcom bubble, according to some observers Silicon Valley has not yet completely recovered and the investor community remains cautious in front of new public offerings (Beltramini, 2013).

4 - Entrepreneurship Policy, Incubators and Ecosystems

There is a growing interest in entrepreneurship policy and the creation of accelerators as a tool to foster innovation and economic growth. Technological innovation and entrepreneurship are now largely considered as crucial elements for the competitiveness and prosperity of nations. Since the Schumpeter’s early reflections on economic cycles, high potential ventures and the underlying process of marked disruption were placed at the center of the economic growth in modern economies.

Venture investments impact both the macro scale, in terms of job creation and economic growth, and the micro-economic level, increasing the profitability of backed companies. Moreover, a mature venture capital industry has significant spill over in terms of increasing the skill premium and helping to create incentives for technological innovation.

In the US, from 1970 to 2008, each dollar invested in venture capital created 6.36$ of revenue, venture backed companies generated 21% of US GDP and contributed to 11% of the total employment in the private sector (NVCA – HIS Global Insight, 2009). The Kauffman Foundation studies show that the top 1% high growth companies created 40% of the new jobs in the American economy (Stangler, 2010).  Before the global financial crisis, 6-7 billion Euros were invested in VC each year in Europe. It is estimated that in Europe more than one million people were employed in venture backed companies and the compounded growth of employment, from 1997 to 2004, in European VC-backed companies reached  2.4%, compared to 0.7% in other companies  (EVCA, 2005 and 2002).  Even if the effect of VC funds on company innovation levels on the post-investment stage is questioned (Caselli, Gatti and Perrini, 2009), the impact on performance is robust and measured in several markets.

These scenarios attracted the interest of policy makers and created the case for governmental intervention. Past cases of positive public programs that created successful innovation ecosystems were, in different time phases and with different industry specialization, the Silicon Valley, Singapore, Tel Aviv-Wadi cluster in Israel, Bangalore and the Zhongguancun cluster. Nevertheless, most of the attempts to replicate the success of the Silicon Valley had limited performance. There are several cases with mixed evidence there are “horror stories”, such as Kansas Investment Fund and the KPERS (Kansas Public Employees Retirement Fund) activities, where public or citizen’s money was wasted and there was a clear misallocation of capital.

From the policy side, in order to replicate the success of accelerator programs such as Y Combinator, the major challenge is the absence or immaturity of some components of the surrounding ecosystem. As pointed in the literature[8], the vast majority of public efforts to promote venture capital and entrepreneurship had poor results: the most common error is the combination of a lack of acknowledgement of the specific characteristics of the target regions and an excessive distortion of market mechanisms, for instance with geographic restrictions in fundraising an hiring. Another significant element is the need to develop the whole financial systems, empowering all the financial institutions needed in the company lifecycle[9]. One of the most successful models of government intervention, designed to develop the capitalization of venture capital funds without distorting market incentives, is the creation of “hybrid funds”, government backed funds of funds (FoFs) that invest in venture capital funds[10].

An additional challenge is the fact that the availability of technical talent, capital, tacit knowledge and potential buyers in Silicon Valley is not common in other parts of the world, even within the United States. Therefore policy challenges will be different for each innovation ecosystem and the Y Combinator model cannot be copied in its original form. On the other hand, from the point of view of human capital, the "war for talent" in the second part of the 2000's decade expanded significantly for software engineering roles. Even with a weak macroeconomic context, due to the financial crisis and the following slow recovery, the compensation for these positions is driven up by the very high demand from technology companies and the expanded number of Internet startups.  Therefore, high quality talent is often wiped out of the market and some of the best people, in terms of a combination of up to date technical talent and pro-active mindset, are “acqui-hired” trough the acquisition of startups. This is also an opportunity for other areas of the world where technical talent can be involved at low cost.

There are several cases of acceleration programs that are trying to model part of Y Combinator success and some example are discussed here. The cases presented in Table 3 are not reflective of the whole accelerator/incubator landscape in general, but they have in common the focus on startups with high-scalability potential and are considered among the most successful ones.

