I believe the correct formula is P=E(squared) / R
That's the correct formula and the reason why the voltage under load is so critical to the amount of heat produced (because the voltage is squared to arrive at the amount of heat produced). So, for example, if the seat element measured 2 Ohms and system voltage is 13.8 volts, you would expect 95 watts of heat. But, if the voltage falls to 12.2 volts under load (for example due to inadequate wire diameter), then only 74 watts of heat is produced (or 22% less heat from 11.6% less voltage).
However, the CX-5 heated seats are thermostatically controlled. So, if your seats don't get hot enough for your preference, the question is whether the current is cycling on/off even on high heat setting.
If the seats cycle on/off when set to high, then the heat can be increased by insulating the thermistor from the seat heat (or moving it further away from the heating elements). Perhaps a more controllable method would be to wire a potentiometer in series with the thermistor to fool the brains into thinking the seat is cooler than it is, thus allowing the elements to be powered a higher percentage of the time.
If the heaters are not hot enough for your preference but they are not cycling on/off when set to high, then the only solution (without adding/replacing heating elements) is to increase voltage to the elements. This could be done by reducing resistance in the feed circuit (for example, by increasing wire diameter) but only if the feed wires have significant resistance to begin with.
Regardless of how the heat is increased, it is critical to insure it doesn't get hot enough to cause skin burns or damage to the seat padding/leather (or cause a fire). Obviously, it would not cost Mazda significantly more money to make the seats hotter but it could cost them down the road in terms of liability and or warranty costs should injury or seat damage occur.