He could tell what record he was looking at by sight by inspecting the groove patterns: number of tracks, track length, and even the patterns of the active part of the song itself. As I recall, he was proud that he once identified the 1812 overture from more than 10 feet away because of the placement of the grooves at the loud parts.
BTW, records didn't necessarily have constant spacing -- the cutter knew the amplitude of the part of the spiral plus/minus one rotation away so it could place the track close as possible without causing collisions. I believe that made it easier to identify songs.