May 11, 2007

Progeny's closure highlights problems of small FOSS companies

Author: Bruce Byfield

Founded by Ian Murdock and John H. Hartman in 2000, Progeny Linux Systems seemed for years like a modestly successful free and open source software (FOSS) company. Although it abandoned its original plans for revolutionizing networks, it survived the dot-com crash while many other companies had gone under, and its efforts to commercialize Debian were profitable after its first year and a half. By early 2005, Progeny had reinvented itself by offering update services and modular components for building specialized GNU/Linux distributions. Yet on April 30, Progeny ceased operations. What happened?

The reasons for the closure provide a case study of the problems that small FOSS companies face. They include difficulties in scaling, an inability to compete with large companies, and a lack of the funding needed to develop.

Branden Robinson, former Debian Project Leader and a Progeny employee from the company's start, makes clear that Progeny's failure was not due to lack of business. According to Robinson, when the company closed, it had half a dozen clients, and was in negotiations with at least one other company. In addition, Progeny was still providing update services for three or four clients. "In some ways, I feel like we had no competitors," Robinson says. One company, he adds, "was really disappointed to hear that Progeny was going out of business, because they weren't really sure who to turn to."

Too big and too small

One problem, Robinson says, may have been that the company was an awkward size for the marketplace. To survive, let alone grow, the company could not afford to take on the small contracts that might have sustained a sole proprietorship. Nor was taking on numerous small contracts a solution, since, as Robinson observes, "You can only distribute an engineer over so many projects before you start losing his efficiency."

Yet, at the same time, with the five engineers who remained until the end, the company had trouble fulfilling larger contracts. In the company's final weeks, Robinson says, "We pretty much had enough people that if we worked really, really hard, we could fulfill the contractual obligations that we had, but if we were to lose just one engineer, we'd be facing deadline slippage and other problems like that. It would have just sent us down the spiral in two or three months." As things were, in the last month, Progeny's director of engineering was doing some work for free, and the company was operating without a CEO.

Competing in the big pond

Varying in size from five to 30 employees, Progeny was also in the awkward position of trying to compete with companies with hundred of employees. Bern Galvin, a member of the board of directors, suggests that competing was especially difficult because "there ain't no IP" [intellectual property] -- meaning that, as a FOSS company, Progeny had no trade secrets to level the playing field. Murdock believes that the company's expertise gave it a similar advantage, but he concedes that the traditional proprietary business models can be easier for a small company looking for an advantage over larger ones.

Another problem, Galvin suggests, is that larger companies operating in the FOSS market are no longer interested in acquiring smaller rivals, the way that Novell, for instance, purchased Ximian. Now, he suggests, they prefer "picking off key employees." He cites Murdock's own departure for the Free Standards Group in 2006 as an example of this tendency, especially since he was soon followed by Jeff Licquia, another of Progeny's original programmers.

According to Murdock, his departure was due to the fact that "I just realized that we had missed the window and it's time to move on." That may be so, but it is hard to imagine that the founder and chair's departure would not make the company's demise a self-fulfilling prophecy, or that the loss of a lead developer could do anything but stretch the already overstretched employees even further.

More directly, with the rise of Canonical and Ubuntu, a large and well-funded rival also working to commercialize Debian, Progeny "has been losing its relevance over the last couple of years," Murdock says. "We tried competing head to head for a little while," but "they executed better than we did. Some of that was that they had more resources than we did, but you can't explain it all that way. And, at some point, we had lost whatever momentum we had."

Lack of capital

However, all those connected with Progeny agree that the largest problem for Progeny was a lack of capital. After exhausting its initial funding in mid-2001, the company was unable to secure additional funding despite repeated efforts. Instead, the company focused its energies on finding the contracts it needed to meet its payroll.

In the short term, this tactic guaranteed the company's survival and even encouraged modest growth. However, as Murdock says, "The problem was, when you are bootstrapping, your first priority is to do the work that pays the bills. You tend to have precious little left over to invest in R & D." The company attempted to incorporate its own development work into its custom contracts, but that tactic was not always possible.

"Progeny was in a business that required a lot of capital to make it work," Murdock says. "We wanted to sell Debian as the preeminent platform for building Linux-based appliances, and that's a big chunk to bite off. If you don't have the capital you need, no matter how good your ideas are, you're simply not going to get there."

The final descent

Progeny seems to have reached its zenith in mid-2005. However, a round of layoffs occurred at the end of 2005, and Murdock became disillusioned and moved on soon afterward. "The company kept going," Murdock says, "but it's just been slowly shrinking since that time. At some point, the remaining people realized it's never going to be any different, and that it was time to move on."

At the end of March 2007, CEO Garth Dickey and other management staff reached the same realization as Murdock and moved on. When contacted, Dickey preferred not to comment, saying that to do so would be inappropriate, but both Murdock and Robinson confirm this sequence of events. For the last month, "the engineering team was having to carry on without management," Robinson says.

The company's board of directors approached one potential CEO with an offer, but it was turned down. At any rate, Robinson says, "Had we hired a CEO, we might have been in the position of laying off another engineer." Faced with such an unsolvable situation, the company had little choice except to close shop.

Despite what must have been an often frustrating experience, those contacted regard their time at Progeny as well-spent. "I was there seven years," Robinson says, "and for practically all that time, it was a fun and really rewarding place to work."

Similarly, when asked if he would consider starting another company, Murdock immediately replied, "Absolutely."

However, in thinking about what he learned from Progeny, Murdock offers two pices of advice for others starting their own FOSS business. First, thinking about the company's sometimes unfocused approached in its first year, he suggests, "Before you take money, make sure you know why you are taking it" -- in other words, have a clear set of goals.

The widespread acceptance of FOSS and the rise of online technologies mean that many companies today do not need venture capital, Murdock says. However, his second piece of advice is, "Once you've decided to take outside capital, make sure that you take it from the right people. One of the problems we had in raising the second round is that the people who initially funded us didn't understand what we were doing. They were basically just throwing money out at a time when that seemed to be a strategy that worked. Having the right backers means that you get more than just money from them. They would have been motivated to help us over the hump and get us to where we needed to be successful. I think that's something that's been lost on a lot of people."

Bruce Byfield is a computer journalist who writes regularly for NewsForge,, and IT Manager's Journal. He was a Progeny employee in 2000-2001.


  • Business