The underlying point is that the short answer is that two competent laboratories on identical material will disagree, and the disagreement is almost always explainable by method differences.
Gradient slope controls resolution, and gentler slopes resolve co-eluting impurities into separate peaks — so the better method reports the worse purity number.
The fraction of your main peak that is actually your target versus isomers, fragments or related sequences is invisible without complementary identity testing.
Proline conformer interconversion kinetics are well-characterised and the half-life is of the same order as the chromatographic peak width at room temperature.
I would be careful about over-reading a single measurement — treat it as a data point, not as ground truth.
The practical summary: ask for the chromatogram and the method, and ignore the headline number until you have both.