Open Source Business Models - What Works and What Does Not?

by Murugan Pal

Many of you might have questions around successful business models. Recent Forbes article on SourceForge and the blog on SugarCRM's functionality differences between open source and pro versions raise interesting questions. If you are contemplating on a new open source venture - my recommendation is not to differentiate (functionally) between the commercial and open source versions, like in the case of MySQL's dual license.

From my conversations with CIOs, enteprise architects and IT developers there are three major reasons for adopting open source:
  1. No vendor lock-in and proprietary code

  2. Freedom to change or enhance - free in libre

  3. Cost effective - pay for tangible services and not for the software

The moment you introduce functional differences between open source and paid versions, you will violate all of the three reasons quoted above. Eventually, an open source free version may have similar functionality that will compete with the commercial versions.

Typically open source is leveraged by software vendors as follows:
  • Packaged and Bundled - Linux Distributions like RedHat or SUSE

  • Maintenance and Support - JBOSS

  • Tested and Integrated as a Stack - SpikeSource

  • Proprietary Products Certified with Open Source Products - Oracle on Linux

If you are a new vendor the odds are so low that your 'functionally incremental proprietary product' offered along with open source version is going to adopt mainstream adoption. This will bring up the debate "good enough vs. the best". For most of the customers a good enough product will suffice, and it is better to base your business model on operational excellence and nominal charges for covering tangible services.

Here is how I would summarize:
  1. Adopt appropriate licensing policy that suits you from the beginning and stick to it.

  2. It will be difficult for the commercial version to succeed provided there are incremental functionality differences between free and commercial offerings.

  3. Customers will pay for tangible services (support, updates, training and documentation) provided the product is usable and the subscriptions are nominal.

Have you ever thought about this issue? Let me know your opinions...


2005-09-06 23:35:47
Flawed assumption

I think there is a flaw in your assumption in comparing an embedded application (MySQL) with a web based system. SugarCRM learned this the hard way, when their product was forked and released as vTiger, causing them to modify their license. This fork was not due to a community feud, but solely for economic gain.

Dual licensing works extremely well, as you know with the SpikeSource model, where anybody wanting to prevent the GPL viral effect, have to pay for a commercial license.

2005-09-07 03:23:39