The Problem with LookML
Blog post from Sigma
Reed Rawlings, a solution engineer and former Looker analyst, discusses the challenges of managing LookML, Looker's proprietary modeling language, which can lead to a bloated and unwieldy codebase as business users demand increasingly complex analytics. LookML requires specialized knowledge, making it difficult to maintain, and as the codebase grows, validation processes can slow down and even crash systems. Despite Looker's introduction of a new feature to improve validation speed, issues remain. Rawlings contrasts this with Sigma's no-code approach, which allows business users to directly interact with data without needing an intermediary modeling layer, reducing redundancy and making data transformation more accessible while leveraging cloud data warehouse capabilities for traceability and documentation.