Case Study: General Mills achieves high-performance, scalable website search and content delivery with Elastic

General Mills - Customer Case Study

General Mills needed a more flexible, high-performance search and content delivery platform for its public sites (BettyCrocker.com, Pillsbury.com, TabRecipes) as it migrated from Endeca to Elasticsearch. The challenge was to index diverse content types (recipes, articles, videos, products, coupons), provide relevant results driven by titles and metadata, and empower content owners to configure classifications and workflows without constant developer support.

The team implemented an Elasticsearch-based solution with a production cluster (5 data, 2 client, 3 master nodes), a SimpleSearch layer, GUID-based metadata and grouping, index aliases for safe atomic re-indexing, and an information/attachment repository plus admin tools (e.g., a configurable notepad). The result was excellent scalability and relevance—peaking at ~3,500 shard searches/sec while ingesting ~350 records/sec with only ~40% max node utilization—better-than-cache performance and a system that content owners can adapt themselves, reducing development overhead.

General Mills

Doug Nelson

Application Architect