Jan 04 2012

Choosing a business model and sticking to it until further notice

Whenever you talk about a company, you often begin by discussing its "business model", rapidly dismissing the product itself. This is particularly true when the product is initially free, when you find the idea a little weak or the market is already crowded. The question the founder is faced with is "how do you plan to make money ?".

For XWiki, there was the obvious precedent of existing Open Source companies which seemed to make things simpler, but that wasn't enough. When I created XWiki, there were quite a lot of choices to pick from:

  • consumer oriented wikis on the cloud
  • a software as a service offering
  • making the solution proprietary
  • an open source business model based on services
  • an open source business model with dual licences
  • an open source business model a la "redhat"
  • an open core business model
  • any combination of the above

Initially I released XWiki under the GPL licence, but I quickly realized I would have many difficulties with this model, as it would make receiving contributions and allowing more embedding of XWiki too complex. Also given wiki pages could contain code, too much complexity would have been created around the licence issue. I dropped the consumer play very early. Although there were no public Wiki farms at the time, I found that the XWiki product was not meant for consumer Wikis and that the advertisement model would require a lot of page views and therefore a big platform. I moved towards an Enterprise target which matched what the product did better. I also moved to LGPL and a business model based on services, which seemed to me a nice balance between protecting our rights and setting our obligation towards our users and contributors. Basically users and contributors can do what they want, including adding proprietary code, but if they want to redistribute the modified software, then they need to redistribute the modification performed to XWiki itself as LGPL (this doesn't apply to their own modules as long as they don't contain LGPL code).

I had long discussions with many people about Dual Licenses, about the Open Core and we continue having these discussions (more on open source licenses in another post). In my view the most important thing is to stick to your licence and to your business model to avoid confusions. You'll have people that like your license, while others will not. You can't please everyone, so the best you can do is to avoid pissing everybody off. 

An example has been the case of extjs which moved from "close to LGPL" to "GPL Dual Licencing", which created a lot of confusion around extjs with many people feeling "robbed" of their time and investment in the community. While the developer has the freedom to choose his model and to find a way that works for his business model, it is plain wrong to experiment with this. If you really need to do it because you have no other business model, you should at least apologize for letting people down (longer story here)

As for the Software as a Service model I did not fully drop it as we did hosting very early, but I realized we would have to wait before implementing SAAS. While the model is very interesting on a longer term, the price point per client is very low (advertisement or low initial fees). Unfortunately as you start a business the cost of a client is pretty high, while on a longer term your cost per client drops very quickly. This makes the consumer and software as a service business model quite capital intensive in the beginning, while the enterprise business model with services is much less demanding from this point of view.

This is where your objectives become important. If your are looking to build a startup, want to have a quick growth and are ready or very interested to sell out quickly, then choose the business models that pleases the VCs and go for VC backing as soon as possible. You will need to be aware that you'll most probably lose your freedom as an entrepreneur and become stuck in the "way things work" for startups.

On the other hand, if like XWiki, you want to remain in control of your future, then you have to pass on VCs and capital intensity. The good news is that it's possible. Service oriented business models or even cloud (to a lesser extend, if your product picks up fast) can be low capital intensive.

Many companies tend to start with the low capital intensive business model and then move progressively to the more intensive one, once they have sufficient traction and they can get capital from VCs. Sometimes this was their plan to begin with, while other times they are pressured by the investors. This is one of the things I was never ready to propose to investors when presenting XWiki. This part is quite disappointing if you view it from the "community side". Any company builds a community around the promise its future holds. When your business model changes you risk betraying your community. You have to be very careful from the ethics point of view, as well as the business point of view.

This problem is not new in the tech industry. Customers need to be more and more careful of the lock-in issue and more informed about the provider's motivation, as well as the sustainability of its business model. More on that in another post.

For XWiki SAS we chose the Open Source business model based on services, complemented with an Open-Cloud Software as a Service offering, and we plan to stick to it.

Ludovic Dubost
XWiki Founder and XWiki SAS CEO