Is Silicon Valley Reproducible?

I’ve linked here to a post by Paul Graham, derived from a keynote at Xtech, that talks about the reproducibility of Silicon Valley. (In addition, here are links to two posts that commenting on it. The first is from slashdot, the second from Curt Monash.)

I have some direct personal experience on this topic, so rather than commenting on Paul Graham’s thoughts, I’ll just tell my own story. My experience derives from five years working in France (of nine in total) at Business Objects.

Most people don’t know that Business Objects is a French company. Fewer still know that it was in part founded as an experiment to reproduce Silicon Valley in France.

Business Objects was founded by a sales manager and a marketing director from Oracle France, both around 27 years old when they founded the company in 1990. The marketing director was Bernard Liautaud, who continues today at the company’s as Chairman and Chief Strategy Officer. The sales manager was Denis Payre, who hired me, and who left the company in the late 1990s.

I joined the company in 1995. I knew the general space (database and tools) quite well. I loved the team. As a native New Yorker, I have never really liked the endless-suburb feel of Silicon Valley and was eager to both work in a real city again and to try to export some of my hard-earned experience to France.

When I joined the company, it was still very much trying to replicate Silicon Valley. At the time:

• Most French software startups were almost entirely founder-owned. Founders preferred control to capital. Most would rather have owned a large piece of a small pie. Business Objects raised venture capital (from both US and European investors).

• Most French startups lacked heavies on the board of directors. We had, among others, Arnold Silverman, an early investor in Oracle and a tremendous resource for the company (who I always thought of as an invisible hand).

• Most French startups didn’t grant stock options. We did. (Subsequent French tax laws have made them much less interesting to your typical employee.)

• Most French startups targeted the French market first and planned go abroad once they had critical mass (which was they rarely achieved). Business Objects targeted international markets from the start. It knew that to “make it” in enterprise software meant establishing leadership in the US.

• Most French startups had all-French management teams. Business Objects directly targeted an international team and imported talent from the USA (e.g., marketing, CFO) for areas where they felt the USA had the best labor pool.

• Most French startups had French names. Business Objects was deliberately picked to sound American.

I’d say this “French transnational” strategy served the company very well in the early days. The company went public in 1994 on the NASDAQ (not the French bourse). But establishing leadership in the US market was elusive. Here’s why:

• Americans aren’t used to being country managers. They don’t even understand the job description. When you hire someone in Germany to be GM or president of Something-Germany, AG, they know what the job is. In America, they think they’re running the show and immediately stumble into issues of corporate strategy and positioning. They simply aren’t used to taking orders from overseas, and acting as the head of a distribution business.

• The US market is the most competitive (strictly speaking: the number of competitors) in software so the “stuff hits the fan” first in the USA. Even if you want to develop your software abroad, your market sensors must be in the USA and it’s hard to do that with most of your corporate body overseas.

And it just got increasingly harder to run the company out of Paris:

• The bigger the company got, the more French the Paris office got. In the early days, I believe local French staff joined precisely because they didn’t want to work a typical French company. Perhaps this was a core values / recruiting mistake, but the bigger we got the more we hired people who didn’t want to work at a Silicon Valley company in France. They wanted Silicon Valley perks like stock options but six weeks of vacation and a labor union. (I had to be peeled off the ceiling when the union activity starting kicking up.)

• The expatriate executives lasted, on average, about two years. A key part of the formula was bringing in experience from overseas. International relocation is very hard on spouses and families and the senior folks, who bring the most experience, typically have them. So it began to feel like a revolving door of senior management expatriates. It’s hard to run a company that way.

• By around 1997 executive staff members started moving to the USA because we’d decided that it USA market leadership was critical and that it just too hard to get it from Paris. Ironically, when I moved back to the USA headquarters in July, 2000, only the VP of R&D remained in Paris.

You can argue that it’s possible to build a great enterprise software company in France (after all, SAP seems to have done just fine in Germany), but that we simply made execution mistakes. While I’m sure we made some mistakes, I am convinced that you cannot reproduce the Silicon Valley innovation engine in France.

Here’s why:

• Cultural attitudes towards failure. In the USA, a failed startup is a red badge of courage. In France, it’s a bit less bad than death. You cannot create a great innovation engine by counting strikeouts. You need to count hits or home runs.

• Talent leakage. Many/most of the senior executives who helped build Business Objects no longer live in France. I’d guess this is largely due to tax policy (e.g., the wealth tax that runs about 2% of net worth). In Silicon Valley, when you “make it,” you either (1) do it again with another startup, (2) become a venture capitalist, (3) “retire” to sit on boards of directors, or (4) become a consultant. Few people leave the system. The vast majority of those who have been successful move to Woodside or Atherton and recycle their expertise.

• Lack of experienced labor pool. It’s not just about hiring smart engineers (which you can find), but smart engineers who have built system software. (This may be one reason why SAP fared better than we did; they were building application software, we were building system software.) But you need experienced managers in all departments and experienced senior executives. My key takeaway from the whole experience: it’s not just about smarts, but experience as well.

• Lack of infrastructure. I’m referring to the vast number of agencies in the Silicon Valley ecosystem such as PR firms, UI design consultants, quality consultants, sales trainers, strategy consultants (e.g., Geoffrey Moore), management coaches, etc.

That’s my take. There are always exceptions: Dassault is a fine company which I think is still run from France. Ilog did quite well, but I think they have moved many leadership posts to the Valley. Cartesis is trying to repeat the Business Objects experience in performance management (and I hope CMO Crispin Read is continuing a few of the fine marketing traditions we had at Business Objects).

But, exceptions aside, can you reproduce Silicon Valley anywhere? Probably not. Can you make it in France? In my estimation, no way.