You are developing a SQL Server Analysis Services (SSAS) cube for the sales department at your company. The sales department requires the following set of metrics: - Unique count of customers - Unique

Placing distinct count measures in the same measure group as additive measures causes severe query performance degradation because the engine cannot apply separate optimized aggregation designs for the two incompatible aggregation strategies.

Placing distinct count measures in separate measure groups is a required SSAS best practice because distinct count measures use a fundamentally different internal storage, partitioning, and aggregation strategy - mixing them with additive measures in the same measure group forces the engine to apply the less efficient distinct-count query plan to all measures in that group, severely degrading response time.

Additive aggregation functions such as Sum and Count include duplicates and therefore cannot produce a unique (distinct) count of customers or products.

Semi-additive functions such as AverageOfChildren and LastChild are designed for time-based aggregations of balance-type measures and cannot perform distinct counting of dimension members.

The Count aggregation function counts all rows including duplicates, so it would overcount customers and products and fail the uniqueness requirement.

Using the Distinct Count aggregation function for the unique customer and unique product measures ensures accurate deduplication at every aggregation level, while using Sum for the sales measure provides correct additive rollup - this combination satisfies all three stated metric requirements with the correct semantics.