The inputs, via an iPad video and scanner, to generate a 3D model, will always be less accurate than testing in a tunnel..
...but it's faster and cheaper with reasonably accurate results, so surely that's good enough?
One of the problems with less precise analysis techniques is they require the user to be more particular about inputs and to have a expertise on when tests will be useful and when circumstances will lose accuracy. GIGO
The problem with Stac, then, may be that it is inherently harder to get accurate results but also lowers the barrier to entry for testing.
On the assumption that their results are as accurate as they say, then I'm keen, even though I know it's not going to be perfect. To me, it seems like a great way to decide which helmet to buy. Or maybe, which bike/frame, or which aero bars (straight vs ski bend) or any number of other things.
It's way more accurate than reading a manufacturers brochure, or listening to the guy at your LBS (he'll just recommend what he has in stock). And that's worth at least a hundred to me. If they really want to sell to us trigeeks, they have to work "carbon" into their sales pitch somehow. (that last bit, pink?)
"Happiness is a myth. All you can hope for is to get laid once in a while, drunk once in a while and to eat chocolate every day"