Most shops I've worked in use squash and merge (or rebase) always. This prevents unnecessary repo bloat as the number of devs, code base and number of commits gets large and removes all those "Typo".... "Typo again".... "Sigh. Working now (I think)" type of commits that really don't add anything meaningful to the context.

With squash and merge you get chunky, meaningful commits that (with discipline) are tagged to a specific issue/feature so it's really easy to understand the history.