Yep.

That doesn't seem like a silly answer.

The point is this, if you take a speaker with multiple drivers (including ports), then measure its output in three dimensional space, you'll measure a variety of interference effects in the crossover region of the frequency response as you move around the chamber. When you put together a speaker, you diddle with the design as a whole to get the on-axis response "right".

Now, if you start fiddling around with the delay to one or more of those "outputs", then the interference pattern moves around in 3d space, and depending on the relative listening position, possibly introduces a big suckout there.

Getting the on-axis response "right" matters (especially with room treatment) despite the nuisance of having a real room and listening in the reverberant (rather than free) field.

[So, Markus, this isn't just group delay, but (potentially) a change in the amplitude response too..]