If the data model is unstable, KPI logic is inconsistent, refreshes are fragile, and the reports are slow, patching around the edges can become more expensive than fixing the structure properly.
A rebuild is often the right call when the business no longer trusts the numbers, when report ownership is unclear, or when the report has become so layered with workarounds that every change creates risk.
The commercial test is simple: will another round of fixes make reporting easier to run, or are you just preserving the same friction with new formatting?