Monthly Archives: May 2012

Freeman Dyson: The Scientist as Rebel

Freeman Dyson is a strange scientific blend of wise and moderate conservatism and pioneer of iconoclasm. He advocates cold analysis but loves what is strange. I just read The Scientist as Rebel, a wonderful book where everyone can find his or her share of intellectual stimulation.

Any relationship with innovation or entrepreneurial high-tech? Very little directly, and the subject is closer to my other articles on the books about reflexions on science (Smolin, Ségalat for example). There are actually many connections between scientific research and technology innovation, not the least being the question of creativity. Another tenuous link: he is the father of Esther Dyson, famous venture-capitalist in Silicon Valley.

Failure is another example. In an interview Dyson gave before writing this book said about its role: “You can’t possibly get a good technology going without an enormous number of failures. It’s a universal rule. If you look at bicycles, there were thousands of weird models built and tried before they found the one that really worked. You could never design a bicycle theoretically. Even now, after we’ve been building them for 100 years, it’s very difficult to understand just why a bicycle works – it’s even difficult to formulate it as a mathematical problem. But just by trial and error, we found out how to do it, and the error was essential. The same is true of airplanes.” From Freeman Dyson’s Brain

The Title “The Scientist as Rebel” also reminds me the quote by Pitch Johnson which I had mentioned in Entrepreneurs and Revolutions: “Entrepreneurs are the revolutionaries of our time.” And he had added: “Democracy works best when there is this kind of turbulence in the society, when those not well-off have a chance to climb the economic ladder by using brains, energy and skills to create new markets or serve existing markets better then their old competitors”

In this book, Dyson writes about ethics, religion, climate change and about scientists as different as Gödel, Erdös, Hardy, Oppenheimer, Feynman of course, Teller the indefensible, and Thomas Gold whom I had never heard of.

The chapter is called “A Modern Heretic.” Gold has covered topics as diverse as
– the physiology of the ear (validated 30 years later despite resistance of many kinds),
– the instability of Earth’s axis of rotation,
– the abiotic origin of natural gas and oil, (i. e. not derived from the degradation of living creatures)
– the existence of life within the Earth’s crust,
– the interpretation of pulsars.
He had no fear of making mistakes on such topics as
– the steady state universe,
– the moon’s surface being covered with a fine rock powder.
Gold was “an intruder but certainly not an ignorant” and added that “science is not fun if the scientist is never wrong.” I just found another blogger article on Gold: The Radical Ideas Of Thomas Gold.

Dyson has written a challenging, exciting book, and I can only encourage its discovery!

A great study on European academic spin-offs

While in Antwerpen (Belgium) where I gave a workshop related to my vision of Silicon Valley described in the book Start-up, I had the nice opportunity (thanks Walter 🙂 ) to meet with Sven De Cleyn whose PhD thesis was published as a very interesting book: The early development of academic spin-offs : holistic study on the survival of 185 European product-oriented ventures using a resource-based perspective. The conversation I had with him convinced me I had to read his work. Although I am not sure I would advise anyone to do so (sorry Sven 🙁 but it is also 335 pages of dense content, including tables and statistical analyses), it is a great piece of work.

I had not seen before such work based on European spin-offs. Of course there is the American equivalent with Shane’s work, Academic Entrepreneurship: University Spinoffs And Wealth Creation and also lesser known but probably as good, the multiple papers of Junfu Zhang, including Entrepreneurship among Academics: A Study of University Spin-offs Using Venture Capital Data. But on the European side?

So let me summarize what I learnt. First an academic spin-off (ASO) is defined as “[1] a new legal entity (company) [2] founded by one or more individuals from an academic parent organization [3] to exploit some kind of knowledge [4] gained in the academic parent organization and transferred to the new company”. Then Sven studied two main questions:
– Research question 1
What characterizes the early development process of knowledge-intensive product-oriented academic spin-offs?
– Research question 2
What are the major criteria that determine the outcome of this process?

