My productivity took a huge jump when I did courses on human-computer interaction and information theory. I realised that the tools that the managers (note, not "the business", but the managers) wanted us to work with were a massive hindrance to productivity, and we ended up replacing them with faster lightweight tools that had been written by people who would have to use them day-to-day.
The most obvious change was binning ClearCase in favour of SubVersion and Perforce (depending on whether the business unit wanted a "proper support" license). Similarly for issue tracking and collaberation software - dumping the "big name" stuff in favour of smaller, lighter and (most importantly) friendly software almost always resulted in huge increases in productivity.
There were also a lot of free beers, as users came to us saying "Thank you for replacing that horrible system with one that works for me". To be fair, most of the replacements were done because we had someone (often me), who would sit with the end-user and see the way they wanted to work. It's not so much about the software itself, but it is about choosing software that is good for the people who are going to be using it, and that does usually mean stuff that "just works" instead of having a "big name".