When a user types "running shoes size 10" into a search box, they expect results in milliseconds. They expect those results to be relevant, ranked by some notion of quality. They expect to filter by brand, price range, and color.

Traditional databases cannot deliver this experience. A SQL LIKE '%running shoes%' query scans every row, ignores word order, misses synonyms, and provides no relevance ranking. Even with full-text indexes, relational databases lack the sophisticated text analysis and scoring that users expect.

Full-text search engines are purpose-built for this problem. They use inverted indexes that map words to documents, enabling sub-second searches across millions of documents.

They apply linguistic analysis to understand that "running," "runs," and "ran" are related. They score documents by relevance using algorithms like BM25. They support faceted navigation, autocomplete, and fuzzy matching out of the box.

The Inverted Index