What is denormalization and when should you use it?

Denormalization intentionally adds redundancy to improve read performance by reducing joins. Techniques: adding computed columns, duplicating data across tables, pre-aggregating values. Use when: read performance is critical, joins are expensive, data changes infrequently, query patterns are well-known. Trade-offs: increased storage, update complexity (must update all copies), potential inconsistency. Common in data warehouses, caching, and reporting systems. Always measure actual performance impact.