Use when designing data-quality coverage for a new data product or auditing an existing one: choosing dbt tests vs Great Expectations vs Soda, column vs table-level checks, severity tiers, SLA and freshness conventions, and anti-patterns to avoid.