Category Archives: Start-up data

The analysis of 500+ startups

Following my traditional analysis of startups through their IPO filings documents (you can check my 2017 analysis on 400+ documents here or the tag #equity on this blog), here is an updated analysis with 500+ start-ups.

You can have a look at the full 500 cap. tables on scribd or look at a shorter synthesis which follows.I hope this is self-explanatory enough.

Uber finally files to go public – Here’s my cap. table

Uber’s S-1 has just been released. I jumped on the opportunity to analyze the shareholding of the startup, a thing I had tried to do in 2017 (with much less information – check here). Here are the figures that I found (subject to errors related to my possible too much eagerness…)


Uber cap. table – from the SEC S-1 published on April 12, 2019

And, if you do not have the courage to read my post What is the equity structure of Uber and Airbnb?, here is what I understood in March 2017:


Uber cap. table – A speculative exercise with very little information available

The Age of Founders of Start-ups – Again!

The age of founders of start-ups is a recurrent topic on this blog. You can just check it through hashtag #age. I have just updated my cap. table database with now 500 “famous-enough” companies for which I have compiled a lot of data. You can check here the most recent update with 450+ companies in mid 2018 – Some thoughts about European Tech. IPOs or a synthesis dated 2017 with 400 companies Equity in Startups.

I just looked at the age of 850 founders from these 500 companies. I think it is interesting. I hope you will agree… I am not even sure I need to comment much. Average age is 37 overall, 45 in biotech, 37 in hardware (electronics, telecom and computers, energy) and 32 in software/internet.

Fascinating data analyses on start-ups by Sebastian Quintero

I just read about Sebastian Quintero’s data analyses on start-ups on his web site Towards Data Science. Thanks Martin H. 🙂 I was really fascinated about his original way of looking at them, their failure rate, the valuation prediction, their runway between rounds, and his Capital Concentration Index or Investor Cluster Score. You should read them.

Of course, it rang strong bells with all the data analyses I have done in the recent past 8see end of the post if you wish)

So as an appetizer to Quintero‘s work, here are a couple of figures taken from his site…

Dissecting startup failure rates by stage

Predicting a Startup Valuation with Data Science

How much runway should you target between financing rounds?

Introducing the Capital Concentration Index™

Where c is the percentage capital share held by the i-th startup, and N is the total number of startups in the defined set. In general, the CCI approaches zero when a sector consists of a large number of startups with relatively equal levels of capital, and reaches a maximum of 10,000 when a sector’s total invested capital is consolidated in a single company. The CCI increases both as the number of startups in the sector decreases and as the disparity in capital traction between those startups increases.

Introducing the Investor Cluster Score™ — a measure of the signal produced by a startup’s capitalization table

As of my own analysis, here are a couple of links…

My papers on arxiv:
– Are Biotechnology Startups Different? https://arxiv.org/abs/1805.12108
– Equity in Startups https://arxiv.org/abs/1711.00661
– Startups and Stanford University https://arxiv.org/abs/1711.00644

or on SSRN
– Age and Experience of High-tech Entrepreneurs http://dx.doi.org/10.2139/ssrn.2416888
– Serial Entrepreneurs: Are They Better? – A View from Stanford University Alumni http://dx.doi.org/10.2139/ssrn.2416888
– Start-Ups at EPFL. An Analysis of EPFL’s Spin-Offs and Its Entrepreneurial Ecosystems Over 30 Years https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3317131

2019, the year of Unicorns IPO Filings: is Lyft the beginning of the end?

Lyft is the first Unicorn which published its S-1 document, i.e. its IPO filing. Is this good news or bad news? Lyft is impressive, two founders who were 22 and 23 when they co-founded their start-up 12 years ago have reached more than $2B in sales with a little less than 5’000 employees in 2018. This is the good part. The less good piece is it took the company more than $5B in equity investment and the reason is simple: Lyft has lost $900M in 2018, and more than $600M in both 2017 and 2016. This is more than $2B cumulative loss. I assume losses were pretty high in the previous years too. YOu can have a look at the cap. table I built from the S-1:

I read recently an article by Tim O’Reilly: The fundamental problem with Silicon Valley’s favorite growth strategy. O’Reilly has doubts about Reid Hoffman and Chris Yeh’s claiming that Blitzscaling would be the secret of success for today’s technology businesses. “Imagine, for a moment, a world in which Uber and Lyft hadn’t been able to raise billions of dollars in a winner-takes-all race to dominate the online ride-hailing market. How might that market have developed differently?” I have the same doubts about this crazy strategy but who am I to say?…

Swiss Startups : new analyzes, without real surprise …

A new and interesting report on Swiss startups has just been published by Startupticker, the Swiss Startup Radar.

It shows a fairly new information, the number of startup created a year, about 300,

and their slow growth …

Interesting testimonials also:

Is it due to the much-cited conditions? (Page 80)
No, Switzerland’s regulatory and fiscal framework is first-rate. But I identify two deficits in the support services available in Switzerland: first, there is a lack of contact points for entrepreneurs in the low and no-tech sectors, and, second, we tend to address young people.