And he used 4 main theoretical frameworks (remember, it is a PhD thesis). I quote him again:
– the Resource-Based View of the firm (RBV), which posits that firm can only achieve a sustainable competitive advantage if they possess valuable, rare, inimitable and non-substitutable resources
– the Human Capital Theory (HCT), which explains firm survival and performance in terms of the firm’s knowledge, skills and experience, which can be created and accumulated through education, training and other experiences
– the Social Capital Theory. The central tenet suggests that a firm’s survival and performance are dependent upon the network to which it has access. The difference between SCT on one hand and RBV and HCT on the other, is that social capital, unlike other resource types, is not located within a firm but in the relationship with other actors
– the fourth and last theoretical stream relates to Life Cycle Models (LCMs). A firm’s resource needs are different in the first years after foundation when compared to more mature phases. The LCM theory builds upon the biological evolution of a living organism, where firms evolve through a number of distinct and consecutive stages and in the transition between the stages a number of important hurdles or junctures have to be overcome.

The heart of the dissertation is formed by a quantitative part. The first phase concerns the actual data collection, using personal interviews (in person, by telephone or using Skype) with 185 top managers (preferably one of the active founders) in nine European countries.

In addition here are some more figures about these spin-offs. Not surprisingly (for me!) the employment and turnover are not very high:

Some interesting not to say surprising results include:

+ on the Business side:
– The results of this study confirm that having developed a formal, written business does not contribute significantly to ASO survival probability.
– However, the results also indicate that several important constituent sections of a business plan (e.g. market study, production analysis …) contribute positively to survival likelihood.
– If the founders of the ASO have discussed the first drafts of the business plan with other entities (mostly externals), subsequent ASO survival chances might be affected. The analyses reveal that ASOs whose business plan has been screened, challenged and altered by risk capital providers face an increased failure risk.
– The results indicate ASOs using patents to protect their core technology do not necessarily enjoy higher survival probability.
– The results indicate that product development teams with (potential) customers aboard increase ASO survival likelihood substantially.

+ on the team side:
– Prior entrepreneurial experience (whether in a high-tech venture or not) has a significant positive impact, but this effect disappears and becomes negative for serial entrepreneurs. The latter result is ascribed to entrepreneurial euphoria (successful entrepreneurs tend to search less for relevant information in subsequent new initiatives).
Heterogeneous team, which are composed of a mixed background, contribute significantly and in a positive way to ASO success probability.
– In first instance, an ongoing active involvement of at least one of the (key) founders has a significant positive effect on ASO survival probability. However, some additional tests reveal that preferably this involvement does not occur as CEO. Secondly, a long-term and strong representation of the founders in the shareholder structure (meaning a shareholdership of at least 50% for the entire group of founders) appears to significantly enhance ASO survival likelihood.

+ about stakeholders:
– The origin of financial resources has no significant impact on its subsequent success (i.e. risk capital does not increase success likelihood when compared to any other funding source). On the other hand, ASOs able to attract subsidies for the development of their business, technology or product have significantly better survival odds than ASOs who don’t.
Academic continued, long-term support (at least up till the moment of interview) turns out to have a positive and significant impact on ASO success probability.
– In the first place, the results have pointed to the importance of the timing of ASO foundation, as ASOs who developed at least a β-prototype at the time of foundation have significantly higher survival odds.

A couple more quotes that I found striking:
– [page 16] European ASOs are reported to be small-scale ventures (mostly one-man SMEs) with limited growth ambitions and clear visions. Yet, the main contribution of ASO does not lie in a fast organic growth, but rather in its contribution to the transfer of technologies and knowledge throughout their networks and in a reasonable rentability. (This is not a result of De Cleyn, but his own analysis of past research)
– [page 52] Founding teams seem to perform better than individual founders. They tend to experience less failure, more mergers and/or acquisitions and a larger employment growth. This superior performance might be due to a better access to resources and network relationships than solo start-ups have. (Same remark)

Now some personal comments. First the lessons are much richer than this short summary. So if you have an interest in the topic, you should read De Cleyn’s book. But most importantly, for me, it dramatically shows some critical differences between the European and the American scene, at least the scene I know well, Silicon Valley. Sven De Cleyn’s definition of success is basically (I hope I am correct) the opposite of failure, which for example implies that surviving is considered as success. I am not sure it is the vision Americans have (the famous Fail Fast that entrepreneurs use even independently of the objectives of venture capital, which hates nothing more than “living-dead”.) You might be interested in comparing Sven’s data with my own analysis of Stanford related start-ups. A last comment. I asked Sven why he did not cover licensing deals with universities. The answer was twofold. He already had enough material to cover and the topic is more sensitive than others. I agree!

