Some people actually think you should do the reverse -- do 2 & 3 before you do 1 & 4.
If you're just looking for ideas that seem like they might work, boy do I have the shameless plug for you.
But jokes aside, selling before you build is basically the best advice anyone can give, and it's repeated quite often but you just have to learn it yourself.
There's also tactical stuff like The Mom Test to teach you how to actually figure out what people need (and will pay you for), but the best strategic advice might be building audiences before you launch.
This is basically what all the "building in public" people do, and while I don't know that I'm a fan of the execution, it's an excellent tactic.
[EDIT] Just to explain what I meant by that first sentence more -- a lot of people believe in not even selling before you build, but being mostly market driven -- as in, spend time talking to 50 real estate agents you know to figure out what you could sell them before you even start selling or (even later on) building.
It's a really powerful strategy and you know it when you see it -- when you see serial founders that seem to go from industry to completely unrelated industry and build huge companies in a relatively short time, it's obvious they've mastered the market-first approach.