That's a good simple solution. However, to get consistent, repeatable data sets, you should keep your efforts the same for each setup. Using a constant power output for each run will show your speed differences between setups pretty accurately. You could also do each run at a consistent HR range, but be aware that HR is your body's response to the workload you're producing, not a measure of the work itself.
I did the same type of tests yesterday, using a two-mile stretch of road. I found that my best position had the arm rests a centimeter higher than I used in my first TT last month. I increased the height by 5mm each run until my speed started dropping off, then re-ran the course in the prior best configuration to validate the earlier run. My best position took 15 second off my second-best position over the two miles.