The entire concept of BC is based on mathematical "models" that are themselves based on real-world observations...in other words...a long time ago somebody threw a bunch of various shaped rocks downrange and took notes on what happened to them.
All of the current ballistic computers are based on those notes, and the degree of exactitude expressed therein should not be taken literally, as it is not an exact science. At best, a model represents a scientific attempt towards a "best guess".
The testing performed to arive at a BC number is very rudimentary, as the equipment necessary to measure BC in a more finite way is generally well beyond the resources of the small arms community. Even if that advanced analysis were to be undertaken, the figures still get plugged back into some mathematical model or another...and then we are right back to making guesses.
Now, some of these guesses can turn out to be pretty close, and often are, which is lucky for us, the shooter, because usually all we are really interested in is how many elevation/windage clicks are necessary to get on target... and that is fortunate.
The best way to figure all that out is to do some range testing and try to find out where the bullets are landing most of the time...just like those rock throwers of long ago. The range method really works, and once you plug some real world drop data back into a ballistic computer, you should be good to go for other environments as well, but only for that particular rifle...that load...and that bullet.
That is because factors related to the individual rifle, like barrel harmonics, site height, and a few others, will also effect drop data, and the resultant BC number.
So in the process of doing your own real-world ballistic testing you can use reverse logic to come up with a BC number that gives drop performance that is close to what the programs predict for the bullet you are using.
If the program you use is pretty close to your real world results, then that is the best BC number to use, regardless of what the manufacturers have published...a bit more testing will determine that. That does not necessarily mean that the BC figure you come up with is will be the correct one for your buddy using a different rifle rig.
Basically, the manufacturers BC rating is really only there for comparison purposes, and just like the EPA mileage ratings manufacturers assign to cars...your mileage may vary!