My approach is to create directories for all the ideas I have. One directory for one idea. Add a readme file to that directory with explanation in what that ide

My approach is to create directories for all the ideas I have. One directory for one idea. Add a readme file to that directory with explanation in what that idea is, how I will execute it and what challenges I might face. Then let it be there for few days. If you think an idea is "doable", try to implement a simple feature of the idea. Make that implementation modular enough so that it can be used in other ideas. If you get bored with that implementation, then move on to another in the list. This takes a few days. But at the end you start liking one of these ideas more that the other ideas on the list. Now rate these ideas on "doable", "likable", "sellable" etc. And pick the one that has the highest rating. This is the continuous process until you get the one idea that you spend more time on implementing. That is the one that you should implement.