You need to design the dimCustomers table. Which design approach should you use?

A reference dimension is a dimension linked to a fact table indirectly through another dimension, which addresses relationship design rather than handling attribute change history.

Type 2 Slowly Changing Dimension adds a new record each time a tracked attribute (such as a customer's address or status) changes, preserving the historical value alongside the new one using surrogate keys, effective dates, or current-row flags. This allows fact table rows to be accurately associated with the customer's attributes as they existed at transaction time. It is the standard approach for dimCustomers when historical accuracy is a business requirement.

A junk dimension consolidates miscellaneous low-cardinality flags and indicators into a single table to reduce fact table width, and has no mechanism for tracking customer attribute history.

A conformed dimension is a dimension shared and reused across multiple fact tables or data marts to ensure consistent reporting, but it does not inherently address how historical changes to attributes are stored.

Type 1 SCD simply overwrites the existing attribute value with the new one, destroying historical data and making it impossible to report on what a customer's attributes were at a prior point in time.