As I've said in a number of past threads, IMO the main usefulness of specs is to allow one to identify and RULE OUT from consideration candidates for purchase that would be poor matches to either the surrounding components or to the user's requirements. An example of the latter would be how much power is needed to support desired peak volumes. Examples of the former would be incompatibilities due to impedance issues, gain and sensitivity mismatches, and various issues which can result in amplifier/speaker mismatches (as Russ/Rcprince alluded to above).
Specs are also useful in troubleshooting problems, in getting to understand the priorities of the designer, and in identifying specs that may be TOO good and thereby may signify undesirable design tradeoffs. A classic example of the latter would be an amplifier having unusually good Total Harmonic Distortion numbers relative to other comparable kinds of designs. That will often signify the use of large amounts of feedback, and consequent adverse effects on things like Transient Intermodulation Distortion which arent normally specified or measured, or that are not even measurable in a standardized manner.
Basically, as I see it, if specs are mostly or entirely disregarded and/or are not properly understood the randomness of the process of assembling a satisfactory system, and the likelihood of expensive mistakes, are increased considerably.
Regards,
-- Al
