Problem #1 would be state of battery charge at the beginning of test. I find bars on navigation display highly inaccurate, sometimes it shows certain number of bars and it stays that way for long time then it drops 2 bars within seconds.
then you have a problem with power demand, even very small change in pedal pressure can increase/decrease amperage used by large amount, in other words the best you could do would be some very rough estimate. Driving first down some longer hill could assure fully charged battery, letting the car roll by itself without pedal pressure from dead stop could eliminate variability due to power demand and throttle pedal angle, but finding place where such test could be done safely would be more difficult I think. But this is very good idea to get base battery charge and compare that later after few years of driving. I need to come up with some "test track" of my own and see if I can do some testing based on temperature and get that base capacity to compare later. The whole test would be much more accurate if we had some way of reading amperage used, that way battery capacity could be calculated and compared to factory spec.