<result> <column_name>UNIQUE VIEWS</column_name> <family>additive</family> <subtype>real_additive</subtype> <confidence>0.93</confidence> <requires_user_confirmation>false</requires_user_confirmation> <sum_meaningful>true</sum_meaningful> <signed>false</signed> <reducers> <primary_reducer>sum</primary_reducer> <valid_reducers>sum, min, max, mean, median, count, distribution</valid_reducers> <excluded_reducers>product (not a ratio), geomean (not a ratio)</excluded_reducers> <average_semantics>arithmetic_mean</average_semantics> </reducers> <inner_outer_measure> <inner_measure_reducer>sum</inner_measure_reducer> <outer_measure_reducer>sum</outer_measure_reducer> </inner_outer_measure> <multiplicative_partner_derivable>false</multiplicative_partner_derivable> <reasoning>The table is STATE with no sequence axis, so the prior strongly favors additive and rules out multiplicative. The column is a non-negative integer count of views (name token "views", BIGINT). It has a heavy right tail (max about 16.4M, median 6, mean about 4101), is nowhere near 1.0, and has no ratio naming. These signals indicate an extensive count where summing across merged volumes is meaningful. A "unique" count may overlap across merged volumes, so sum is an upper bound if uniqueness is defined per user. Still, it is treated as a standard extensive count at the entity grain. The high log_symmetry_gain only reflects the skewed count distribution, not multiplicative behavior, so I discounted it.</reasoning> <table_type_conflict>false</table_type_conflict> </result>