Should I choose star or snowflake schema for fast queries in a data warehouse?

[Brian.J]

I’m trying to decide between a star and snowflake schema for our new sales data warehouse, and I keep going in circles. My team needs fast queries for regional performance dashboards, but the product dimension is heavily hierarchical with multiple slowly changing attributes. I’m worried the snowflake’s normalized dimensions will simplify some ETL but hurt query performance when we start joining all those tables.

[Noah.H]

We tried a star because dashboards needed to be fast. Product dimension got a wide, denormalized path with SCD Type 2 for changes, and it worked for the most part. ETL was messy but predictable, and queries on regional aggregates were snappy.

[ZacharyQS]

We flirted with a snowflake approach to tame the product hierarchy, but the team worried about all those joins hurting dashboard latency; is the extra normalization really worth it?

[Stella_S]

We kept the star but added a separate product rollup table to speed region filters, and used a small amount of precomputed joins when loading the cube. It felt heavy to maintain but the dashboards loaded faster.

[Aria96]

We measured latency and saw a 30 percent drop after adding the pre-aggregates, but the ETL pipeline became a bit more brittle and scheduling drift started showing.