Get email delivery of the Cadence blog featured here
Normally the Stanford VLAB meets in Menlo Park, but occasionally they make a foray up to the city, as they did last Thursday for a session with Ann Winblad, who was, I believe, the first female venture capitalist. The location of the meeting was kept secret, for some reason, and only revealed to people who had signed up to attend the day before. Only in San Francisco would a meeting about business models have to worry about gate-crashers.
The meeting was billed as A Fireside Chat with Ann Winblad: Disruptive Business Models in Enterprise Software. Apart from the lack of a fire it was exactly as billed. It was moderated by Lars Leckie, the managing director of Hummer Winblad (HW). Unlike many other VCs, HW only invest in early stages and only in enterprise software companies. In fact, they were the first people to focus on software, back before there was really a software market to focus on.
Ann started her career working in a bank as a programmer. She left the bank when they decided not to reimburse her studies since it would make her the most qualified person at the bank and, as a woman, that would never do. She resigned the next day. She started a software company called Open Systems and they gave the source code away. This was in 1975, long before anyone was thinking about software that is "free as in freedom" or coined the phrase "open source." This was before Microsoft or Oracle were founded. When she tried to raise money, there were no VCs, especially in Minnesota where she was. However, there were 10 banks in Minnesota and after going around 8 of them, she got to the point where she could predict when they would say 'no.' So she practiced bursting into tears at that moment with the result that she walked out of the 10th bank with a $75,000 check. Needless to say, she does not recommend this approach for pitching today!
When they started Hummer Winblad, software was untouched. Microsoft and Great Plains and some others started around the same time. HW must have been quite a sight since John Hummer is a 6' 10" first round draft pick for the NBA and I would guess that Ann has to be around 5' 4".
Ann talked about one of their first investments, a company called MuleSource. They noticed that when they talked to Oracle, IBM and so forth one name kept coming up. The whole company was built around a technical guy in Malta (the island in the Mediterranean, not the town in upstate NY). There were two guys in Buenos Aires. A guy in Gary, Indiana. The co-founder was in San Francisco. The first customer was JP Morgan. Banks were building systems on it, airlines building reservation systems. It already had 150,000 downloads into corporations. They invested, eventually put in a new CEO. Last year they did $100M, will do twice that this year. There are 700 employees, with 200 in Buenos Aires.
Ann went through this company in a certain amount of detail for a reason. It is typical of what an enterprise software company needs to do: start as a product company and become a platform company. Somewhere along the way it is important to establish a business model. It is also good to be in an area of software that is not overfunded because then, as Ann put it, "the pond becomes polluted." That is one reason they focus on very early investments, before others have noticed that there is a segment to invest in. As another VC said to me once, "VCs make sheep look like independent thinkers."
HW tends to invest very early, not just in companies but in market segments. Their first fund invested in the company that became Hyperion when it was just two guys. There was no segment called business intelligence. They were in client-server, but they had no server. So HW bought them a server and gave it to them. This was an era in which large corporations were FedExing diskettes to calculate important metrics like their book-to-bill ratio. So they grew fast to what they are today. The challenge in these sort of companies is to position themselves in the new segment and become the de facto standard, which is the gold ring available for being the first.
Several startup CEOs then came and discussed the issues they were having, and their potential public embarrassment was counterbalanced by free advice from Ann.
One company was dealing with small businesses as customers. Ann calls them the prosumer, since you need to learn a lot of the tricks of the consumer market (you are not going to have a big expensive sales force, for example), but the product is like a big business product. Ann advised not going vertical too early, pointing out that only this year is Salesforce going veritcal.
These days it is easy to have a web front end, but the challenge is whether the users can get value in 10-15 months with almost no help from the company (since help does not scale). But becoming a platform too early (Ann used Github as an example) can be a mistake since you can get real revenue but not get huge. Ann doesn't like proof of concept from government or academia since it is always slightly off what the market really needs longer term. She prefers acceptance at a company like Goldman Sachs. After all, they have 22,000 developers.
Another company was a typical "double bottom line" company, which is also sometimes called a two-sided market (think trying to get people to use Airbnb when there are hardly any places to rent, at the same time as encouraging people to sign up when there are few users). Sometimes getting things going requires "unnatural acts." Actually, Airbnb did that in the very early days: they would photograph your apartment and put together the listing for you. So these acts are unnatural in the sense that they don't scale, but things that don't scale may be the only way to get started. One advantage that companies have entering the market now is that it is all mobile first. Companies that have been in the market for a long time, like OpenTable or Salesforce, have great web interfaces but their mobile interfaces are poor.
A few statistics about Hummer Winblad.
Previous: Open Server Summit: How to Install 5,000 Servers Per Day
Next: Phil Moorby and the History of Verilog