You have to be pretty brave to be a venture capitalist and keep an “anti-portfolio” page on your website. This lists the deals that you were offered but turned down. Bessemer Ventures is the only VC I know that does this. They’ve had some great exits over the years (such as Skype, Hotjobs, Parametric Technology or going back, Ungermann-Bass). But they also turned down FedEx (7 times), Intel and Paypal. And here’s their description of another one that got away:
Cowan’s college friend rented her garage to Sergey and Larry for their first year. In 1999 and 2000 she tried to introduce Cowan to “these two really smart Stanford students writing a search engine”. Students? A new search engine? In the most important moment ever for Bessemer’s anti-portfolio, Cowan asked her, “How can I get out of this house without going anywhere near your garage?”
Most of us who’ve been in Silicon Valley for a long time also have our own sort of anti-portfolio: the companies that we could have worked for and didn’t. I’ve never interviewed and then turned down a job that turned out to be a really huge mistake, but I have been invited to come and interview and refused.
When SDA (the fore-runner of Cadence) was founded, a friend of mine, Graham Wood, actually someone I shared an office with doing my PhD at Edinburgh, relocated to California from Bell Labs to join them. He was about employee number 20 and went on to be the creator of SKILL, still significant today in the battle over Cadence’s Virtuoso franchise. He asked me to come and interview. But at the time I was happy at VLSI Technology and thought we were going to change the world. In a way, of course, we did. But I wasn’t smart enough to see that the real money in ASIC was not in building the chips but in building the software to build the chips. Plus, if you are going to write EDA software, you are may as well do it at an EDA company where you are regarded as valuable, rather than at a semiconductor company where you are regarded as a weird item on the expense line.
Wes Patterson ran VLSI’s design centers and he left to be CEO of Xilinx. Somebody, not Wes himself, invited me over to interview but I was too blind to see that they would become a big success. “RAM-programmable gate-arrays? Who’d use one of those? How big a market is that?” Again I failed to realize that the cost structure might not have been great at first, but it was only going to get better. If you only want a handful of parts then FPGAs are the only sensible solution. For high volume, ASIC is the way to go. But over time, the cutover point creeps up and up and FPGAs serve more and more of the market.
Many years later I was headhunted by Xilinx to come and interview to run their entire software business and make it profitable. I wasn’t sure how feasible this was. Based on my experience at VLSI you can’t run a real software P&L inside a semiconductor company since you are an enabler for silicon and, if a customer is important, the company will just give everything for free, but not compensate your P&L with some of the silicon profits you enable. I interviewed with Wim Reolandts, the CEO who’d recently joined from HP. They asked me to come back in but that same week I was asked to become CEO of Compass. Who knows which would have been the better opportunity? Being a CEO is really hard unless you have the one qualification that everyone wants: you’ve been a CEO before. So it is always a good idea to take it when fate offers you that opportunity.
After Ambit was acquired by Cadence, Al Stein, VLSI’s CEO, tried to interest me in running a venture capital portfolio for VLSI. It was all the rage then for companies to take some of their cash and try and use their specialized inside knowledge to pick some winners for investment. I went and talked to the guy who ran the equivalent fund for Adobe, who’d started the trend. He’d invested in Netscape and other early internet successes and been wildly successful. But in the end I stayed at Cadence since I wasn’t convinced I’d be that good a venture capitalist. In retrospect I should probably have taken the job. I’m sure it would have been really interesting irrespective of how successful I turned out to be, and jobs are really interesting when you are learning a lot. In fact, if I’d taken it, it would have been fairly short lived since soon after Philips Semiconductors (now NXP) bought VLSI.
One job I did interview for and eventually turn down wisely, was to help run the software arm of European Silicon Structures. This was a company set up in Europe (duh!) to use e-beam technology to do very small runs of wafers cost-effectively. I turned the job down when the CEO couldn’t convince me of a good reason to have a large well-funded software division since clearly the company stood or fell based on how good the e-beam technology turned out to be. By then I’d got smart enough to know that you don’t want to be in an “expense” department in a semiconductor company. It turned out the e-beam technology didn’t work that well and the company failed. I think Cadence picked over the bones of the software division.
I was never offered a single digit badge number job at Google or anything like that. But it is always hard to tell which jobs are going to be with companies that turn out to be wildly successful. I asked a friend of mine who worked for me briefly as my finance guy before going on to be the CFO of Ebay and lead the most successful IPO of all time what was the most important criterion for success: “Luck.”