6 years ago or so, Zinn included this Pro Tip of mine in a book he was writing:
"In my observations at the San Diego Wind Tunnel, the particular brand of aero helmet that works the best seems to be specific to the individual. The geometry of the helmet and how it integrates into the shape of a person's back and shoulders seems to be critical; the optimal combination for an individual can only really be determined in the tunnel! Despite that caveat, based on a sample size of 17 the typical time savings for using an aero helmet versus a standard vented road helmet is approximately 41 seconds (+/- 20 seconds) for an effort that originally took an hour."
I checked my raw data for this claim, and that works out to be about a 3.1 +/- 1.5% reduction in CxA.
It's entirely possible that this quote is where the current meme of "optimal helmet selection depends on the rider" got started. I'd like to revise my quote based on new information over the years:
"a good TT head position is pretty robust when considering the aero helmet variable."
That probably doesn't help you much, but at least it reflects my current thinking (...and makes you think a bit, hopefully?)! ;-)