A friend and I were talking about the relative valuations for open-source companies. Over on his blog, JBoss founder Marc Fleury had joked that, after seeing MySQL's valuation, he realized he clearly had "undersold."
But as it turns out, JBoss' valuation multiple was not much different from that of MySQL--which was snagged by Sun Microsystems this week for $1 billion.
The 451 Group is reporting MySQL's numbers as follows:
While MySQL has not disclosed revenue for 2007, it is believed that its revenue in 2006 was $34 million, up from $16 million in 2005 and $14 million in 2004. We estimate trailing 12-month (i.e., 2007) revenue to be about $48 million, and if we assume slightly slower growth in 2008, we get to about $65 million....Given that, it would mean Sun is paying 20.8 times trailing 12-month revenue and 15.4 times projected 2008 revenue.
I think this is roughly accurate. I've heard on good authority that MySQL's 2007 bookings were close to $60 million (not sure about revenue--I'll take The 451's word for it). That's 16.6 times 2007 bookings or 20.8 times trailing 12-month revenue. Not bad.
JBoss? 2005 revenue was $17 million and bookings were roughly $27 million. Red Hat acquired JBoss in 2006 for $350 million. This represents a multiple of 20.6 times trailing 12-month revenue or 13 times 2005 bookings. That's not a wide divergence from the valuation multiple MySQL got (but both are a lot better than the multiple that BEA Systems got for its stock from Oracle).
Zimbra? $6 million in bookings that fed into a $350 million valuation, though it's probably better to use the $20 million in bookings it was on track to hit in 2007 before the acquisition. At that number, it's a 17.5 multiple on bookings projecting to the end of 2007. More likely, it was a multiple on bookings in the range of 12 to 15. In other words, very close to the same multiple for MySQL and JBoss.
Given how slowly revenue accumulates in an open-source company--assuming it is recognizing subscriptions over 12 months--bookings is probably the valuation metric being used or at least strongly considered. It surely is the metric by which the start-up wishes to be measured. [See also this update which tracks forward-looking revenue as a realistic valuation metric.]
So while Savio suggests we open-source entrepreneurs may be "sleeping with dollar signs in (our) eyes," there's clearly a lot of work to do before most open-source companies are worth selling. It's not worth selling out for $100 million. Not for the venture-backed companies, anyway.
But for $500 million? OK. Twist my arm. :-)
UPDATE: Some have asked for the valuation multiple on XenSource. I purposefully didn't include it because it had zero relation to reality. XenSource had around $1 million in 12 months trailing revenue. It didn't have much more in bookings. So, at a purchase price of $500 million, that's roughly a 500x multiple. Few are going to see that....
What XenSource did have was a looming OEM win with a major software company and the VMware hype. The lesson from this (as well as JBoss and MySQL) is that it's easier to get a high valuation if you're in a hot market and/or are a category winner. Not everyone will be in this position, so including XenSource didn't seem to make much sense.