What are the Benefits of having separate layer/a dedicated schema for each layer in ETL (extract , transform and load) ?

In a typical cloud data warehouse, For batch inserts, There are three minimum layers:

1. Cloud Storage Buckets - Where the files land

2. Staging Layer - Where the files are inserted into tables. (1:1 mapping)

3. Target Layer - Where the Dimensions and Facts are stored

The CDC (Change data capture) happens between Staging and Target Layer.

Why do we need separate layers ? For audit and control

1. We can easily restrict the access with multiple layers (Your business users may get confused if you expose them to staging layers)

2. Its easier for backups (we can only take backups of Target tables and ignore the Stage layer)

3. Its easier to purge (Stage tables are kill and fill, so everyone knows its safe to delete the data in the stage tables)

4. Its much simpler to have a staging database, instead of having two similarly named tables in the target layer