A few days ago, I had the opportunity to present a video conference on equity sharing in a startup, between founders, investors and employees. I’ve done it many times in the last few years like the one Slideshare here, but I had never recorded it. It’s now done:
“Prediction is very difficult, especially about the future.” attributed to Niels Bohr.
I was asked yesterday which startups I knew were the most promising, not to say the greatest. So I prefer to refer you to the quote above as I did not understand the potential of Google and Skype when I first heard of them. I am less shy of my lack of talent as this difficulty in predicting has been acknowledged by others.
First from the book The Business of Venture Capital on page 207:
Legendary investor Warren Buffet admired Bob Noyce, cofounder of Fairchlid Semiconductor and Intel. Buffet and Noyce were fellow trustees at Grinnell College, but when presented, Buffet passed on Intel, one of the greatest investing opportunities of his life. Buffet seemed “comfortably antiquated” when it came to new technology companies and had a long-standing bias against technology investments.
Peter O. Crisp of Venrock adds his misses to the list: One “small company in Rochester, New York [came to us, and one of our junior guys] saw no future [for] this product… that company, Haloid, became Xerox.” They also passed on Tandem, Compaq and Amgen.
ARCH Venture Partners missed Netscape – that little project Marc Andreessen started at the University of Chicago. An opportunity that, according to Steven Lazarus, would have been worth billions! “We just never knocked at the right door,” he would say. Eventually, ARCH decided to hire full-time person to just keep tabs on technology coming out of the universities to “make certain we don’t miss that door next time.”
Deepak Kamra from Canaan Partners comments on his regrets: “Oh, God, I have too many … this gets me depressed. A friend of mine at Sun Microsystems called and asked me to meet with an engineer at Xerox PARC who had some ideas to design a chip and add some protocols to build what is now known as a router. The drivers of bandwidth and Web traffic were strong market indicators, and he was just looking for $100,000. I really don’t do deals that small and told him lo raise some money from friends and family and come back when he had something to show” That engineer was the founder of Juniper Networks. He got his $100,000 from Vinod Khosla. Khosla, then with KPCB, added an IPO to his long list of winners. Juniper slipped out of Kamra’s hands because it was too early.
And of course, those were frothy times when everyone was deluged with hundreds of opportunities each day.
KPCB missed an opportunity to invest in VMWare because the valuation was too high: a mistake, according to John Doerr.
Draper Fisher Jurvetson (DFJ) was initially willing but eventually passed on Facebook (ouch!), as the firm believed the valuation was too high at $100 million pre-money.
KPCB, not wanting to be left out of an opportunity like Facebook, invested $38 million alt a $52 billion valuation.
Tim Draper of DFJ, turned down Google “because we already had six search engines in our portfolio.”
K. Ram Shriram almost missed his opportunity to invest in Google when he turned the founders away. “I told Sergey and Larry that the time for search engines had come and gone but I am happy to introduce you to all the others, who may want to buy your technology. But six months later, Ram Shriram, who had once turned Google down, now invested $500,000 as one of the first angel investors.
Now some examples of the updated BVP antiportfolio:
AirBnB: Jeremy Levine met Brian Chesky in January 2010, the first $100K revenue month. Brian’s $40M valuation ask was “crazy,” but Jeremy was impressed and made a plan to reconnect in May. Unbeknownst to Jeremy, $100K in January became 200 in February and 300 in March. In April, Airbnb raised money at 1.5X the “crazy” price.
Facebook: Jeremy Levine spent a weekend at a corporate retreat in the summer of 2004 dodging persistent Harvard undergrad Eduardo Saverin’s rabid pitch. Finally, cornered in a lunch line, Jeremy delivered some sage advice, “Kid, haven’t you heard of Friendster? Move on. It’s over!”
Atlassian: Byron Deeter flew straight to Atlassian in 2006 when he caught wind of a developer tool from Australia (of all places!). Notes from the meeting included “totally self-financed, started with a credit card” and “great business, but Scott & Mike don’t ever want to be a public company.” Years and countless meetings later, the first opportunity to invest emerged in 2010, but the $400m company valuation was thought to be a tad “rich.” In 2015, Atlassian became the largest tech IPO in Australian history, and the shares we passed on are worth more than a billion dollars today.
