I couldn't possibly guess what each individual manufacturer/designer does, but the reason for any changes to FR with cables is due to impedance matching (or mismatching). IEM impedances can get very low, which means that even subtle shifts in impedances of cables, connectors, and amps all matter. Also worth remembering that not all IEMs have a flat impedance, so the impedance ratings that manufacturers give us are just averages across a narrow (and often unstated) frequency range--say, 500Hz-5kHz.
If impedance graphs of IEMs were as commonly shared as frequency response graphs and if impedance measurements of cables were similarly shared (and accurate and consistent), we'd have a much better sense for how/why cables affect different sets differently. But these responses also change with power, gain, etc. from amps, so it quickly becomes complicated. The best answer IMO would be if cables were all spec'ed and manufactured the same, then we could just focus on the IEMs themselves. Alas. For me, rather than taking shots in the dark with cables, I try to keep my cables as similar in build, length, and connections as possible. It at least takes one major variable out of the equation. It's also quite easy to measure the impedance of a cable without any fancy equipment. A basic multimeter is all you need. Not the most accurate measurement, but good enough.