The G7 reference projectile is a boattail shape with 10 caliber radius tangent ogive. This shape is simply closer to the bullet shapes most people use these days (though the G5 is closer for `68 grain SMK") than the old G1 reference projectile (flat base and paltry 2 caliber tangent ogive). Modern flat base bullets are actually closer to the G8 projectile which has a flat base with a 10 caliber tangent ogive. The old G1 is probably best for flat base cast bullets with modest tip radius and for round nose bullets.
The whole idea behind a reference projectile originally was to be able to fire thousands of rounds of this one bullet and document its velocity loss and time of flight over each successive yard until you have a very complete table of its performance. Then, with any other projectile the exact same shape, the difference in how quickly it will lose velocity compared to how quickly the reference projectile lost velocity is just the ratio of the sectional densities of the two. So you can scale the trajectory of the reference projectile at the same starting velocity down to the performance of your actual bullet. In turn, this can be listed in tables. Reference projectiles are standardized at 1 lb weight with one-inch diameter, making the sectional density equal to 1 as calculated in ballistics (lbs weight/diameter in inches squared; not as in physics or aerodynamics where you would divide pounds weight by the actual cross-sectional area of a shape). So, if your bullet matches the shape of the reference projectile, and it has a sectional density of 0.5, then 0.5 will be its BC with respect to that particular reference projectile, and it will lose velocity as 1/BC, or twice as fast as the reference projectile.
Where this all gets messy is when the shapes don't match. I might have a sleek 30 caliber bullet that weighs just 1/40th of a pound but that has a BC of 0.5 because it slows only half as fast as the G1 projectile due to its more aerodynamic shape. The ratio of the sectional densities doesn't work because of the shape difference, so that error is adjusted by a form factor. The form factor is a multiplier of the square of the diameter that makes the adjustment.
The problem with that approach has always been that when the shapes are too different, the ratio of the drag coefficients of the two isn't the same at different velocities. So then you need a series of BC velocity adjustments to correct it. [/b]However, with a modern boattail bullet using the G7 BC, you can usually use just the one number and not have to make the bracketed velocities and BC tables.[/b] That's the whole deal: how you can make the ballistic information for your bullet as compact and simple as possible, and that's what the G7 does where it is available.
All that said, things are changing. Hornady and Lapua now have Doppler RADAR sets strong enough to follow a bullet for 300 yards. They can fire them at different velocities and come up with the individual bullet's drag function. With that, rather than having to compare the bullet to a reference projectile (a 19th-century workaround to having to fire thousands of every bullet made to know their actual trajectories), modern ballistics programs can work directly with the actual drag function of the bullet shape. This is intrinsically the most accurate approach. Hornady has been building a library of drag functions for use with their online 4 DOF ballistics calculator. The free JBM calculators can use individual drag functions when they are available. I don't know if Hornady is publishing theirs or not; I'll have to ask, but their calculator is fine for their bullets, so you have access either way. The QuickTARGET Unlimited program that comes with QuickLOAD has all the Lapua bullet drag functions and some from the BRL, such as the 168-grain MatchKing and the M1 Type bullet used in military match ammunition for a long time.
The RSI shooting lab will do it and now has links for deriving drag functions from the Labradar instrument directly. In the 19th century chronographs were electromechanical and you had to have a lot of space between the start and stop break wires, so finding a velocity for one particular yard meant working it out from averages of numerous results between those two wires (usually close to 150 feet apart with the first wire at 6 feet; this is why many military cartridges are given at 78 feet instead of at 15 feet like SAAMI uses. The original instruments, including those with electronic detection coils, were just too slow and imprecise to do that accurately).
So, basically, we are seeing things move from 19th-century tech, which was upgraded by the late Robert L. McCoy at the BRL into 20th-century tech, and now we are moving to 21st-century tech. Changes march on.