Table 3: Examples of successful accelerators

Techstars	Acceleration program with multiple locations: Boulder, New York, San Antonio, Seattle, Boston. The essential elements are the mentorship support and a $ 100,000 convertible note given to each startup from a syndicate of VCs. The initiative is very selective, with 1% acceptance rate, and in comparison with other incubators publish publicly its metrics: among the 137 companies admitted to the program from 2007 to 2012, 90% received funding, 16 failed, 18 were acquired and 98 are alive.
Seedcamp	It is considered the major European accelerator program, based in London but with a strong international focus. It has a structure more suited to less developed ecosystems, it has a rolling application process and provides €50,000, office space and mentoring in exchange for 8-10% equity stakes. From 2007 to 2012 73 companies were accepted and 80% received funding after the program,
Startup Chile	Six-month program created by the Chilean government that provides funding to approximately 100 startups each year. It is aimed to attract foreign entrepreneurs; the teams that pass the selection are awarded with $40,000 of funding, a networking program and a one-year visa. The program is based on a strong connection with the Silicon Valley ecosystem and has both the objective to promote innovation internally and build links with the world markets.

Sources: Techstars, Seedcamp and Startup Chile datasets.

The top business accelerators add value to their companies in a measurable way (Hoffman, Radojevich-Kelley, 2012): accelerator-backed companies tend to acquire subsequent funding rounds much earlier than other companies and, on the same time, have an higher risk/return profile with an higher probability of a successful exit. On the other hand, the same companies have concurrently an higher probability of failure (Smith t al, 2013), putting the accelerated startups in a “high success or failure” track. One of the most common criticism is also that accelerator programs had often poor performance in their aggregate portfolios and only few, such as Y Combinator, fulfilled the promise of building companies with valuations greater than $1 billion. None of them was able to generate, up until now and including Y Combinator, “super-unicorns"[11], technology companies with valuations greater than $100 billion such as Google, Facebook, Apple and Microsoft.  

One of the reasons for the success of the Y Combinator model is shown in historical structure based subsequent pivots and adaptations: the constant evolution of the incubator helped its performance and was aligned with the rapid changing structure of the Silicon Valley ecosystem. For instance, in the most recent pivots in 2012 and 2013 involved the adaptation of the program to the new Silicon Valley landscape: the new context is characterized by a limited number of series A rounds in comparison with the number of companies funded per year, especially in the consumer internet area, with on the same time an increase of the average valuations that doubled from 2010 to 2013[12].

As shown in Table 7 and Figure 1 in the Appendix, from 2003 to 2013, in the USA 39 companies reached a $1 billion dollar valuation. The figure could be larger, given the fact that recently founded companies will often need a five year window in order to develop fully their potential[13]. Y Combinator produced 2 out of these 39 companies: Dorpbox, from the summer 2007 batch, and Airbnb, from the Winter 2009 batch.  Other companies, such as Stripe, from the Summer 2010 batch, are likely to join the club of the companies valued over $1 billion (Table 7 and Figure 1 in the Appendix list the US software companies founded between 2003 and 2013 with valuation higher than $1 billion).

The performance of the Y Combinator model does not imply that it must be copied completely. Incubators in aggregate demonstrated the skewed return distribution and time consistency that is typical of VC funds (Kaplan, Schoar 2005), leaving space for new players mainly if specialized in market niches or if associated with highly differentiated investment strategies. Moreover, new experiments such as funding platform like AngelList and the crowdfunding regulation related to the JOBS Act in 2012-13 are changing the landscape and will potentially disrupt low-tier investors.

On one hand, most of the modern accelerators were born after 2005 and it is still to early to drive conclusions, on the other hand there is enough empirical evidence to say that most of these structures will not generate big hits. The first element is that several highly successful companies, such as Square, Quora, Palantir, Zynga and Evernote, were created by serial  entrepreneurs. These actors have a higher probability to build successful companies (Gompers et al., 2006) and, by definition, they do not need the help of an accelerator. Moreover, the network effects and the location effects are extremely powerful factors in the growth of technology companies.

