The way I go about this is to first record whatever idea I get onto a document that I keep. From my past experience, I tend to sometimes get very excited about

The way I go about this is to first record whatever idea I get onto a document that I keep. From my past experience, I tend to sometimes get very excited about ideas, and that usually stops me from thinking through properly about its viability. So, I generally force myself into a 2-3 week "pause and ponder" period in which I resist the urge to start working on it immediately. Even if the idea is trivial to implement (trust me it always looks like that), it's a distraction from stuff you are already working on, and you don't want to rush into implementing it and end up nowhere. I must add that, after 2-3 weeks, most of these exciting ideas sound really bad to me. There will be a few that will still sound exciting. These might be ones that are worth pursuing.