Both of these companies have very different product architectures and business strategies than the incumbents in their markets. For example, QlikView operates completely in memory without any need to create OLAP cubes. Maybe this is a lucky accident sprung from naivete, or maybe it's a brilliant idea that came from having a fresh perspective. But it's notable that the view came from outside the traditional approach to BI.
Similarly, there have been lots of teams that have built storage engines for MySQL, but perhaps the one that could be the most interesting comes from a Y Combinator startup called ReThinkDB. I'm pretty sure none of these guys were even born when relational technology was invented, so they aren't likely to be too set in their ways about how a database should work. Instead they're focused on how to optimize database performance for emerging SSD-based systems.
Maybe the longer your exposure to the traditional ways of building products and businesses, the less disruptive you will be. So here's to naive startups that can afford to be radical in their architecture.