When your hardware project is "done" - that is, it works as you'd like, maybe you've made a handful of them - you're only really 30% of the way. First you need to test it - EMC testing can be painful and unless you really designed it carefully you'll probably need to redesign parts of it.
One you've got past that, you need to manufacture it. To get a project to volume manufacture means you need to work out all the little details - how will you program it, how will you test it, how much stock do you hold, who builds it. You'll probably find there's things you overlooked, or a part you can't buy any more, and you have to do a board respin.
When you see polished products on Kickstarter that manage to ship in the schedule they initially proposed, that's because they've already done all of this. They've probably spent 6 or 7 figures to get to that point.
I'm not saying you shouldn't do it (and there are paths you could take which aren't so onerous, like maybe making just a circuit board which you sell mainly to the hacker community) - I find the whole process really enjoyable. It's just that it's a lot of work and can be very, very capital intensive. Plan things carefully and be realistic, not optimistic about costs and schedule.