How We Reached The Open Source Business Decision

by Paul Everitt, Digital Creations

The scene: it is the first week of November, 1998. Digital Creations has just closed an investment with Verticality Investment Group for first round funding. Hadar Pedhazur, a principal in Verticality, whom we've gotten to know quite well over the last year, is talking with us about laying out our roadmap.

At the same time, I'm working on getting our message in gear. The focus of the investment is to legitimize our jump from a product company to a consulting company, where we do consulting using our software. I am sorting through our software story in light of this. Here's what I saw.

We are a small but growing company with two strong offerings: Bobo, a free, opensource toolkit for web-object applications, and Principia, a commercial web application platform. Principia used code and ideas from Bobo. But as I look at it, I see there is an unbridgeable gap.

It is too hard to have two separate messages to articulate, two different groups of customers to nurture, two documentation efforts, two engineering efforts, etc. While Bobo is popular, we were focusing the company on Principia. It made sense to find ways to reduce duplication of effort and message.

Starting the week before the Python Conference we formulate the idea of consolidating Bobo and Principia into something new. The stuff in Principia would just be a free (but closed source) module that plugged into a more common platform. I then exchange emails with everyone on the Bobo list that had a post count higher than 20 to solicit ideas.

Well, that just didn't make any sense at all and didn't bridge the gap. It frankly was confusing. Jim Fulton here floated the idea, "Why don't we just open source Principia?" That was on Thursday, five days before the conference. We all kind of laughed and got back to the important stuff -- bickering about the intellectual property value of the tree tag :^)

Suddenly an unexpected force emerged: Hadar from Verticality, the guy that had just laid down a stack of the little green things. He coached us through the decisions about what is the real value of the company and how to best deliver on that value. Thus, the question was, "Can going open source increase the value of our company?" Here's what we saw:

• Going open source will increase our user base by a factor of 100 within three months. Wider brand and stronger identity leads to more consulting and increased valuation on our company.
• Open source gives rock solid, battle-tested, bulletproof software on more platforms and with more capabilities than closed source, thus increasing the value of our consulting.
• Fostering a community creates an army of messengers, which is pretty effective marketing.
• This is not the last innovation we'll make.
• In the status quo, the value of packaging the software as a product would approach zero, as we had zero market penetration. What is the value of a killer product with few users? The cost to enter the established web application server market was going to be prohibitive.
• The investment grows us into a larger, more profitable company, one that can make a credible push to create a platform via open source. Since our consulting is only on the platform, a strong platform is imperative.
• Open source makes the value of our ideas more apparent, thus the perceived value of the company is apparent.
• Our architecture is "safer" for consulting customers. With thousands of people using it, the software is far less marginal. The customer is able to fix things themselves or reasonably find someone to do it for them. Finally, the software will "exist forever".
• Dramatically increasing the base of users and sites using it gives us a tremendous boost in "legitimacy".
• The exit plan isn't about the golden eggs (the intellectual property) laid last year. It is about the golden goose and tomorrow's golden eggs. The shelf life of eggs these days is shrinking dramatically, and the value of an egg that no one knows about is tiny. Give the eggs away as a testament to the value of the goose and a prediction of eggs to come.
• The community can work with us to dramatically increase the pace of innovation and responsiveness to new technical trends, such as XML and WebDAV.
• Ride the coattails of the nascent Open Source community and its established channels such as RedHat. OSS has a certain buzz that is greater than its real customer-closing value, but this buzz is getting hot. Moving aggressively towards Open Source can make us a category killer for the web application server market segment.
• We believe like hell in what we're doing. Others believe in us as well. We should follow our instincts.

Well, all this decision-making and internal haggling happened on Tuesday while I was on the way to the airport. Gulp! Wednesday, the first day of the conference, we were still coming to grips with giving away the intellectual property we had labored over for two years. It became crystal clear that it was a winning business strategy, oh, around 2:00 Wednesday, the first day of the conference. Me in Houston, Rob Page and Jim Fulton in Fredericksburg, Hadar in New York, all trying to scale the mountain.

I was unfortunately in a jam. There was a feeling that I was sitting on a golden opportunity. Eric Raymond gave the keynote and, during the Q&A, I asked him questions about IP, exit plans, open source, and the investment community. This sparked a lively discussion. Afterwards a Washington Post writer who was following Eric around for a few weeks interviewed me.

Point is that I wanted to make the announcement and let Eric know that a credible web application server was moving over to the open source community. Also, I had to get Eric to sign off on the three strings we are attaching: gotta have a "Powered by whatever-the-hell-we-call-it" button, gotta have a credits page, and gotta tell us what you are doing with it. He said the third one went right up to the line, but as long as we respected projects with legitimate confidentiality issues, it sounded OSD-ish to him.

With that mission accomplished, I wound up making the announcement at the end of the demo session. Still having my Witty Mouth installed, I asked Jim Hugunin if I could make an announcement and said something like, "Just wanted to let everybody know that we are making Principia free and open source, and positioning it as the open source alternative to Cold Fusion and other application servers." I was surprised when everybody applauded -- I didn't think that Bobo and Principia had that much of a name awareness.

Afterwards there was a large number of people that came over and said that they wanted to get their stuff onto the platform. Definitely good news as making a bigger tent is a high priority.

Now the problems began. Sean Summers came up and told me that he had just gotten off the phone with Liz Coolbaugh, co-editor of the Linux Weekly News and that it was going to be in the next edition the next day. Yikes, our website was completely blank regarding any of this!

I got back from the conference, announced it on the Bobo list, and watched the traffic explode. I had to stall while we came up with a name for the unified thing. This consumed around 60 hours, combing through Mayan dictionaries and Encarta looking for obscure terms. No matter what I found, it was already taken. Michael Hauser suggested the Zope name. It survived the test. Viva la Zope!

The Wednesday before Thanksgiving we had a board meeting and elected Hadar as board chairman. We also had a robust conversation about where we would define the "line" between the intellectual property we held close and that which we threw over the fence and into Zope. Well, from that conversation we decided to throw Aqueduct, our relational database integration software, over the fence as opensource and free for non-commercial use. A few days later we decided to drop the non-commercial use part and just make it free, allowing us to legitimately be a "free, opensource alternative to Cold Fusion."

We also decided to look hard at the conditions, aka strings, and make sure they were effectively accomplishing our business goals. I had gotten some useful feedback from a few people in the community. We decided that making the registration system a requirement was bound to fail and showed bad intent, so just encourage registration instead. We also decided that we should allow flexibility in the "button" idea and make sure people have different ways to accomplish our goal of promoting the brand.

So it is now two days before the launch of the source. Lots to do, but this decision is sitting well with us. It feels like the one milligram of catalyst that you drop into the solution and, well, watch the whole chem lab explode. Err, bad analogy. Watch the room-temperature fusion reaction begin...much better analogy.