It's easy to write a branching path story. You write one path, finish the game. Go back and write the next path. Literally the easiest way to do it and doesn't require much organization.
That only works with simple trunk-branch concepts - before the decision point you have a shared line, and after the decision you have 2 (or more) lines that don't rejoin. Upside - very simple, downside - each decision point doubles the amount of work you need to progress.
There are commonly two ways this approach is used:
- The classical shared trunk-split branch approach: in which there is a shared/common route from the start of the game until the decision point, at which the game will branch off into heroine-specific routes. This makes the amount of additional work berable, because it will be limited to the end of the game and by the number of heroines. This is traditionally how japanese VNs were structured, with some addition of:
- Continue Trunk-Bad End Branch: There's a correct choice, which will progress the game towards the conclusion of the game/next choice, and a bad choice, which will result in the main character dying or otherwise force a reload/selecting the other option. This limits the cost, as instead of having to do twice the work for the following parts, you just have to do once the work + handling that short bad/other end.
The annoying thing of it is... that's rather limiting. Choices either have to be a big deal (and dev-expensive) or a dead-end.
So another approach is split-rejoin-reactivity. Upon reaching a decision point, you give players a number of options, which will then play out in a number different ways, and sooner or later rejoin most or all of those threads back to the main thread(s), storing all those decisions. Then later down the thread, you can check which decisions the player takes, and react accordingly (or limit/introduce new choices for them in later decision points):
- Hey, you helped me there, so let me help out here. Or give you the option to get closer to me.
- You expressed interest in this topic, so here's an option to learn even more about it (and if you pursue this, you might unlock a route option down the line)
- You didn't go in guns blazing, which isn't just expressed in some various lines here and there, but also I don't know you're supposed to be my enemy! Here's some banter which gives you additional approaches to your current problem!
Because branching in this is limited to just the actual (short) sequences following from the decision point and at the points where the reactivity activates, it's much more economical. However, it also makes "just do one branch first" practically impossible, as it requires re-adding and re-re-adding new branches to the existing progression, reassessment of existing dialogue etc., it's just a mess.
The "replayability doesn't matter" part honestly doesn't even deserve a detailed analysis - Yes, most people will not replay games, but those that do are more devoted, how should they be weighted? Eh. Replayability, it also means that people playing a game for the first time have more options of how their game will go, more ways of expressing themselves, more ways and chances of them playing the game like they want to. But eh. You were pretty clear that you think games should be developed just according to your taste, to your preferences. And that's about as sad as your take on gamedev was ignorant.