In the accelerator space there are the same Paretian distributions and power law typical of the venture capital industry: the top 1% incubators, such as TechStars and Y Combinator, are attracting most of the high-quality dealflow, leveraging their capacity to attract the best entrepreneurs (Isabelle D. 2013) and letting few space for the new incubators.  The latter tend do be successful when they specialize in industry verticals, such as medical technology in the case of RockHealth or hardware in the case of Lemnos Labs, of for specific underrepresented demographics, such as women or minorities such as NewME. A new wave of accelerators related to social business and impact investing is also confirming some of the key characteristics of the YC model, such as the key role of partner selection and selectivity (Barn et al, 2013).  It is also shown from most of the datasets that top accelerators are generating most of the new job positions, and also for this metric Y Combinator is placed at the first position among similar programs[14].

Public sponsored programs often demonstrated difficulties in their implementation and promoted distorted incentives: one of the most common errors is to replicate practices that worked in other areas without taking account the local market needs.  One of the best cases of public funded acceleration programs is Startup Chile. One of the key ideas of the program is related to the creation of hubs of talented entrepreneurs with a global mind set, an approach aimed to create the minimum critical mass for the creation of a private innovation ecosystem. Therefore, the Startup Chile attempted to attract 1,000 high-tech startups in four years with an overall cost of $40 million. The impact was significant: a country that was known for traditional industries or for its troublesome political history is now considered one of the hubs of global entrepreneurship. It is too early to wholly assess the performance of the program, we know that from the first year batch 55% of the startups are now profitable or were able to raise funding. If this performance will be consistent in the future, it will be a remarkable result for a government sponsored program, even if the outcomes will be lower in comparison with the top private incubators.

Startup Chile is an interesting example of global talent pooling, a mechanism were a country was able to attract people that otherwise would have been under-utilized in their respective countries or would have migrated to the traditional innovation hubs. Startup Chile attracted human capital and build an investment hub, something very similar to the Y Combinator key elements, but generated a longer and more structured program that was tailored to the specific characteristics of the local business environment.

5 - Conclusion

From the business history perspective, the Y Combinator model succeeded thanks to its capacity to evolve rapidly and change its strategic direction in at least two phases of its development. Moreover, Y Combinator integrated effectively with the surrounding environment because not only leveraged its existing strengths but compensated the scarce elements in its ecosystem. In particular, the market, in this case VC funds and corporate development offices of big IT players, needed an high quality dealflow of small companies with proven metrics and robust growth rates. These  elements are easy to build in the Silicon Valley ecosystem, but in other environments there are additional challenges to be addressed. The policy maker must accept the fact that longer programs, different acceleration packages and training are fundamental elements in less mature ecosystems.

Successful examples around the world, such as Techstars, Seedcamp and Startup Chile have elements in common with the YC model but they are structured with their own DNA.  In particular, the key similarity between Startup Chile and Y Combinator is the extreme focus on attracting the best human capital on a global scale and the design of the program, structured to combine the best elements of the local ecosystem and supplying the missing ones externally. It is too early to see if Startup Chile, as a government program, will show the same evolutionary properties of Y Combinator. The dynamic approach should be a key element of the structure of future incubator programs: the ability to pivot and to maintain complementarity with the surrounding ecosystem are  fundamental aspects to reach success.

It is debated if the role of public policy should be to create “black swans” with disproportionate returns in comparison with average deals. The alternative view stands for the public as the catalyzer to the creation for average, but sustainable, companies. From this point of view, the search for disproportionate returns should be left to private investors. Moreover, the chances for the creation of very high growth companies are higher in the ecosystems where the hubs of technical talent and capital are already formed.

There are also other limitations to the startup enhancing programs: they can accelerate investments bubbles and distort incentives, for instance deviating resources from other industries were a specific country can benefit most. But, on the other hand, the positive externalities for the creation of entrepreneurial hubs are significant and seed accelerators are increasingly considered part of mature innovation systems. Moreover, even if most of the accelerators are unable to systematically build $1 billion companies, they can became the focal point for the evolution of local innovation ecosystems and add significant value.

Successful entrepreneurship policies should acknowledge that the most successful startup and teams are "born global" and that there is the need to compensate the specific limitations of the target region of the program. The experience of private accelerators such as Y Combinator and public seed programs such as Startup Chile tells us that it is possible to create new institutions to promote innovation in a relatively fast way. They are experiments, some of them will fail, but this exploration process will provide guidance for the evolution of innovation ecosystems in other geographical areas.

