I've been trying to do this for a few years (admittedly failing).
You can come up with complicated models as some do, but I prefer simple models (3 dimensions or less): What do you want? What are you willing to build/iterate on for years? What can you find feedback on fast?
(1) because if you're using it you'll build something nice and have a lot of insight internally.
(2) because you'll need to work hard, fast, and for a long time to make something great.
(3) your reward systems depend on feedback.
Sometimes answering these questions isn't straightforward so I recommend building a lot of smallish things without an outcome in mind before answers to these questions become clear. You tend to know when you're well suited to something because you can't stop yourself from coming up with ideas and working on it.
Example: I spent a long time building/understanding http://choremate.co because it's something I NEEDED. Despite finding insight, (1) I found the work hard to do because it was boring (I don't enjoy CRUDing web apps), (2) I'm not innately interested in the research domain (behavioral and institutional economics), (3) feedback is best found living in a large shared household like a coop or fraternity and I don't live in that situation.
Comments
I've been trying to do this for a few years (admittedly failing).
You can come up with complicated models as some do, but I prefer simple models (3 dimensions or less): What do you want? What are you willing to build/iterate on for years? What can you find feedback on fast?
(1) because if you're using it you'll build something nice and have a lot of insight internally.
(2) because you'll need to work hard, fast, and for a long time to make something great.
(3) your reward systems depend on feedback.
Sometimes answering these questions isn't straightforward so I recommend building a lot of smallish things without an outcome in mind before answers to these questions become clear. You tend to know when you're well suited to something because you can't stop yourself from coming up with ideas and working on it.
Example: I spent a long time building/understanding http://choremate.co because it's something I NEEDED. Despite finding insight, (1) I found the work hard to do because it was boring (I don't enjoy CRUDing web apps), (2) I'm not innately interested in the research domain (behavioral and institutional economics), (3) feedback is best found living in a large shared household like a coop or fraternity and I don't live in that situation.