It's a classic switching cost problem, to make it worth switching the effort either needs be below a threshold or the outcome has to be much better otherwise people won't switch - i.e. a killer feature/features.