Tesla: In 2006 Byron Deeter met the team and test-drove a roadster. He put a deposit on the car, but passed on the negative margin company telling his partners, “It’s a win-win. I get a great car and some other VC pays for it!” The company passed $30B in market cap in 2014. Byron paid full price for his Model X.
eBay: David Cowan passed on the Series A round. Rookie team, regulatory nightmare, and, 4 years later, a $1.5 billion acquisition by eBay.
My naive and obsessive quest for startup cap. tables has led me today to a thriller-like research! First I will let you have a look at Darktrace cap. table which I decided to study as it belongs to the short list of UK unicorns together with Revolut and Graphcore.
Well the first surprising information is the founding structure, ICP London. Why a British Virgin Islands structure? To hide who the founders are? Then I discovered surprising board members, Michael Lynch, the founder of Autonomy and Sushovan Hussain, the former Autonomy CFO… Autonomy was always a puzzle to me before becoming a scandalous HP acquisition and then the cause of a huge trial, not decided yet… And in law, you are innocent unless proven guilty.
Another strange side of the company is its links to secret services, MI5, CIA, NSA. Probably not so surprising when your industry is cybersecurity… Being based in Cambridge, it is not surprising that many Darktrace employees were at Autonomy before. The board members I did not know, but the investors are famous: Summit, KKR, Insight. Less maybe is Invoke Capital board members Vanessa Colomar and Andrew Camper. Lynch and Hussain are not on the board anymore and this is probably linked to the HP Autonomy litigation.
Then I got it: Invoke Capital Partners… ICP! So Darktrace was indeed founded by the former Autonomy people and its new investment structure, Invoke. I had to do a little more search and found two quite fascinating articles:
I read again today about the importance of the lean startup movement. I have never been a big fan. Of course you need to interact with customers (at least to sell something) but you should not become a slave of your customers and pivot as soon as you can not get validation from them.
Do not get me wrong, I am a big fan of Steve Blank and customer development, I use his work a lot. But there is so much uncertainty, the tool should not replace the vision and intuition of the entrepreneur. Let me quote again Horowitz for example: “Figuring out the right product is the innovator’s job, not the customer’s job. The customer only knows what she thinks she wants based on her experience with the current product. The innovator can take into account everything that’s possible, but often must go against what she knows to be true. As a result, innovation requires a combination of knowledge, skill, and courage. Sometimes only the founder has the courage to ignore the data.”
It reminded me I had read something about this from Peter Thiel. I found it again in a 2014 post: Should entrepreneurs have start-up skills? Two counterintuitive answers. Here is what Thiel had said: “What do I think about lean startups and iterative thinking where you get feedback from people versus complexity that may not work. I’m personally quite skeptical of all the lean startup methodology. I think the really great companies did something that was somewhat more of a quantum improvement that really differentiated them from everybody else. They typically did not do massive customer surveys, the people who ran these companies sometimes, not always, suffered from mild forms of Aspergers, so they were not actually that influenced, not that easily deterred, by what other people told them to do. I do think we’re way too focused on iteration as a modality and not enough on trying to have a virtual ESP link with the public and figuring it out ourselves.”
In a nutshell, “Lean Startup is best used as a teaching tool for those who need a little help in learning how to use their mirror neurons to feel the real needs of the real people they are seeking to serve. It can help to reduce waste. It can help to slow the rate of decline of organizations that are being disrupted.”
I recently published an updated version of a database of capitalization tables of 600 (former) startups. I obtain the data most of the time from the IPO prospectus of the company (that is the document the company publishes when it is listed on a public stock exchange, and in general Nasdaq.
These documents are an amazing source of information of all the business components of the companies even if I focus only on the shareholding and funding history. They are sometimes a little frustrating though as they do not cover the full history of the company, but only 3 to 5 years in the past so it is not simple to get the founders’ data for example.
Some countries do however provide access to the full company data, often for a fee like in France. A few cantons in Switzerland (Basel, Zurich) and the United Kingdom provide it for free and this is just great.
I have done some research for Revolut and Graphcore recently. Today, I revisited the data I had built for two British companies: Autonomy founded in 1996 and had gone public on Easdaq in 1998 and Bicycle Therapeutics, a biotech company with links to EPFL (Lausanne, Switzerland) founded in 2009 and public since July 2019.
The IPO documents did not provide enough for me about the founders and early rounds. So here are my new tables:
Bicycle from the IPO data
Bicycle from the UK register data, the updated cap. table, the funding rounds and its growth over time:
The funding rounds
The growth of revenues and jobs
A final post (for now) about the data about 600 (former) startups. So what have they become today in April 2020?
