Electronic switching is great for circuits that can afford a little resistance along the way. Circuits such as electronic ignition and logic can afford a little in-circuit resistance because the result shows little (if any) difference to that resistance. That, and there is usually a path to dump some heat along the way. Ignition circuits often uses wire with resistance built-into it (mainly to reduce electronic magnetic interference) and the loses from the primary side of the coil are relatively negligible when you consider what you get at the secondary side The coil itself is a heat sink to the losses it absorbs. Logic circuitry uses 1's and 0's with a relatively wide range in-between so the loses there are unnoticed. Circuits such as an injector (basically a solenoid coil) are categorized as a logic circuit because they are either full on or full off with a mechanical delay in-between. Still, you have to consider the insertion loss of the switching device and what to do with that lost power if it's significant.
But with circuits with high demand, such as motors you want all you can get because a small current loss is noticed in performance loss.
In the old days when we were forced to use Transistors, particularly Silicon, where a minimum 0.6V loss was observed -which was a lot when you start drawing current.
We could use Germanium transistors to reduce that to a 0.3v loss, but it was still a lot. (remember those old Delco ECM's?) Today, we have power MOSFET devices as shown in the early part of this post, which has a typical 0.1V loss, which is a lot nicer when you consider the other choices, but even that 0.1V loss can add up when you're dealing with high current devices. You still have to heat-sink the device, as drawing 10 amps at a 0.1V loss through such a device will produce 1 watt of thermal energy that you have do dissipate somewhere, else lose the MOSFET device to thermal breakdown. For intermittent devices, this is not usually a problem, but for continuous duty, it becomes a big problem.
One way around this is to use pulse width modulation. This is used a lot in automotive dimmers where you send the full voltage to the device (usually lamps or LED's) by using variable width pulses. Our eyes can only discriminate a 40 millisecond pulse, so keeping the pulse width much faster makes the light appear to ramp smoothly and not flicker. Pulse width modulation is used in fuel injection systems as well. This method can also be used for varying motor speeds, but the switching frequency wants to be higher than our hearing range so we don't hear the switching frequency come through the motor windings.
Regardless of your application, you need to keep in mind that when using Power MOSFET devices to switch your devices on & off, you will also need to dissipate the heat generated from the loses. In some cases, this can take a considerable amount of physical space to do. LED Lamps for example, struggle with these power losses a lot.
Sometimes, for simple on/off applications, it's best to use a relatively zero loss relay, but for high speed switching such as fuel injection and the likes, that's simply not an option.