The system was also touted as being so precise that a skater could improve his/her performance by simply stuyding the protocol from his/her last skate. The protocol was supposed to tell him/her exactly what to improve upon and how. That implies that if skater does X in the same way next time, the same score should be achieved, but if skater does X better, a better score should be achieved and figuring out what better is is a matter of looking at the protocol.
I never heard that these concepts only applied to the technical side of things.
When IJS was introduced to the U.S. the true believers from the ISU would mock (and still do) the kind of things judges would tell skaters after a 6.0 competition about what was wrong with a program. The protocol under IJS does only a slightly better job. Now it is (hopefully) clear to a skater if a jump was cheated or had the wrong edge. But if a spin or sequence does not get the level expected, the skater does not know from the protocol why. What features did they get? What features did they miss? Why? For PCs the skater can see how they compare to the others, but why did they get a specific PC mark? Which criteria (of the 38) were good? Which were poor?
In the end, if the skater really wants to know what to work on, they still have to talk to the judges or technical panel for specific guidance. One benefit of IJS, though, is now the judges and technical panels can more clearly point to very specific things for the skater to work on (specific features or PC criteria) when they talk to the skaters.