I totally understand the having to pick from two evils there, and yeah, I agree that there is no general best choice for it. Your approach of going with whichever of the two options seems best suited for the particular situation at hand seems most promising. I've played plenty of games who tried to stick to only option one or option two the entire way and neither worked too well imo. Option one (minimal changes depending on choice) does really feel massively low-effort and reduces replay value to almost nothing (from my experience it reduces a replay with different choices to skimming the text for those 1-2 lines that are different while ignoring the rest). Option 2 (doing vastly different scenes depending on past choices) is the great one in theory, but as you said it's a massive pile of work. It is amazing if a dev has enough manpower/time to pull it off and games that did that were among the best I played, but for smaller projects/solo devs it usually just ends with them being stretched too thin and the game never going anywhere because there is so much lateral expansion that horizontal (aka main story) progess is nil.