Your performance may or may not improve in this instance. If you are running the VM locally than you are sharing resources with the host. If the VM is remote, than you gain a bit by offloading the client bits (i.e. browser) and it's resources.
In both instances, if the VM is taxed on the server side, than you wont see much improvement regardless of the where you initiate the browser session.
"Best Practice" is to try both, and see what works for you.
That being said you can look at few things on the VM to gauge server side performance.
The obligatory: CPU / Memory Utilization / Disk IO.
Clearly if your CPU is running or Mem. is running high with no load, than you may not see an increase regardless of where you are running the browser. However, if it is high but not critical, you may see benefits running browser in an outside context, but you may just be able to bump up the VM resources as well.
If Disk I/O is the issue, and you have latency access SQL, than it won't matter where you run your browser as that is not the bottleneck.
IDERA has a free monitoring tool, that will tell average page load times. They have on for SQL as well. They both installed on a client side and access the servers (provided your VM's allow for it).