How to start with anything?
Start new things is easy. You just... start doing something, right? Wrong!
Every project requires some clearly defined phases, e.g. :
(1) Ideating, i.e. the time when we collect ideas and brainstorm rapidly. You could call it the "green thinking phase". That's when the ideas are born, and we allow ourselves to think about all the good outcomes and possibilities, without paying attention to the plausibility, difficulty, nor costs.
(2) Research phase is more strict, since it's the time to evaluate ideas, find out the requirements, asses risks, and score our ideas based on the real world factors like: cost, difficulty, required time, legal framework. Many ideas won't survive this phase.
(3) Planning is the time for strategical thinking: every idea should be estimated, and ordered in the general plan of action. It's no time for phrases like sometimes—every idea must be split into tasks, and each task must be an actionable item: collect documentation, create template for something, write test for particular feature, are all good actionable tasks.
(4) Execution is the doing time—either by going one by one, or strategically clustering tasks, it's the phase when ideas are executed. The more attention and care were paid during the previous phases, the easier execution will be: obvious mistakes can be avoided, as well as potential surprises, and unaccounted workload will be easier to tackle.
These 4 phases are how you can start with anything, from doing a breakfast, to creating new international companies: ideate, research, plan, execute!