One thing to understand is that a typical cell phone speaker has a frequency response from roughly 300 or 400 to 3.5 or 4K. It can also have a pretty high deviation in that range, on the order of +/-5dB from some of the measurement charts that I've seen. The addition of any low bass will be not be present, nor will any higher harmonics. So any significant musical content outside of that range will be absent. Since voice content is centered in that range, it will naturally be more prominent when played over a cell phone. But be aware that you are throwing away roughly 3 octaves of music on the lower end. That's a lot.
The question then becomes what will things sound like when you put them on a reasonably good system that goes down to 30 or 40Hz? Will the bass and low midrange that you didn't hear on the phone overwhelm the rest of the mix, or will it be non-existent and the whole thing sounds like a cell phone.
I'm not typical of today's listeners but the amount of time that I listen to music on cell phone is probably less than 0.1%. Maybe 1 or 2 songs a month, and those are more to let someone else hear something when a real system isn't available. I would never judge any mix by the way it sounds on a phone speaker.