[1] The popularization of the lean start-up concept has its most significant example with Reis Eric, The Lean Startup: How Today's Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses, Crown Business, 2011.

[2] VV.AA., "The Mida's List - 2012" Top Tech Investors, Forbes, May 2012.

[3] Data are available from Y Combinator, CrunchBase, YC Universe, CB Insights and Seed-DB.

[4] Meaning that YC partners pushed for a culture focused on engineering and experimentation, with the aim to build product and services with strong growth metrics and fast iteration cycles.

[5] There are very good narrative accounts of the YC experience, such as Randall's Inside Y Combinator, Silicon Valley's Most Exclusive School for Startups, and some attempts to systematize the fragmented sources of information, as Jed Christiansen work. The founders of Y Combinator are themselves prolific bloggers and many of the companies that participated at the program released accounts of their experience.

[6] Hwang Victor W., Horowitt Greg, The Rainforest: The Secret to Building the Next Silicon Valley, Regenwald, 2012, pp.31-33.

[7] Grimaldi Rosa, Grandi Alessandro, "Business Incubation and New Venture Creation", Technovation, n°25, 2005, pp. 111-121.

[8] Lerner Josh, Boulevard of Broken Dreams: Why Public Efforts to Boost Entrepreneurship and Venture Capital Have Failed--and What to Do About It, Princeton University Press, 2009.

[9] For a description of the Italian Case, see Gervasoni Anna, Bollazzi Francesco, “Developing venture capital activity to promote entrepreneurial competitiveness: Italy’s case history”, Sinergie Nr. 87, January-April 2012.

[10] For a comprehensive assessment, Murray G., Cowling M, Liu W., Kalinowska-Beszcynska O., Government co-financed ‘Hybrid’ Venture Capital programmes: generalizing developed economy experience and its relevance to emerging nations, Kauffman International Research and Policy Roundtable, Liverpool, 11-12 March 2012.  Look also Feola Rosangela, I fondi misti di venture capital, Convegno Aidea Giovani - “L’Innovazione nella Pubblica Amministrazione: Teoria e Prassi”, Università di Roma Tor Vergata, 15 July 2005. For a discussion of the recent Italian experiment with the Fondo Italiano di Investimento, look at Montanino Andrea, Sattin Fabio, Il nuovo Fondo Italiano di Investimento per le PMI: una riflessione generale, EntER - Centre for research on Entrepreneurship and Entrepreneurs,  Università Bocconi, WP n°1, 2010.

[11] The expression was coined by Lee Ailen, Welcome To The Unicorn Club: Learning From A-Dollar Startups, TechCrunch, 2 November 2013.

[12] PitchBook, Venture Capital Database.

[13] Wilson Fred, The Billion Dollar Valuation Club, AVC blog, 3 November 2013.

[14] Seed-DB collects information on start-up employment, but data related to the workforce levels in new companies are hard to collect and often not reliable.

References

VV.AA., Accelerating Success: A Study of Seed Accelerators and Their Defining Characteristics, Chalmers Institute of Technology, 2012.

VV.AA., "The Mida's List - 2012" Top Tech Investors, Forbes, May 2012.

A.A.V.V., Venture Impact: The Economic Importance of Venture Capital-Backed Companies to the U.S. Economy, National Venture Capital Association - IHS Global Insight, 2009.

VV.AA., Employment contribution of Private Equity and Venture Capital and Europe, EVCA Research Paper, November, 2005.

VV.AA., Survey of the Economic and Social Impact of Venture Capital in Europe, EVCA Research Paper, June 2002.

Baird Ross, Bowles Lily, Lall Saurabh, Bridging the “Pioneer Gap”:
The Role of Accelerators in Launching High-Impact Enterprises, Village Capital, 2013.

Ben-Ari Guy, Vonortas S. Nicholas, "Risk financing for knowledge-based enterprises: mechanisms and policy options", Science and Public Policy, 34(7), August 2007, pp. 475–488.

Barrehag Lisa, Fornell Alexander, Larsson Gustav, Mårdström Viktor, Westergård Victor, Wrackefeldt Samuel, Accelerating Success:
A Study of Seed Accelerators
and Their Defining Characteristics, Chalmers University of Technology.

