In a post that is perfectly applicable to open source in its pithy conclusion, ReadWriteWeb suggests that vendors look beyond "viral" to ensure a product's success:
It's been said over and over again, but maybe one more time would help: "viral" can be encouraged, it can be prayed for, but it can't be engineered. Your only hope is to create engaging, compelling content, and tell a terrific story... and then hope.
Yes. Yes. Yes. In open source, the license is simply a start. An open-source license, however, is no guarantee of success. It's simply a description of what will happen to derivative works, but no guarantee that anyone will care about your project enough to actually create one.
To get more people to care, open source needs marketing. Lots of marketing. IBM brought marketing to Linux and made it a safe enterprise choice. JBoss' Marc Fleury created a guerrilla marketing campaign that first made the world notice JBoss, so that people would then notice the code behind its creator. In both examples, marketing mattered.
Of course, even marketing won't guarantee success. Witness Microsoft and its Vista problems. Microsoft's marketing blitz may improve Vista sales, but it probably won't be enough to turn Vista into a winner.
No, "viral" is a confluence of good licensing, good marketing, good product, and luck. This last one shouldn't be underestimated....