Open-source enterprise applications can be sold just as profitably as proprietary software
Larry Augustin of Medsphere, the open-source electronic health records company, gave a talk last week at the SDForum SIG on open source. It was a great talk, especially for the unfortunate few of us whose entire creative energy revolves around enterprise software. There was lots of audience interaction, and we barely got out of there by 9pm. The thesis of Larry's argument is that open source based companies can operate with profit margins similar to those at typical enterprise software companies that rely on license revenue. And the time is right for open source products at the application layer, which is significant because applications are why customers invest in the whole technology stack in the first place.
There was a general acknowledgment that in most areas open source applications are not challenging existing enterprise software companies yet. But as more open source applications come out with wider distribution and lower cost, they will put pressure on the more expensive software providers from below. It was noted that a few application-level open source companies are already doing quite well, including Compiere, SugarCRM, Asterisk, and Medsphere.
Expanding the open source area, even when it means reducing license revenues, actually expands the total size of the enterprise software market due to the larger number of businesses that can afford lower-cost enterprise software systems.
Enterprise software vendors are like vacuum-tube radio providers, in an era where the transistor radios represented by open source are still of a lower quality. Eventually, transistor radio technology moved upstream in quality, redefining the market. We know how that turned out. Pretty exciting stuff.
Some highlights of the talk:
- The cost of sales at a typical enterprise software company (he used Siebel as an example) equals about 75% of total license revenues
- Therefore, companies that buy enterprise software are essentially paying for the sales cycle
- A successful open source project that has a high volume of downloads, such as SugarCRM, has greatly reduced cost of sales, approaching zero in some models
- Enterprise software companies have very high general & administrative costs, due to their wasteful operations and big profit margins
- Research and development costs are typically lower in open source companies
- Larry did some magic with the numbers, removing license revenue, and showed how the gross margin % remained the same without license revenues
Some critical comments:
- One of the key questions in this assumption is whether the sales cost could be kept down as larger deals are pursued. Large deals would be critical for legitimate enterprise software companies, open source or not. Assuming that an open source product will have a huge distribution and no sales cost is not an accurate expectation. Companies such as RedHat that sell to enterprises (although not in the application space) have high sales costs, similar to standard enterprise companies
- In enterprise software sales, in my experience, price is rarely the primary decision factor. It almost always comes down the features, or the ability of the vendor to sell the features. Only when an open source application can compete head to head with a commercial application and win on merit will there be significant traction. It’s a fair prediction that there will be more and more applications that can do that. JBoss can beat WebLogic now, and JBoss Inc. has an extremely low sales cost. That is the ultimate model, but it takes great software.