First a quick point of caution: I counted some companies twice, because I had looked at their equity strutcure at different points in time: Alibaba had two IPOs in 2007 in HK and 2014 in the USA, Esperion had 2 filings in 2000 and 2013. This is not a big deal, except if you count Alibaba’s value twice!
So out of the about 600 former startups, I found that
– 20 were still private (they may have recently filed for IPOs though)
– 12 were private again after an IPO
– 13 had stopped their activity (often through bankruptcy)
– 225 had been acquired or merged with another company (Merger and acquisitions – M&As)
– 331 were still public.
So let us have a closer look at M&As and public companies:
On the M&A side, the main acquisition value comes from biotech, with a $5B average value whereas software or internet is a little les below $3B.
On the public side, I will let you discover depending on your interest about, given the field, the number of companies, employees, cumulative market capitalizations, sales, profits, then age of companies and current average price to sales (PS), price to earnings (PE) and an interesting personal metrics, price ot employees in $M (Pemp).
Graphcore gave me concerns. How is it possible that the two founders, Simon Knowles (58) and Nigel Toon (56), two serial entrepreneurs, who founded Icera Semiconductor in the past (sold to Nvidia in 2011 for $435 million or $367 million depending the sources – after having raised $258 million) and Element14 (a 1999 spin-off of Acorn – or its new name – sold to Broadcom in 2000 for $640 million), each owns only 4 shares of the startup? Are they so rich that they don’t need more? !!
All this follows my recent discovery that the UK gave open access to all company and in particular startup data. I began with Revolut a few days ago and now Graphcore. There had to be something wrong. The startup could not have only investors as shareholders. And then of course, I had forgotten the ESOP, the employee stock-options. So my only explanation is that the founders are part of this too and have a minimal number of shares. Still intriguing!
Sixth post of comments on the 600 startup data. Today, it’s about the valuation of startups.
I had touched the topic on page 615 of the pdf. Here is the data again.
Be aware that these numbers are not typical of traditionnal companies. They show that startups going public are of a speculative nature, with a promise of very high-growth in the future. The multiples are very high and in the case of earnings, in fact most startups do not make a profit at IPO – about one out five!
Interestingly Silicon Valley does not have the highest multiples, but Europe is behing the USA.
The tables give the PS ratio (price to sales – ratio between valuation and sales) and PE ratio (price to earnings – ratio between valuation and profits) and the number of startups taken into account each time. The following curves show the PS values by 5-year periods and by year.
If you are lost, here are slides I have used in the past and if you want to get the excel file, send me an email.
A colleague of mine (thanks Agnès!) informed me that the United Kingdom made public its data about startups. This is just amazing!
So I checked about Revolut and found all the data I could dream of. Founders, rounds of funding, shareholders.
Two young founders from Eastern Europe origin, 29 and 30-year old at the time of founding.
Some big, somewhat strange, rounds and here is today’s cap table. However series E is a best guest whereas previous rounds were publicly availale.
This morning (April 13), I discovered an important inaccuracy, nothing wrong but still: what about the ESOP, the stock-options. They are mentioned in the company documents, so here is a modified cap. table, and see the difference! I must add this is the ESOP in Dec. 2018, so the number is probably bigger today.
Fifth post of comments on the 600 startup data. Today, it’s about the ownership of non-founding CEOs (compared to the founders).
I noticed a few months ago that in a majority of cases, the CEO was a founder. This was a surprise. The data confirms this: there are a total of 229 startups with a non-founding CEO out of 600 (38%). Again, fewer in the digital domain, and more in the health-related fields.
As you may see in my initial post, these CEOS have on average a 3% ownership (median value is 2.7%).
Is all this useful in practice? The median value of the ratio between the CEO & founders’s ownership is 0,5 (the average value is 1 because there are big outliers). Does this mean that if founders want to hire a CEO at foundation, he should have about 33% of the company, and at series A about 15% if you have read all my posts before!
Here is a more granular illustration.
The next image shows the founders’ ownership on the horizontal axis vs. the CEO’s on the vertical axis (with a zoom on the right).
A final illustration as food for thought, the founders’ ownership in the digital and health-related fields, relatively to the presence of a non-founding CEO or not. (Note that the vertical axis does not have the same scale for the two domains).