Beltramini Enrico, “Tech Entrepreneurship in the Post-dotcom Bubble Era. The Case of Early Stage Start-ups in Silicon Valley”, Journal of Economic Philosophy, 2013,  1(2), 24-3.

Bugamelli Matteo, Cannari Luigi, Lotti Francesca, Magri Silvia, Radici e possibili rimedi del gap innovative del sistema produttivo Italiano, Banca d’Italia, Aprile 2011.

Caselli Stefano, Gatti Stefano, Perrini Francesco, “Are Venture Capitalists a Catalyst for Innovation?”, European Financial Management, Vol. 15, No. 1, 2009, 92–111.

Christensen L. Jesper, "The Development of Geographical Specialization of Venture Capital", European Planning Studies Vol. 15, No. 6, July 2007.

Christiansen Jed D., Copying Y Combinator: A Framework For Developing Seed Acceleration Programs, Copenhagen Business School, 2009.

Cooper Arnold C., “The role of incubator organizations in the founding of growth-oriented firms”, Journal of Business Venturing, Volume 1, Issue 1, Winter 1985, Pages 75–86.

Hochberg Yael V. and Kamath Kristen, U.S. Seed Accelerator Rankings, Kellogg School of Management, Northwestern University, 2012.

Hoffman David Lynn, Radojevich-Kelley Nina, “Analysis of Accelerator Companies:
An Exploratory Case Study of Their Programs, Processes, and Early Results”, Small Business Institute Journal, 2012, Vol. 8, No. 2, 54-70

Hwang Victor W., Horowitt Greg, The Rainforest: The Secret to Building the Next Silicon Valley, Regenwald, 2012.

Feola Rosangela, I fondi misti di venture capital, Convegno Aidea Giovani - “L’Innovazione nella Pubblica Amministrazione: Teoria e Prassi”, Università di Roma Tor Vergata, 15 July 2005.

Gervasoni Anna, Bollazzi Francesco, “Developing venture capital activity to promote entrepreneurial competitiveness: Italy’s case history”, Sinergie Nr. 87, January-April 2012.

Gilani Aziz, Dettori Gianluca, Incubators in US and Europe - Speed and Scale in Capital Formation, Kauffman Fellows Program, 2011.

Gompers Paul, Kovner Anna, Lerner Josh, Scharfstein David, “Skill vs. Luck in Entrepreneurship and Venture Capital: Evidence from Serial Entrepreneurs”, Journal of Financial Economics 96, 2010, 18-32.

Graham Paul, How Y Combinator started, March 2012. http://ycombinator.com/start.html

Graham Paul, Y Combinator Numbers, June 2011. http://ycombinator.com/nums.html

Grimaldi Rosa, Grandi Alessandro, "Business Incubaton and New Venture Creation", Technovation, n°25, 2005, 111-121.

Grimaldi Rosa e Sobrero Maurizio (2000), "Le strutture a supporto della creazione di nuove imprese", in Lipparini, A. e Lorenzoni, G. (eds.), Imprenditori e Imprese, Il Mulino, Bologna (Bologna, ISBN 88-15-07422-8).

Hansen, M.T., H.W. Chesbrough, N. Nohria, and D.N. Sull (2000). “Networked Incubators: Hothouses of the New Economy,” Harvard Business Review, 78 (5), 74-84.

Hochberg Yael, Ljungqvist Alexander, Lu Yang, “Whom You Know Matters: Venture Capital Networks and Investment Performance”, Journal of Finance, 62, 2007, 251–302.

Isabelle Diane A., “Key Factors Affecting
a Technology Entrepreneur's Choice of Incubator or Accelerator”, Technology Innovation Management Review, February 2013, 16-22.

Kaplan Steven, Schoar Antoniette, “Private Equity Performance: Returns, Persistence, and Capital Flows”, Journal of Finance, vol. 60(4), pages 1791-1823, 08.

Lee Ailen, Welcome To The Unicorn Club: Learning From Billion-Dollar Startups, TechCrunch, 2 November 2013.

