In the late 1990s, while working as a sales engineer at Sun Microsystems, I decided that I wanted to do something more technical, more on the programming side of things. I was interested in developing an application server, a program that manages the applications for an enterprise, using Java, a Sun programming language with which I was familiar. “Open Source,” whereby a developer licenses the code to users at no cost, was just coming into its own at the time. Users could run the program, and access the source code or even modify it, at no charge.

Given the business climate at that time, I felt that there was a good likelihood of finding venture capitalists to fund the project. In other words, once I firmed up the product, I could stay focused on the product while the suits could figure out how to make money and build a real company around it.

The prospect of working for myself was also appealing. I had a friend who felt the same way, so together we founded Telkel, Inc. in the fall of 1999. I was the chief technology officer, while my friend, who had a financial background, was chief executive officer, in charge of pitching our product and company to potential investors.

Adopting the Business Model Du Jour

My partner and I developed a business model, which consisted of building an application hosting business on top of JBoss, our Java application server. Hosting applications, or running programs on our server for a fee, would provide the revenue for our company.

Two problems quickly became apparent when we approached VCs with our plan. The first was that we had no credibility in the application hosting business. We had strayed from our core competency, which was developing the technology, and jumped on the application hosting bandwagon because it was the only business model we could come up with for making money on it.

The second problem was that our product reached its first stage of maturity in the spring of 2000. Up to this point, we had funded the development ourselves so that we would have a viable product with which to approach investors. At the very moment when people might have been interested in funding our product, the heyday was over, venture capital funding had dried up, and open source had become the pariah of high tech investment.

At this point, we had a great product but no revenues. Since the product is free, we had users but hadn’t figured out what we could sell them. Meanwhile, we were carrying significant overhead for salaries, office space, and equipment. The money ran out, and the company folded in November 2000. My partner and I parted ways.

You can probably see by now that when it comes to starting a company, the above is a recipe for “doing it wrong.” Fortunately, from my experience to date with the launch of my second company, JBoss Group LLC, in August 2001, I can also tell you how to do it right.

Keep the Best; Toss the Rest

I had a very strong belief in JBoss, the application server that was the core product, and, unlike the money people, I believed in open source. I had also really enjoyed being my own boss. In short, I just wasn’t ready to give up yet on a company of my own. My first company had failed, but I had still awakened the next day. If I gave it another shot, I felt fairly confident that I would still be able to find a salaried job in the tech industry if the entrepreneurial experience didn’t pan out.

Thus, in approaching my second venture, I took a good, hard look at what went wrong with my first, and segregated those factors that were within my power to change. The product wasn’t the problem. The macroeconomic environment was wrong – VC money had evaporated almost overnight. Tech companies had fallen out of favor. I couldn’t change either of these things, but there were some things that I could change. Chief among them were my own understanding about what running a business is all about and a willingness to adapt my business model to the post-1990s reality.

I decided that the company had to turn a profit from the start. I had to keep revenues ahead of expenses at all times. I would not seek VC money this time – the environment wouldn’t allow it. Since I had already lost my savings, I had to make money on my venture from the get-go. I had no choice.

When Opportunity Knocks, Open the Door

While I was analyzing my mistakes and missteps, a funny thing happened. Users of the JBoss application server began contacting me through the Web site and user forums. They wanted to know if onsite training was available for JBoss. Being a firm believer in giving customers what they want, I told them it was. Then I scurried to write a training class. One company commissioned me to deliver a training course, which I then presented to that company and to others. When I began to get inquiries for open-enrollment training, I scheduled a course in Atlanta. At $3,000 per trainee, the first course quickly sold out, grossing me $60,000 for one week’s work.

Listening to the customer and the market seems like Business 101, but the allure of the dot-com boom caused this to be less apparent the first time around. And, frankly, it had to hit me over the head the second time around. In retrospect, it seems like a no-brainer, but it can be pretty tough to hear opportunity knocking at the door when you’re looking out the window.

Find Good Collaborators

As training took off, I began to get requests for support, consulting and documentation as well. For the first half of 2001, I worked as an independent consultant, but it soon became apparent that I would need to find people to work with me if I wanted to exploit the opportunity I had. In August of 2001, I formed JBoss Group LLC, in order to provide training, support and consulting services by the programmers who were most familiar with the product.

In order to accommodate the needs of the very best programmers, I developed an employment model where developers work half time developing the free, open source product and half time providing associated services. JBoss Group paid them for the time they spent doing consulting, training and support, but at a consulting rate. It took some time for business to ramp up, but, under this scenario, the money they earn as part-time consultants has come to equal and often exceed what they could earn as full-time employees. They spend the rest of their time programming in open source, which gives them a lot of creative freedom and experience. They can improve their skills working on a project they enjoy, while still making a comfortable living.

We’ve since moved away from this model a little, as the success of the company now allows us to offer programmers a minimum salary guarantee, with an upside based on the amount of consulting work they do. This method provides customers with access to people with the highest level of product knowledge. Everybody wins, yet developing this strategy also required some ‘thinking outside the box’ to come up with a scenario that made sense.

If At First You Don’t Succeed….

JBoss Group now employs 30 people, and competes with companies like IBM and BEA Systems in the Java application server industry. The company is still self-funded, and revenue has doubled every nine months since our founding. The company has been operating profitably since day one.

I found that working from the bottom up – building the product, going after the business – is better than working from the top down – looking for the money first to fund a raw idea. For fledgling entrepreneurs, the lesson is clear: focus on your core competencies, and do the work yourself – don’t count on someone else. And keep your overhead under control.

Very early on, I knew I had a product that people wanted. I just had to figure out a way to make money from it. After a false start, I was able to do that by believing in the product, listening to the customer, and finding the right people to work with me. The money followed pretty quickly from that point. Since then, the challenge to building a self-funded company has been to manage growth in stages so that revenues stay ahead of overhead.