What's the use of fixing something in a cost effective way if your startup is going to fail anyway?
You need to validate and test first. Get customers ASAP. If that means writing shitty code and using a shitty database or not even a database at all, than that's great. You can always convert to a better system afterwards.
I do agree that the longer you wait, the harder it gets to rewrite code and convert to different systems, but the most important thing for any startup is to get their idea validated as soon as possible.
You need to validate and test first. Get customers ASAP. If that means writing shitty code and using a shitty database or not even a database at all, than that's great. You can always convert to a better system afterwards.
I do agree that the longer you wait, the harder it gets to rewrite code and convert to different systems, but the most important thing for any startup is to get their idea validated as soon as possible.