More money is one thing, but is it spent differently? Page 89)
In Switzerland, I observe a strong focus on the survival rate. Startups are encouraged if they have collateral, such as patents, and take a cautious course. As a result, eight out of 10 startups from ETH Zurich are still active five years after their foundation. In Israel, on the other hand, more attention is paid to the economic impact. What matters when assessing a project is the prospect of growth and the creation of new jobs.

The awareness that investing in startups can lead to losses is undoubtedly more pronounced in Israel. This is particularly evident in the financing of very young projects. In Switzerland, seed rounds are worked on with thick business plans, PowerPoint presentations and sales projections. In Israel, this paper war has been largely dispensed with. The business angels and VCs accept that there can be no absolute security in the high-tech segment.

In an article by Techcrunch, 30 European startup CEOs call for better stock option policies, we also talk about the gaps in the framework conditions in Switzerland:

with the following recommendations:
1. Create a stock option scheme that is open to as many startups and employees as possible, offering favourable treatment in terms of regulation and taxation. Design a scheme based on existing models in the UK, Estonia or France to avoid further fragmentation and complexity.
2. Allow startups to issue stock options with non-voting rights, to avoid the burden of having to consult large numbers of minority shareholders.
3. Defer employee taxation to the point of sale of shares, when employees receive cash benefit for the first time.
4. Allow startups to issue stock options based on an accepted ‘fair market valuation’, which removes tax uncertainty.
5. Apply capital gains (or better) tax rates to employee share sales.
6. Reduce or remove corporate taxes associated with the use of stock options.

Google is 20 years old

Google was incorporated in California on September 4, 1998 so the company is just 20 years old today. The technology is older, it was called BackRub initially (in 1996) and was an internal web site at Stanford University, google.stanford.edu and in September 1997, google.com was registered as an independant web site. You can see below some historic images

and the various logos.

There’ve been many books about Google, some of them are great. I blogged about most of them, Work Rules! a few weeks ago, In The Plex in mid 2015, How Google Works in late 2014, Dogfight in early 2014, I’m Feeling Lucky in 2012. Indeed I blogged a lot about the company as you may see from the Google tag.

If Fairchild was the emblematic Silicon Valley company, founded in the 50s, it was followed by Intel in the 60s, Apple, in the 70s, the 80s have seen Cisco and Sun Microsystems, and Google symbolizes the 90s (Yahoo might be forgotten soon). Facebook belongs to the 2000s, the 2010 decade is still open I think. But the lessons learnt from the years of Google are just unique. The technology, the product, the startup growth, the teams have just changed the way we look at business for good and sometimes bad….

Work Rules! by Laszlo Bock (part IV) – Managers

Google has been famous for defiance of authority. Bock develops this further.

At google, we have always had a deep skepticism about management. This is just how engineers think: managers are a Dilbertian layer that at best protects the people doing the actual work from the even more poorly informed people higher up the org chart. But our Project Oxygen research, which we’ll cover in depth in chapter 8, showed the managers in fact do many good things. It turns out that we are not skeptical about managers per se. Rather, we are profoundly suspicious of power, and the way managers historically have abused it. [Page 118]

Acton who said “Power corrupts; absolute power corrupts absolutely” also wrote: Great mean are almost always bad men, even when they exercise influence and not authority: still more when you superadd the tendency or the certainty of corruption by authority, there is no worse heresy than that the office sanctifies the holder of it. That is the point at which … the end learns or justifies the means. [Pages 119-20]

It was such a deeply held belief that in 2002 Larry and Sergey eliminated all manager roles in the company. We had over three hundred engineers at the time, and anyone who was a manager was relieved of management responsibilities. Instead every engineer in the company reported to Wayne Rosing. It was a short-lived experiment. Wayne was besieged with requests for expense report approvals and for help in resolving interpersonal conflicts, and within six weeks the managers were reinstated.
[Page 190]

Still Project Oxygen initially set out to prove that managers don’t matter ended up demonstrating that good managers were crucial. [Page 188]. I will let you read Chapter 8 and here are the 8 rules from the study [Page 195]:
1- Be a good coach.
2- Empower the team and do not micromanage.
3- Express interest/concern for team members’ success and personal well-being.
4- Be very productive/results-oriented.
5- Be a good communicator – listen and share information.
6- Help the team with career development.
7- Have a clear vision/strategy for the team.
8- Have important technical skills that help advise the team.

I cannot finish this new post without mentioning a link given by Laszlo Bock about the history of Silicon Valley: “Silicon Valley’s Favorite Stories”, Bits (blog), New York Times, February 5, 2013.


