The problem with the "if it sounds good it is good" philosophy, and the "graphs don't matter" philosophy is coloration. To many listeners, especially people who seldom listen to unamplified live music or voices, which includes most of us, a saddle-shaped frequency response can initially sound pleasing. Everyone likes a plump deep bass and mid-bass, and many people associate brightness with what is actually the upper midrange, so they like it depressed a bit, and highs above 4KHz can sound more exciting to some people if they don't roll off naturally as they would in a large recording venue. Coloration doesn't go away; it lends a sameness to every recording played back through the highly colored speakers. I've found we can be trained to identify coloration by ear, by looking for differences between recordings, but it is difficult. I think properly generated graphs can tell you a lot about a speaker, and especially taking in-room measurements at your listening seat. Yes, it takes some equipment. Yes, it takes a little research to learn about how to take the measurements, but it isn't expensive and it opens up your mind.
Sure, every speaker is colored to some degree, but there is a wide variance between them.
As I've posted earlier in this thread, and many times before, I'm a great believer in listening before buying. I normally insist on that even with electronics, not because I expect them to sound obviously different, but I look for product design anomalies that will annoy me later. With speakers, however, I don't think listening is enough. It is necessary and IMO mandatory, but insufficient. If you don't care about coloration, that you'll never notice that sameness between recordings, fine, just listen before buying. I've just never been able to tolerate substantially colored speakers for very long. BTW, the least subjectively colored speakers I've ever heard are the Sound Lab electrostatics. Amazing devices. Just too expensive and huge for my current circumstances.