Of course different transducers, electrical and acoustic tuning, will result in different amounts of distortions which, while usually not disturbing because they tend to be much quieter than music, do remain audible. So EQing one IEM into the other one will always have some remaining level of sound difference coming from distortions(THD being only some of it).
But that's absolutely not what you're hearing when you try to EQ the IEM to match the other one using graphs! If the frequency response did match at your eardrum, you would agree that the difference is somewhat subtle and perfectly acceptable and enjoyable IMO. It might not be true for IEMs with vastly different FR or if you tried to correct some massive frequency roll off at either end of the spectrum (because then the distortions might be quite loud or the FR adjustment might just plain fail to reach the target). But here with your example, I wouldn't expect much issue, beside perhaps in the upper treble.
Sadly, that's only a thought experiment at this point because you or I, do not know how to EQ one IEM to make it sound like the other one at YOUR eardrum. The graphs you show are the frequency responses measured when inserted a certain way into a coupler, and I'm guessing they are different pairs than those you have. So you have FR differences from using your ears and your preferred/imposed insertion instead of a particular coupler, and you have some more FR differences from using different pairs(unless a brand puts in particular efforts to try and limit the deviations between pairs, you can expect a few dB here and there between pairs of the same model if not just between left and right pieces).
The bass can be different from measured because the seal quality has a huge impact on the low frequencies.
The 3kHz area changes depending on the insertion depth and length of your own ear canal.
Treble changes significantly from the tiniest change in insertion. Graphs must not be taken at face value for trebles, never!
So from a given measurement, you can sort of count of the 500 to 800Hz area to be just what you see. Cool.
For bass, you need to trust that the measurements were made with the best seal, and that you get the same in practice. We can sort of tell when we get a seal, but that's valid for the moment we insert the IEM, 5mn later might be another story. There is also some issue depending on the coupler and IEMs(vented or not being the main variable) where the measured low end is not what you will perceive even with everything done right. At least there is some hope that between 2 pairs, if all is done well, and you get similar seal quality on both pairs, then the FR difference measured will be as you hear it.
The 3kHz area is going to be whatever because your ear canals aren't like the coupler, and they're the main cause for resonance in that area. The measurement only tells you that if there is little difference between pairs with similar insertion depth, then you'll hear little difference in that area.
Last but not least, above 5kHz or so, you're completely on your own and trusting a graph is pretty much guaranteed to give you the experience you described because what you're actually hearing could easily have up to 10dB deviations at various places in the treble. The graphs are nearly useless when it comes to how you'll hear the upper frequencies. If the graph shows loud overall treble, then of course you will hear that IEM as having more treble than one with a rolled off graph, but the actual amplitude at each particular frequency is a very different story. What's even more troublesome is that all your efforts to correct say 13kHz based on the graphs, will go down the drain if the IEM moves a little in your ear over time, which most do! Resonance shifts from insertion differences get bigger the higher the frequency. I personally tend to attenuate the upper trebles or not touch them at all. A boost is just too much of a gamble(high freqs are higher energy and more likely to damage your ears if too loud) for frequencies that ultimately don't do all that much for music.
The long conclusion of an already long and disappointing post is:
1/ frequency response, in the absence of something else gone very wrong, is a main variable for your experience and more importantly for your sound preferences. If you do manage to EQ to a FR you like with reasonable settings, music will sound better. The price of the IEM isn't remotely as significant as getting a FR you favor for your overall listening pleasure. Again, +15dB to try and get something physically blocked or cancelled, are more likely to create garbage. EQ isn't there to turn a donkey into a race horse. But with an already fairly smooth and well extended tuning and THD below 1% in the midrange, EQ(without clipping the signal!!!!!!!!) is a very powerful tool to improve listeners' enjoyment.
2/when you match some online graphs, or really even graph from your own measurements like I do, the result sounding different is not evidence that the other sound variables are much more important than FR. Because FR is not actually well-matched in such a scenario. People just falsely imagine it is, and jump to false conclusions based on false assumptions about matched FR.
3/ You kind of have to trust your ears when EQing, which makes it a difficult tool to use well. Because while Audiophiles love to gloat about all the tiny things they can notice, and how perfect their memories are, when faced with having to use those golden ears and robotic memories for actual work, there is predictably almost nobody left. Learning how to EQ takes time and practice, and even then, we're only humans. It's hard to keep an accurate memory of a tuning long enough to take off one IEM and put another one in(properly!!!!) then EQ just right.
This might be a better starting point than online measurements
https://alexiy.nl/eq_chart/
