"We will sort that out later" in a software project

Some postponed things cost double later. Others resolve themselves. The difference is predictable.

What gets more expensive

Data structure, access rights and user management. These are foundations: changing them later means touching everything built on top.

What can wait

Visuals, secondary features, reporting and optimisation. Those can be added without breaking what exists.

The special case

Security and logging. Technically addable later, but in the meantime you are running a system without knowing what happens in it. That cannot be recovered retroactively.

How to decide

Ask: if we add this in six months, what has to change? If the answer is "nearly everything", do it now.

Write down what you postpone

An unrecorded decision becomes a surprise. Keep a short list of what was deliberately deferred, with the reason.