If you mean that it bottlenecks communication with it, yes, the mailbox can kind of be treated like a 'mutex'. Depending on how you are operating on the data and how it can be referenced then a pool of these managed by a supervisor (or library like poolboy) or operating on multiple messages at once inside the process or even raising it into ETS with a library for operating on the data within ETS can be significantly faster. However a good bottleneck can help in many cases as well (most I dare say) so it depends on the use-case as to whether you want it or not.
TL;DR: A process generally operates on one message at a time, but there are alternative methods.