A map that includes everything is no longer a useful map. Models work by leaving things out. The question is whether the omissions fit the decision we are trying to make.
Name the purpose first
A model for explaining a mechanism has a different job from a model for making a forecast. Confusing those jobs makes a clean explanation look like a promise about the future.
What question is the model answering?
Which assumptions carry the result?
What changes if an assumption fails?
What evidence would make this model less useful?
Make the omissions visible
Write the assumptions next to the result. This small habit makes it easier to notice when the setting has changed. A model is a tool for thought, and tools need to be matched to the work.
Use more than one view
Several small models can reveal different aspects of the same situation. The disagreements between them are sometimes more useful than any single answer.
Keep a record
A short note about why a model seemed useful can make later evaluation much more honest.