After reading Free by Chris Anderson I have some thoughts on how Red Hat Enterprise Linux can claim much of the unpaid Linux market share.
In his book, Free, Chris Anderson tells a story of a game developer who asked Slashdot, a technology forum, why pirates continued to steal his games when he charged just $20 for them. While many open source advocates and others might expect the responses to be filled with arguments against intellectual property the response was very different. Most of the individuals who commented back to the game developer said the piracy was occuring for two reasons. First, because the small amount of copy protection and difficulty navigating his web store provided a barrier to playing the game. Second, the individuals generally didn’t feel that the games being provided were worth $20. This meant the game developer had to create a better game or lower his prices and also provide an easier means of access to the game.
A similar situation exists in the Linux market today. Red Hat Enterprise Linux, a paid Linux distribution held roughly 62.2% of paid market share while CentOS, a free copy of Red Hat Enterprise Linux, and other free Red Hat based Linux distributions hold 50% of the unpaid market. In 2008, according to IDC, unpaid Linux subscriptions accounted for 1.2 million of the 2.8 million servers deployed. According to this data 42% of the server based Linux deployments used unpaid Linux in 2008. So why has this situation occured? It’s difficult to draw a conclusion because there are many different variables and situation, but one possible reason that an organization might run unpaid Linux is similar to why a pirate would decide to use a cracked version of a game. First, because they cannot gain access to Red Hat Enterprise Linux quickly. A system administrator may not feel it’s worth speaking with his manager, going through procurement, and then waiting for his/her Red Hat Network account to be created. These are all barriers to scratching their Linux itch, so to speak. Second, they don’t feel that $2500 is an approrpriate price for an operating system. Maybe they just want to serve up some web services to a LAN in their office.
So how can Red Hat remove these barriers to capture the unpaid Linux market?
First, remove the barrier to Red Hat Enterprise Linux and allow any users to download and use Red Hat Enterprise Linux. Certainly this would not entail providing support to consumers who do not pay for a subscription, however updates, erratta, and other information should be provided. More reasearch is required around the economics that would make this possible, but there is such a small marginal cost to providing updates to an unpaid audience that we could afford to give updates to all consumers. The Fedora project already performs similar functionality and the infrastructure could be leveraged to provide free updates to Red Hat Enterprise Linux as well.
Second, lower the price of supported Red Hat Enterprise Linux to a price that the unpaid Linux market finds more reasonable. Some revenue is better than no revenue. Provide pricing to markets that creates a psychology of responsibility and value. If a consumer is willing to spend a certain amount of money for a copy of Red Hat Enterprise Linux wouldn’t it serve our best interest to provide him/her Red Hat Enterprise Linux? When consumers in the unpaid market decide to move to a paid product it would be in Red Hat’s best interest tf the consumer was already running Red Hat Enterprise Linux. This would provide a clear upgrade path. It is important to note that the demand at free is much higher than even a penny and will cause a huge spike in the demand curve. Both of these theories need to be validated further.
The psychology of free may have also hampered mainstream adoption of Linux on the home users desktop. Could free be associated with lower quality to the average consumer? Does the fact that JBOSS costs 15 percent of the total cost of Websphere cause buyers to assume it is of lower quality? Would raising the price change consumer psychology and increase sales in certain market segments? This is another theory that should be explored.
Update
For some reason I’m not able to reply to comments on my own blog (I’m still figuring out docspace).
Per a new IDC report:
Nonpaid Linux server operating systems continues to increase in importance on the overall market dynamics, with nonpaid Linux server operating system deployments accounting for 43.3% of the worldwide total, up from 41.4% in 2007.
There is no debating that CentOS is an exact binary copy of Red Hat Enterprise Linux, but the psychology of using CentOS instead of RHEL might make a huge difference in the long run.
1. Many users might not realize that CentOS is a copy of RHEL.
2. When a user looks to upgrade to a paid version of Linux they might think a leap from CentOS to another distribution.
3. If a user had the option of downloading Red Hat Enterprise Linux for free they might run it rather than another distribution (excluding CentOS). This would make them more likely to begin paying for it later.
