Case Study: Instructables achieves faster, more relevant, and customizable search with Elastic Site Search

Instructables Making Search Faster and More Customizable

Instructables, founded at the MIT Media Lab and now part of Autodesk, hosts over 100,000 user-created DIY projects and handles hundreds of thousands of searches daily. Their existing search solutions (Solr and an external provider) returned generic, title-based results, struggled with misspellings and long or specific queries, and couldn’t take advantage of the rich project metadata the team had collected.

They implemented Elastic Site Search, whose language modeling and relevance controls immediately improved results—automatically handling misspellings (even domain-specific terms like “arduino”) and multi-word queries. Instructables indexed metadata such as popularity, step counts, and page views to tune relevance, deployed the same engine across web and mobile so changes propagate instantly, and now use search analytics to find zero-result queries and guide content strategy.

Instructables