A messy data model

If adding a new measure feels like defusing a bomb, the model is the problem, not you.

One huge flat table instead of facts and dimensions
Relationships that fight each other, ambiguous paths, inactive links nobody remembers
Measures scattered over tables, with names like "Measure 2 final v3"
No documentation, so only one person can change anything

Assess Review tables, relationships, keys and measures against star-schema practice.
Redesign Propose a target model and a safe migration order.
Rebuild Implement the new model and repoint reports without changing the numbers.
Document Model documentation and naming standards the team can keep using.