To maximize query performance and minimize storage costs, a well-designed summary table should aim for a data reduction rate of at least 97% to 98% compared to the raw source data. Achieving this level of compression transforms unwieldy terabyte-scale datasets into manageable gigabyte-sized tables that support sub-second response times. If the reduction rate is significantly lower, you should re-evaluate your aggregation strategy to ensure you are not carrying unnecessary granularity.