Robert Noyce, right, set up an atmosphere of openness and risk at Fairchild Semiconductor.Credit Courtesy of Wayne Miller/Magnum Photos

Work Rules! by Laszlo Bock (part III) – (the invisible) women in the workplace

The gender gap has become a much more visible issue in 2018 and Bock is no exception (even if his book is older). But before I mention what he says about it, here are two recent and very interesting references:
– The New Yorker just published an article about the gender gap at work and particular at BBC: How the BBC Women Are Working Toward Equal Pay.
– France Culture tells the story of Margaret Hamilton, a software programmer on the Appolo project: (in French) Margaret Hamilton, la femme qui a fait atterrir l’Homme sur la Lune.


Margaret Hamilton during the Apollo program.• Credits : NASA

Now Bock: In one study conducted by Maura Belliveau of long Island University [1], 184 managers were asked to allocate salary increases across a group of employees. The increases aligned nicely with performance ratings. Then they were told that the company’s financial situation meant that funds were limited, but were given the same amount amount of funds to allocate. This time, men received 71 percent of the increase funds, compared to 29 percent for the women even though the men and women had the same distribution of ratings. The managers – of both genders – had given more to the men because they assumed women would be mollified by the explanation of the company’s performance, but that the men would not. they put more money toward the men to avoid what they feared would be a tough conversation. [Page 170]

[1] “Engendering Inequity? How Social Accounts Create vs. Merely Explain Unfavorable Pay Outcomes for Women” Organization Science 23 no 4 (2012) 1154-1174 published online September 28, 2011, https://pubsonline.informs.org/doi/abs/10.1287/orsc.1110.0691

Bock mentions another study on page 137 about graduates from Carnegie Mellon that is also mentioned in the New Yorker article as “As the economist Linda Babcock and the writer Sara Laschever explain, in their book “Women Don’t Ask,” women are less likely than men to negotiate for higher salaries and other benefits. At Carnegie Mellon University, for example, ninety-three per cent of female M.B.A. students accepted an initial salary offer, while only forty-three per cent of men did. Women incur heavy losses for their tendency to avoid negotiation. It is estimated that, over the course of her career, an average woman loses a total of somewhere between half a million and a million and a half dollars.” Additionally “Even when women do make it to the bargaining table, they often fare poorly. In “What Works: Gender Equality by Design,” the behavioral economist Iris Bohnet examines data from a group of Swedish job seekers, among whom women ended up with lower salaries than their equally qualified male peers. “Not only did employers counter women’s already lower demands with stingier counter-offers, they responded less positively when women tried to self-promote,” she writes. “Women, it turns out, cannot even exercise the same strategies for advancement that men benefit from.” When women act more like men, she suggests, they are often punished for it. Lean in, and you might get pushed even further back.

Work Rules! by Laszlo Bock (part II) – the GLAT

In Work Rules!, Bock mentions briefly the GLAT (Google Labs Aptitude Tests) that were also mentioned in David Vise’s Google Story. But he just quickly says they may have been overused and sometimes a waste of time and of resources. But let me refer to his page 73:

That page begins with the image above which can be also found on google blog’s page Warning: we brake for number theory. It’s never too late solve math problems… If you solved it at the time, you got access to the following one:

The second puzzle:
f(1)=7182818284 
f(2)=8182845904 
f(3)=8747135266 
f(4)=7427466391 
 f(5)= __________

Again feel free to try… you will find answers here. Bock just adds this: The result? We hired exactly zero people.

Maybe this will help you:

2.71828182845904523536028747135266249
7757247093699959574966967627724076630
3535475945713821785251664274274663919
3200305992181741359662904357290033429
5260595630738132328627943490763233829
8807531952510190115738341879307021540
8914993488416750924476146066808226480
0168477411853742345442437107539077744
9920695517027618386062613313845830007
5204493382656029760673711320070932870
9127443747047230696977209310141692836
8190255151086574637721112523897844250
5695369677078544996996794686445490598
7931636889230098793127736178215424999
2295763514822082698951936680331825288
6939849646510582093923982948879332036
2509443117301238197068416140397019837
6793206832823764648042953118023287825
>0981945581530175671736133206981125099

as well as this:

x = 1
2.71828182845904523536028747135266249
7757247093699959574966967627724076630
3535475945713821785251664274274663919

x = 2
2.71828182845904523536028747135266249
7757247093699959574966967627724076630
3535475945713821785251664274274663919

x = 3
2.71828182845904523536028747135266249
7757247093699959574966967627724076630
3535475945713821785251664274274663919

x = 4
2.71828182845904523536028747135266249
7757247093699959574966967627724076630
3535475945713821785251664274274663919

x = 5
2.71828182845904523536028747135266249
7757247093699959574966967627724076630
3535475945713821785251664274274663919
3200305992181741359662904357290033429
5260595630738132328627943490763233829
8807531952510190115738341879307021540
8914993488416750924476146066808226480
0168477411853742345442437107539077744
9920695517027618386062613313845830007
5204493382656029760673711320070932870
9127443747047230696977209310141692836
8190255151086574637721112523897844250
5695369677078544996996794686445490598
7931636889230098793127736178215424999
2295763514822082698951936680331825288
6939849646510582093923982948879332036
2509443117301238197068416140397019837
6793206832823764648042953118023287825
0981945581530175671736133206981125099