The decisions worth writing here are mostly the negative ones, and they are the ones founders reconstruct badly under pressure.
Record, each with the alternative rejected: the amount and what it buys — in months and milestones, not dollars; the structure and why; who you are deliberately not approaching; and what you will not accept — a board seat, a particular term, a strategic investor whose real interest is a future acquisition.
Write them in week one, because of how long week one is from the end. With 50% of successful 2023 seed raises taking 13 to 24 weeks, a decision made at the start will be tested by a version of you who is three to six months more tired, has heard thirty-eight versions of the same doubt, and is being offered a way to stop. The pressure runs steadily toward the thing you already decided against, and it peaks exactly when your judgement is worst.
A decision note with its rejected alternative visible is the cheapest defence available. It does not stop you changing your mind — it stops you changing your mind without noticing, which is the failure that actually happens.