As a short summary, I had not seen such a deep analysis of European start-ups, with so much statistical analysis. And… unfortunately, it seems to confirm the culture gap we have with the USA. It might be that we have a different way of doing things or it might be that we have not really understood what Silicon Valley is really about…

Andy Bechtolsheim talks at Stanford about the Process of Innovation

This morning, I got up at 4am for an unusual event, a talk by Andy Bechtolsheim back at Stanford University. And it was great! I took a couple of screenshots and notes. For those who would not know Andy, here is more below. And I should also add that Bechtoslheim is from Germany, I had mentioned him in a past article: Europeans and Silicon Valley. There should be the full video on Stanford Youtube in a few days…

More than 30 years ago as a Stanford graduate student, Andreas “Andy” Bechtolsheim designed a simple but powerful computer workstation that would help define the modern technology era and launch Sun Microsystems. He’s since founded three more startups, including cloud-networking company Arista Networks, where he is now chairman. His investing foresight is legendary. Not only was he the first major backer for Google, but he’s also been an early-stage investor in VMware, Brocade and others. Bechtolsheim will discuss the process of innovation and describe its importance to Silicon Valley.

Bechtolsheim began his talk with some historical background on innovation. If you want to only read about the lessons, jump to the end! (I am aware some of the screenshots are low-res…). Recent (I mean in the last 50 years) innovations have their roots in semiconductors, networking and the Internet and (Open-source) software as well as in an acceleration in technology development (including Moore’s law and a faster adoption cycle of products.) These are two slides about the semiconductor roadmap:

Then he showed how the Internet from Arpanet, to the browser and finally to social networking has accelerated the innovation cycle.

More importantly, he gave some clues about what innovation is about, and why start-ups have an advantage here. Innovation is not about R&D not even about marketing. It is about bringing a needed product to customers:

And he illustrated his arguments with the Apple case:

– Apple does not make a lot of R&D

– Apple does not really study customers

So how does innovation work. Here are some clues:

And very importantly, he finishes with the innovation culture at Apple, Google, Amazon and the lessons learnt:

In conclusion, Andy had great lessons:
– Innovation is not about R&D or customers, it is about products.
– Timing is critical, so focus.
– Big companies are about evolution not revolution.
Be the expert in your field and understand the market, both gives you (self-)confidence (to attract people.)
– Failure is not an issue in SV but fail fast. However outside of SV, you may have to hide for 30 years when you fail. In SV, not trying is the risk, not failing.
– He also discussed patenting, “a sore topic”.
– Following another question, he considered a major threat to innovation is the current weakness of venture capital (there is money, but the returns are not good and a lot of money goes in narrow fields – cleantech a few years ago, web2.0, etc)

Well the title was misleading. Innovation is not a process, it is a culture! If you like this, you have to watch the video…

The New Facebook Legacy

Facebook will not only produce new millionaires with its IPO next Thursday; it has already created a new generation of entrepreneurs and start-ups. The New York Times just published an article A Circle of Tech: Collect Payout, Do a Start-Up and a related video Facebook’s Network of Tech Tycoons which illustrates the fact.


A few Facebook alumni entrepreneurs

I had already showed the power of networks when I commented Once you’re lucky, Twice you’re good a book subtitled The Rebirth of Silicon Valley and the Rise of Web 2.0. You can check again the web of people connected ten years ago or so. I had done the same with the older and now mature EDA industry. This new NYT article shows new connections illustrated by the new figure below:

Let me just quote the article: “The history of Silicon Valley has always been one generation of companies gives birth to great companies that follow”…”This is the story line of Silicon Valley, from Apple to Netscape to PayPal and now, to Facebook.” … and finally, “the social fabric of Silicon Valley is a dense set of overlapping spider webs, meaning everyone is connected.”

In my article on the web2.0, I had also shown the value creation. There had been $800M of VC money invested for a $17B value creation (mostly paper value). The new table below adds another $100M of VC money, and the value creation is now… $113B!!

Semiconductor start-ups. Is it the end?

