Be good at something and help someone with that.
If you're not good yet, do it for free first until you get good, cause it will take you awhile to debug and troubleshoot and work out the kinks of what you're trying to do.
Then do it for cheap until you're an expert. You have little to no overheard a the beginning (other than time invested), so you can do it cheaper than anyone else. Think of it like you're taking a 50% pay cut: 50% is paid to you now, and the other 50% that you don't get now is in the form of experience because you got to do a job now that you wouldn't have had the opportunity to do, because 100% cost to them would be too much. Then that 50% comes back to you in 5 years, because now you have one more tool in your toolbelt. The next time someone needs help with that, it will take you 50% less time (or probably more like 80%).
People value their time, but they won't pay for something that saves them 10 or 20 minutes, especially if it just takes a quick Google to solve (although you'd be surprised how many people can't do a simple Google to solve a problem. Maybe start there?)
Become so good that you save them an hour, or 3, then maybe half a day, or a few days.
Then charge an equivalent rate for the amount of time or hassle you've saved them.
I learned a great piece of wisdom from some random soul on Hacker News that I'll never be able to thank: increase your rate with each new client, until you start getting push back. Then you'll know your worth, or at least the appropriate range. I went from $25/hour when I first started, to $175/hour now. It took 10 years, but now the $175 is a bargain for most people, because I'm no longer saving them an hour or two here and there; I'm the difference between their business working or not.
