> The point is to exactly avoid this kind of overhead when you have zero paying customers, as that's premature optimization. It should take a day at most to wo

> The point is to exactly avoid this kind of overhead when you have zero paying customers, as that's premature optimization. It should take a day at most to work out the data model for a startup. I don’t see that as a lot of overhead. Do as you will, but from my perspective, I’d rather do it correctly from the start.