[aerohead]

I'd give up on trying to quantify delta-CdA for anything small. That's what wind tunnels are for.
If, say, you have 'smaller', 'higher fineness ratio', and 'further-distanced' mirrors from the car, you might just presume that there's an advantage.
You really can't afford to 'know.'
Old-school road testing requires a thermal stabilization period of 30-minutes, driving 50-mph, before you ever begin the actual testing.
1) Calm ( or not to exceed SAE specs )
2) Dry
3) Straight
4) Level
5) 10 runs, back-to-back, with each configuration
6) Continuous time-stamped weather monitoring
7) 1/10th-second recording resolution.
8) Coasting through 70-mph, down to 20-mph
9) 5th-wheel data transmitter ( mechanical or optical )
10) Data-logger/ strip recorder
11) Done to SAE Handbook specifications.
12) You may prefer root canals to the mathematics required for data reduction.
------------------------------------------------------------------------------------
I bailed, and just paid CAR and DRIVER to do the tests at the Chrysler Proving Grounds. Between that, and Bonneville, and lots of tank mileage, I eventually got some numbers that told the tale.