Without understanding the endpoint, one of two things almost always happens.:

Overengineering — they do it "for growth", lay down complex architectures that will never be needed. This is bad for business: it takes longer, is more expensive, and is unnecessarily difficult.

Underengineering is done "on the knee", without architecture and a foundation for development. This is bad for development: difficult support, chaos in the code, expensive revision.

And here's the paradox: the second case is worse for developers, but easier for businesses.

Why? Because:

— A project from scratch is rarely technically evaluated — the customer does not know that "good code" is cheaper in the future.

— And no one pays much for minor edits, even if it takes 10 times longer to deal with the project than to solve the problem itself.

Therefore, vision is not a luxury, but a necessity. Even a basic roadmap greatly improves the quality of architecture and saves money.

And what was your experience — did you have to work in projects "blindly"?