I was lucky to be invited as a panelist at the Global Semiconductor Conference in Geneva on May 8-9. The round-table topic was “how to create more successful start-ups”. But before mentioning that discussion, I’d like to mention the previous panel, which gathered Stan Boland, former CEO and co-founder of Icera Inc, Dennis Segers, CEO of Tabula and Remy de Tonnac, Chief Executive Officer, INSIDE Secure. Stan has sold Icera to nVidia for $367M after raising $250M (a nice but small 1.3x return). Dennis has raised about $200M for Tabula and apologized for preventing start-ups to get that money, whereas Remy just took INSIDE Secure public on the Paris stock exchange, raising €70M after the company raised more than €100M in venture money since its inception.

I was very impressed by INSIDE Secure long history (it was founded in 1994), including unfortunately washout rounds. What was great is de Tonnac’s message stating that start-ups can only survive if they keep their entrepreneurial spirit and innovation DNA. Here is my usual cap. table format for the company. [The history, the list of investors and number of financial rounds are so long that the numbers might be approximation only…]


click on picture to enlarge

This shows once again that it is possible to try and succeed in Europe, but it seems to take much more time than in the USA. Now back to my panel. The motivation for the topic is the smaller and smaller number of funded semicon start-ups as the next figure shows.


click on picture to enlarge

And apparently, the main reason for this “crisis” comes from the huge amount of money these start-ups needs to reach profitability.

click on picture to enlarge

Well, if it was just about the amount of financing needed, biotech would be dead too, and my recent post Biotech IPOs, not so different shows it is not true. There might be at least two other reasons which explain the difference:
– you cannot go public without any revenue as it is the case with biotech, and I am not sure why (is it because the life cycle of semicon products is much shorter?) and
– the financial ratios of semicon companies are not great (Intel, the market leader is worth 2.5x its sales and 10x its profits).
But I am not fully convinced by the argument.

In fact I had another argument which might be simply said a lack of creativity combined with a culture of collaboration which has been lost. Indeed, the day before, another panelist said “why the heck would I share it, if I had the killer app”. Well, one might not share a killer app, but in Silicon Valley, there has been a lot of sharing:

Even today, people at LinkedIn and Facebook help each other even if they compete. I am less sure this happens at Google or Apple though! And here is what the NPR Radio “Morning Edition” had to say about the Silicon Valley ingredients. I strongly believe and agree with de Tonnac that we need more Entrepreneurial Spirit and Innovation DNA.

Edouard Bugnion at EPFL: back to Switzerland!

If you do not know Edouard Bugnion, you can read the article I had published in 2010, a Swiss in Silicon Valley. Edouard was at EPFL last week and gave a great technical talk about VMWare and virtualization, in fact a summary of his PhD thesis. The full text comes below.


Edouard Bugnion with the author in the middle of « cubicles » at Nuova in May 2006 (Picture: Mehdi Aminian).

For the anecdote, you can notice than Ed began his PhD in 1994 and will only finish it in 2012! As Martin Vetterli, dean of Computer Science at EPFL, said “at least he is finishing it, contrarily to the Yahoo or Google founders!”. The reason why it took Ed so much time is that he co-founded VMWare and Nuova in between… I have such a friend from Stanford who was doing his MS / early PhD in 1990-92 and obtained finally his diploma in 2004. He also had his start-up journey in the middle. (Faster though, Michael!) This back and forth adventures are not very common in Europe…

I noticed 2 other great lessons from Ed’s talk:
– as mentioned below “Virtual machines quickly lost popularity with the increased sophistication of operating systems” and it did not prevent VM to become a great market through VMWare success. Market dynamics are never simple and great opportunities may come from less explored areas of technology.
– Ed also explained that they did not or could not partner with established players (microprocessors or OS vendors) for various reasons. You can imagine the big players did not care, or would not change / adapt their (strategic) products or features. So when Ed was asked if this was not risky, he answered, risk is (sometimes) good.

Once again, this proves first that SV success comes also from immigrants and second, we need these people and their experience back! Hopefully we will have him at our ventureideas conferences. Invitation launched!


