I do some electronics related freelancing aside from my day job as a way to get some extra income, fund my hobby and reduce the woman annoyance level. In general there are two ways: you can either specialize in something (eg. power supplies, led lighting and so on) and come up with a product that stands out on the market in some way, so that you have customers. Another way (the one I'm going) is to offer various services to people who who want to pay.
Bright side is you'll never get bored, because every project is different. Doing projects from A to Z (spec, design, prototyping, testing, and so on) is most pofitable but it will also take a lot of your time, which translates to high cost for the customer, so you won't have many, as people who have that kinda money will rather hire a full-blown engineering outsourcing company rather than an individual and private individuals or small companies donlt want to spend that much. Also, there is a lot of places where something may go sideways and you'll end up at a loss.
I think there's more money in small jobs, especially is this is your side activity. I'm talikng about stuff like debugging some part of the circuit (eg. my power supply will shut down from time to time), repairs of specialistic equipment (eg. replacement of transistors and caps in a high-power smps for which an authorized repair shop would charge an arm and a leg), designing small to medium pcb's (it's easier to screw up a schematic than the layout), doing some projects for desperate students when end of semester is approaching (if you want this stuff done tomorrow, you will pay this $50 for half an hour of my work and the napkin on which I drew the circuit), Change a connector in a pcb design to a different one, because the original one is hard to source.
Consulting is a good way to make quick buck, but you need to be prepared for interactions with VERY stupid people. Companies tend to hire consultants when they are in a pinch which usually means that you land in a sticky situation (customer claim), monstrous design screwup (people will later be able to explain to their bosses that "they even hired a consultant, but he wasn't able to help" and they have to respin $10k worth of pcbs), company has low credibility because of previous screwups ("this time it's ok, we even had an external consultant look at this design"). Anyway, from few experiences I had with consulting it's a well paid, but stressful and often frustrating job.
If you want to concentrate on designing stuff from ground up, then speciality items are the way IMO. There are situations when people need a device that is very specific and has to be designed specially for them because there is nothing on the market that fits the bill. This includes exotic devices that will be produced in low quantities or even 1-offs. I did numerous jobs that went like: "we want something like that [a link to some product] but powered from [something] instead of [something else] or with output that will soure [number] amps more". Or "we need [link to a product] but with 50 outputs on a small connector instead of 8 on a terminal block that will fit [LxWxH] slot in out product". Those jobs are often simpler than ground-up design, because you can typically buy (or get from the customer) a sample base device to analyze.
And don't hestitate to work for foreign customers if you live in a poor country. It pays off. My fees are rather low compared to most people living in the western EU or US, but they make for a really nice addition to the home budget in my home country.