You are correct. Whether you use backup or replication, the VM would return to the point in time of the backup / replication. The difference is that with replication a VM is in "standby" mode so when you do a failover, the amount of time to get the VM back online is the time necessary to start that VM. In case of backup, there are two options. You can recover the entire VM and that will take more time, or you can use instant recovery which will be the time to start the VM that lives on the backup storage. However, there will be a bit of performance loss here as most likely your backup storage will be not as fast as your production storage. And after doing an instant recovery, you still need to migrate (in the backend, while online) the VM back to the production storage.
Hope it helps