Through my work I have tested many greases under demanding industrial conditions, and my experience has been that many (most) greases are still good after there has been some separation. Mix the oil back in and apply as needed. I expect there is a point where the oil separation becomes a problem - like when the substance left behind under the oil is dry and cracked, but we never tested grease in that condition. But a little bit of separation can happen to even the best performaing greases after a relatively short time, and the performance of the grease does not suffer.

For example... we have a particular grease often used for greasing the threads on rising stem valves that has always proved consitent in testing. When we were concerned that our instruments were not functioning properly we would use that grease to conduct baseline tests and we could tell by the measured coefficient of friction between the valve stem and nut if the instruments were reading properly (this is when everything is in calibration, but we just want to convince ourselves that an outlyer test sample really was behaiving strangely and we are not crazy). We would go back to the cupboard where the sample greases were kept, took out our favourite, mixed in the separated oil and ran the test, and it performed fine.