That got us going. In 1994 we spent two or three months developing the idea, and invited a friend, Ed Scott, to join us. Ed’s background was in enterprise selling, and his input was a very critical insight. Most of the startups we see in the Valley are built this way: you “invent fire” then the fire sells itself, and everybody goes away happy and makes a lot of money. But distribution is a very critical element, and that changed the angle of how we wanted to get the company started.

What was the approach?

In mid-1994 we had gotten to the point where we had a preliminary business plan, and we were shopping for funding. Because Ed was a sales guy, he gave us dose of reality. He said, “This is what you need to sell this stuff effectively.” And then we did the math backwards. If we wanted to be a publicly held company we had to build out a distribution engine. And that’s a problem that a lot of Valley companies ignore. You see a lot of startups go out of business, not because they don’t have good ideas, but because they can’t sell. These days, so many startup companies want to partner with us to sell stuff through our channel. I tell them: “My channel is to sell my stuff, not to sell your stuff!” You have to build a sales channel.

To make a long story short, we magically came up with the number: We needed $50 million to get our business going.

That was a lot of money at that time.

Oh, a huge amount of money in 1994. Nobody in software had ever raised that kind of series A. We met with the usual suspects — Mayfield, Kleiner Perkins and others. But that’s not how the Valley funds companies. Obviously, it’s changed since then, but the funding used to be $2 million to $5 million series A and then you go to the syndicate model and get to the magic $10 million mark. Then some time before you go public, you get another $10 million. That way we were still $30 million short. Valley VCs are very focused on technology. In our case we wanted a balance between the people, the technology and the distribution. And investors don’t care about that stuff. Nobody wants to fund to build a sales force!

We met a guy named Bill Janeway at Warburg Pincus. At the time they were much more into M&A and buyouts. He was willing to commit the “pre-first round” of funding, and kick in once our seed money ran out to finish the idea. And he would have the first right of refusal to actually fund the business for $50 million. In January 1995 we quit Sun and officially founded the company.

At the time Bill Coleman and I shared an office in East Palo Alto, because it was the cheapest we could find. I was an executive at Sun with very high pay. We left all of this stock on the table. In those days you could still get an office with someone to answer the phones and pretend that you were big for $450 a month.

We knew the technology we wanted to sell. This was before the Internet and Java, but we saw that people lived very nomadically. So eventually laptops and home computers would be the way a lot of transactions happened, through home banking and so on. We wanted to enable the back end of that, in terms of distributed transaction processing.

I did the initial prototyping. The amount of time it was going to take to build the technology from scratch, mature it and then prove it to the world would be huge. We found a piece of code called “Tuxedo” that was invented in Bell Labs. It was designed for a distributed telephone switch. At the time it was owned by Novell, and Novell didn’t want to sell it.

We finally went to them and said, “You tell us what kind of revenue you can generate in the next three years with the technology and we can guarantee you those revenues.” Obviously, they came back with an astronomical figure. This was a technology that wouldn’t sell more than $3 million a year and they came back at us with a number like $50 million. But I guaranteed them that revenue, and on top of that they still had rights to the technology, on condition that within three years if they didn’t use the technology we would have a buyout right. We got the guarantee from Warburg Pincus and got the technology. We modernized it, and used the Tuxedo name because it was a mature technology.

That way we grew the company to $100 million from 1995 to 1998. In 1997 we went public, at the very bottom of the market. In 1998, clearly the Internet had taken off. We were starting to become much more of a niche technology and the momentum of the Internet was not catching on fast enough that people really wanted to do sophisticated transaction processing.

Somehow our technology wasn’t catching on fire. People perceived that it wasn’t naturally built for the Internet. It wasn’t the trendy thing any more. We went to Sun, and said we wanted to take Java to the enterprise. We prototyped a system.
We bought a small company called WebLogic, and created a subsidiary with 150 people. And that company essentially became the company that BEA is today. We gave away the software free on the front end to enterprise programmers. They adopted the software, went into production, and we sold licenses. That’s how we went to a billion-dollar company, from 800 employees to about 4,000.

What about the challenges of keeping the company on track as the market struggles?

The joke that I tell people is that you’re not going to see Citibank, Wells Fargo, and Bank of America saying, “You know we tried to go do the home banking thing, and distributed teller machines, but we were just kidding. We’re going to go back to serving you in very nice bank branches. We’ll buy some expensive Italian sofas, nice marble floors, and mahogany tables. And we’ll have beautiful men and women to type for you on tiny little green screens.” I just don’t think that’s going to happen. The train has left the station.

Enterprise software has always had a very poor distribution model. The consumption is so unpredictable that if you’re buying a large enterprise application (like Oracle, SAP or PeopleSoft) your anticipation is that you are going to roll that out to, say, 1,000 people at your company. Up front you’re buying a very expensive license. You don’t buy gradually because you want to know your total cost of ownership. Hopefully everything will get implemented and all the licenses will be fully consumed over the next two years. But it never happens that way and licenses end up sitting on the shelf. In a market like this it’s very hard to go back to a customer and sell when they have several million dollars worth of stuff sitting on a shelf to be implemented.

Because we have free downloads, we get consumption from the developers and we always grow incrementally, down to the individual CPU. An enterprise software company our size usually does 100-200 deals a quarter. We do more like 3,000-4,000. It’s like selling hamburgers! This makes us more resilient. We do do bigger deals, but those deals will be hard to do right now.

I’m challenged along with everybody else. This quarter in particular. For example, we won’t be selling anything to the airlines. I certainly don’t have the courage to go ask for an order. It is a challenging time.

I’ve seen it before in the Valley. Nevertheless, I have lived here for 23 years and I have never seen a bubble this big and one that burst so hard. It needs some time to heal, but I believe it will heal in 12 to 18 months.

How do you sell the company to employees in these tough times?

Stock options are a challenge. We did a reload on options three or four months ago when the stock had dipped from around $70 to $30. All of those are underwater again. So we’ll need to reload those for employees. It’s a very uncertain time. During easy times everybody can make things happen. In tough times, you have to re-assemble the team, and communicate a lot more.

What about selling the company to Wall Street?

The analysts have their own estimates because the market is going up and down and there are so many factors affecting the psychology of buying. We have more than a billion dollars in cash and 11,000 customers, so we’re not going out of business. I have to explain to the Street that this is a long-term play. You can’t just look at this quarter-to-quarter. It’s important to set expectations, but we have to preach very differently. si