It's a great time for experimentation (outside the UCI straitjacket) to see what works; there's scope for lone inventors (harking back to Obree) to figure out what works, perhaps not so equally there are people designing stuff on intuition ("looks fast") but unable to practically measure the difference. Sounds like you're up on this, stuff like VE, Chung, DFDM etc as ways to meaningfully measure the effect of the many ways to improve bike aero.
