The half-hour decision that costs months

The database schema, the auth model, the pricing structure, the brand name -- the decisions that take half an hour at the start cost ten to a hundred times more to change once they’re sitting in production with customers leaning on them, and almost nobody recognises the moment they’re in while they’re in it.

The decisions that look the most boring and the most obvious are almost always the most expensive ones in disguise, and “later” is a word that almost always resolves to either “never” or “hugely expensive when we finally do it” -- which is why the best discipline on a new project isn’t moving fast, it’s knowing which half-hour to spend three hours on instead.

The same principle that applies to a database schema applies to a hiring decision, and a half-hour call on those first three becomes a decade-long reality -- the kind that’s invisible in week one and load-bearing by year three, when you’re trying to figure out why a culture you didn’t mean to build is the one you have.

Spent half a day this week unwinding a Firestore decision I made in roughly the time it takes to drink a coffee, and the lesson keeps being the same one -- the early shortcuts you can’t see the cost of are the ones you end up paying back later at a far worse exchange rate, with interest, on a day you needed for something else entirely.

Hit reply -- I’m collecting these for a future deep dive and the most useful answers always come from people still in the middle of the bill arriving, not the ones who’ve already filed it away as a lesson.