EPFL IC Seminar : “Using Virtual Machines in Modern Computing Environments with Limited Architectural Support”
By Edouard Bugnion, Stanford University

Abstract
Virtualization has gone through a full “popularity cycle”. Originally conceived in the mainframe era, virtual machines provided an efficient, isolated, and compatible duplicate of the hardware of the underlying machine. Virtual machines quickly lost popularity with the increased sophistication of operating systems, and subsequent processor architectures were designed without consideration for virtualization.

In this talk, I propose to use virtual machines to address limitations of commodity operating systems on modern architectures, even in the absence of architectural support for virtualization in the hardware. The primary technical contributions of the work were developed as part of two systems, each built for platforms with limited architectural support for virtualization. First, Disco ran commodity operating systems on scalable MIPS multiprocessors. Disco enabled virtual machines to form a virtual cluster that could transparently share the resources of the underlying multiprocessor. Second, VMware Workstation is a successful commercial product that allows multiple, unmodified operating systems to run concurrently on the same x86 system, allowing users to decouple their guest operating systems from the underlying hardware. VMware Workstation was the first 32-bit virtual machine monitor for the x86 architecture, and demonstrated that the x86 architecture was indeed virtualizable, despite a lack of architectural support.

Today, and in part because of the impact of Disco and VMware, virtual machines once again play a foundational role in Information Technology, and current-generation hardware provides architectural support for virtualization, similar to what already existed decades ago on mainframes.

Biography
Edouard Bugnion started his Ph.D. at Stanford in 1994, and is expecting to finish it this month. In the meantime, he co-founded two successful companies: VMware and Nuova Systems (acquired by Cisco). At VMware from 1998 until 2005, he served multiple roles including CTO. At Nuova/Cisco from 2005 until 2011, he built the core software team and became the VP/CTO of Cisco’s Server, Access, and Virtualization Technology Group, a group that brought to market Cisco’s Unified Computing System platform for virtualized datacenters.

His research interests include computer systems, datacenter and cloud networking, as well as technology entrepreneurship. For their work, Bugnion and his colleagues have received the ACM Software System Award (for VMware) and the ACM SIGOPS Hall of Fame Award (for Disco). Edouard was raised in Neuchatel, Geneva, and graduated from ETHZ.

Biotech IPOs, not so different

I just read Biotech IPOs Start to Show Some Modest Signs of Life from Xconomy. It’s an interesting article because it focuses on Biotech, a field that many people consider as very different from other high-tech start-ups such as Internet, Software or IT in general. The general idea is that it takes much longer to succeed in biotech. You should read the article if biotech is of interest for you and I will not comment it more than mentioning that the good news is that there have been recent biotech filings and IPOs, the less good news being that the market capitalizations are not huge.

What I am more interested in is updating my regular analysis of start-up data (I have now 131 start-ups; see my latest analysis in March 2012 for example with 116 companies) and see how biotech behaves. Here is the synthesis (if you are interested the detailed list is provided at the end).

So what do I see as specific to biotech start-ups? First it does not take them longer to go public. 8 years vs. an average of 7 years. The difference is not in the exit time. They raise $98M on average, but this does not look so special either. But, and here is the but, their sales are only $11M when they go public. So, it takes them much longer to reach revenues. But it does not prevent them from going public (or even be acquired when they begin to have good results in clinical trials).

Another specific element is about founders. The founders’ average age is 41 (similar to medtech and semiconductor) whereas it is 35 on average. Why is that? because many founders are established, recognized university professors. Often times, they do not work full-time in the start-up but have a role of chief scientist. Indeed, the ownership of founders in the start-up is smaller than average (8% vs. 15%).

I should also add that the founders/employee shares ownership is much smaller too (25% vs. 40%) and the reasons are manyfold:
– founders have fewer shares as I just mentioned
– investors have more equity (50% vs. 45%)
– IPO shares are higher (25% vs. 16%). This comes from the fact (I think) that in order to raise the same amount of money, it is more dilutive for a company with less revenue…
– I did not mention another statistical element, which is they have fewer employees. The detailed table below imples about 100 employees (and you may see many of them have even less than 50 or 20 employees). This induces a smaller amount of stock options… (On average my 130 companies have 500 employees when they go public).

I thought this data was of some interest. Please react or comment!

Appendix: detailed data (notice that I am missing the Amgen data)


click on table to enlarge