Lerner Josh, Boulevard of Broken Dreams: Why Public Efforts to Boost Entrepreneurship and Venture Capital Have Failed--and What to Do About It, Princeton University Press, 2009.

Merrifield Bruce, “New Business Incubators”, Journal of Business Venturing, Volume 2, Issue 4, Autumn 1987, Pages 277–284.

Mian Sarfranz A., “US university-sponsored technology incubators: an overview of management, policies and performance”, Technovation, Volume 14, Issue 8, October 1994, Pages 515–528.

Miller Paul, The Startup Factories: the Rise of Accelerator Programmes to support new Technology Ventures, NESTA discussion paper, June 2011.

Montanino Andrea, Sattin Fabio, Il nuovo Fondo Italiano di Investimento per le PMI: una riflessione generale, EntER - Centre for research on Entrepreneurship and Entrepreneurs, Università Bocconi, WP n°1, 2010.

Murray G., Cowling M, Liu W., Kalinowska-Beszcynska O., Government co-financed ‘Hybrid’ Venture Capital programmes: generalizing developed economy experience and its relevance to emerging nations, Kauffman International Research and Policy Roundtable, Liverpool, 11-12 March 2012.  

Reis Eric, The Lean Startup: How Today's Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses, Crown Business, 2011.

Sherman, H.,“Assessing the Intervention Effectiveness of Business Incubation Programs on New Business Start-Ups,” Journal of Developmental Entrepreneurship, 4 (2), 117-133.

Smith Sherley, Hannigan T.J., Gasiorowski Laura, Accelerators and Crowdfunding: Complementarity, Competition, or Convergence in the Earliest Stages of Financing New Ventures?, University of Colorado-Kauffman Foundation Crowd-Funding Conference, Boulder, CO, July 12-13, 2013. 

Stangler Dane, High Gowth Firms and the Future of American Economy, Kauffman Foundation Research Series, Firm Formation and Economic Growth, March 2010.

Stross Randall, The Launch Pad: Inside Y Combinator, Silicon Valley's Most Exclusive School for Startups, Portfolio, 2012.

Van Huijgevoort Thomas, The ‘Business Accelerator’: Just a Different Name for a Business Incubator?, Utrecht School of Economics, 2012.

Wilson Fred, The Billion Dollar Valuation Club, AVC blog, 3 November 2013.

Appendix

Databases 

AngelList https://angel.co/

CB Insights http://www.cbinsights.com/

CrunchBase: The Free Tech Company Database http://www.crunchbase.com/

Statwing Blog, A Statistical Portrait of a Y Combinator Batch, August 2012.

http://blog.statwing.com/a-statistical-portrait-of-a-y-combinator-batch/

PitchBook – http://pitchbook.com/Venture_Capital_Database.html

Seed-DB http://www.seed-db.com/

Y Combinator Universe http://ycuniverse.com/ycombinator-companies

Tables and Figures

Table 4: Y Combinator investments, exits and returns organized by batch, 2005-13.

Source: Personal elaboration, data from Y Combinator CrunchBase, TechCrunch, Y Combinator Universe and Seed-DB.com.

Table 5: Major acquirers of Y Combinator companies

Source: Personal elaboration, data from Y Combinator CrunchBase, TechCrunch, Y Combinator Universe and Seed-DB.com.

Table 6: Comparison of the top 30 business accelerators worldwide.

Source: Personal elaboration on Seed-DB data.

  

Table 7: Valuations of software companies founded between 2003 and 2013 with at least $ 1 billion in valuation. YC companies are highlighted.

Source: Personal elaboration from NYSE, NASDAQ, CrunchBase and Cowboy Ventures dataset.

Table 8: Top 25 Y Combinator companies by subsequent funding.

Source: Personal elaboration on Seed-DB data.

Table 9: Top 25 Y Combinator companies by exit value.

Source: Personal elaboration on Seed-DB data.

Table 10: Complete list of Y Combinator’s investments, 2005-13.

Source: Personal elaboration on Seed-DB data.

Figure 1: Valuations of software companies founded between 2003 and 2013 with at least $ 1 billion in valuation.

Source: Personal elaboration from NYSE, NASDAQ, CrunchBase and Cowboy Ventures dataset.

---