High cardinality can cause data “explosions” that severely degrade summary table performance and increase storage overhead. Best practices suggest limiting aggregations to columns with lower cardinality or using techniques like data normalization and string transformation to keep unique attributes within a manageable range. If high-cardinality data is required for specific metrics, it is often better to isolate those fields in separate, bespoke tables to avoid slowing down broader analytical queries.