A lot of people don’t like measurements, and I can understand why. This stuff gets confusing and is sometimes a bit too ‘into the weeds’ - I’m guilty of that for sure. But, FR measurements from industry standard rigs are currently the best visual indicator for sound quality - and they’re also better and more reliable than sound demos. They’re not all there is to it, but at the very least they can tell you if something is going to sound bad. More importantly, if you’re unsure what to make of a given perspective (we’ve all got our biases and preferences), measurements do a good job of cutting through that.
So far, in my experience the biggest limitation on measurements is that they can be difficult to understand and analyze properly. Now, I’m not an ‘objectivist’, or at least not the hardcore variant of that - but I’ve seen much of the resistance to measurements come from a lack of understanding for how they work. Even just distinguishing between raw and compensated measurements asks a lot of the audience, and so a big part of my project is to help people be able to make sense of them better. It would be nice if we didn’t have to just rely on alignment with a reviewer’s preferences to be able to make decisions on equipment.
One last note on that - I don’t think we’re at a point where we can just look at data and say “here’s everything there is about how the headphone sounds”, but the folks over at Harman research have shown that at the very least measurements are predictive of preferred sound quality - with a strong correlation between even spectral balance or ‘neutral’ and listener preference (what that ends up being is still up for debate).
For anyone wanting further reading on this, this paper is a good place to start: http://www.aes.org/e-lib/browse.cfm?elib=16486