The mistake here is believing in efficient market theory.

It's not so easy for a competitor to show up, because they have to work against heavy first-mover advantage and network effects in software. Those who do a crap job get first to the market and set the trend (and the expectations). Moreover, there's a heavy component of tragedy of the commons here, with commons being users' computing resources - the software is often designed with implicit belief that it's the only thing running on user's machine. It's enough to make a sale, and you don't get a reward for making it so that your users can run many other stuff simultaneously with your application.