Denormalisation
Deliberately duplicating data across records to make reads cheap, accepting the write-time cost of keeping copies in step.
Normalisation optimises for write correctness — one fact, one place, no possibility of disagreement. Denormalisation optimises for read cost — the data is already shaped like the query, so no join is needed.
It is the default in wide-column and document stores, where you model per access pattern rather than per entity, and it is a targeted optimisation in relational ones.
The cost is not storage, which is cheap. It is that a fact now lives in several places, so every writer must update all of them and any missed path produces silent divergence. Before denormalising in a relational store, verify the join is actually the bottleneck — an appropriate index frequently removes the problem without introducing a consistency obligation you will carry forever.