When it comes to using a product, people rely on mental models – for better or for worse.
The mind draws on prior experience to create expectations which guide behaviour. We did this last time and it worked, so that’s what we’re doing this time, too. Mental models reduce cognitive overhead because you don’t have to make a series of decisions more than once – just retrieve a successful outcome from memory, and use the same template.
The base example of a mental model is one that every web designer learns to respect from day one:
Put the logo in the top left, and make sure it links to the homepage.
People expect this, and if they don’t get it then certain things might happen. They might become irritated at your brand, with its stupid logo sitting up there all non-linky and such, which is bad. At the very least they’ll wonder if your web designer is incompetent – not much better.
But just because a mental model exists, doesn’t make it mandatory. You can conform your UX to users’ existing (presumed) models, but you can also teach them new ones – if you have the courage. When Blockbuster said people come to the video store to rent a movie, it’s all part of the experience, nobody disagreed. But Netflix started mailing discs instead, and almost overnight the entire movie-watching population completely changed its behaviour.
Challenging mental models comes with risk, but the rewards can be considerable. When fresh thinking yields innovation, entire markets can shift fast.
Personally I wouldn’t recommend you move your logo link from the top left, though.