When I'm reviewing gear, audio analysis software like FuzzMeasure [this issue] gives me the ability to resolve differences in sound quality and consequently use objective language when conveying those differences. But I would argue that all recording engineers would benefit from learning how to use an analyzer and decipher the data that it presents. Even if we're not doing formal product evaluations, we're still constantly presented with a choice of recording tools. Knowing how single pieces or whole signal-chains of gear react to certain sounds or techniques can steer us toward more efficient workflows - mic'ing, processing, and mixing.
But shouldn't we trust our ears over numbers, specifications, and charts? Of course we should. But can we absolutely trust what we're hearing across space, time, and memory? And more specifically, can we trust the monitoring system that our ears are hearing, or the room that we're mixing in, without having some kind of baseline to accurately describe that system or room? At the very least, an audio analysis package can highlight the anomalies in how our rooms react to the sounds coming out of our speakers, and if we choose to go further, analysis is the first step in identifying what kinds of acoustic treatment we should consider to offset those anomalies.