Semantic Integrity
Default Aggregation: When SUM Is the Wrong Assumption
Default aggregations can distort results when a sum is not meaningful.
Rendering diagram...
A flow showing rate column summed incorrectly vs explicit measure.
TL;DR
- • SUM is not always correct.
- • Set explicit aggregations or measures.
The problem (layman)
- • Columns default to SUM even when the right behavior is average, max, or distinct count.
- • AI queries often rely on default aggregation.
Why it matters
- • Wrong aggregation leads to wrong answers with high confidence.
- • Errors are subtle and often missed.
Symptoms
- • Rates or percentages are summed.
- • Balances are aggregated across time.
Root causes
- • Default aggregation left untouched in model.
- • No explicit measures for key fields.
What good looks like
- • Explicit measures for critical metrics.
- • Columns set to correct aggregation or “Do not summarize.”
How to fix (steps)
- • Audit columns used in AI and reports.
- • Define measures for rates and ratios.
- • Set default aggregations in the model.
Pitfalls
- • Relying on report visuals to override defaults.
- • Leaving raw columns exposed for AI.
Checklist
- • All ratio metrics use measures.
- • Default aggregation settings reviewed.
- • No critical metric depends on implicit SUM.