Warehouse Migrations without Breaking Looker (Case Study)

Learn how one company switched from Snowflake to BigQuery without Looker disruptions.

🏔 The Challenge

How does a company migrate warehouses confidently, without worrying about broken Explores and dashboards in Looker? A new warehouse comes with different SQL syntax, schema, and permissions, and data teams want to catch those errors before their customers or executives.

This is the challenge an education-tech company was facing when migrating from Snowflake to BigQuery while leveraging Looker for internal and customer-facing, embedded analytics.

✅ The Solution

This company implemented Spectacles for continuous integration, which automatically tested their Looker environment on the new warehouse. Specifically, Spectacles checked for Looker errors (content validation, data test failures, and SQL errors) on both warehouses and schemas.

They were able to test their Production Looker environments against their new warehouse and proactively learn of errors (whether a result of schema changes, SQL syntax differences, or permissions issues) before pointing their production Looker instances to their new BigQuery warehouse.

🚀 The Upshot

Here's Bryan Kaufmann, Analytics Engineering Manager,

"Anyone doing a warehouse migration with Looker should use Spectacles to validate. It helped us identify both schema and bespoke-SQL differences. We couldn't have successfully done the migration without Spectacles!

The company successfully migrated warehouses without breaking Looker, and currently leverages Spectacles to ensure that their Looker environments are 100% error-free.

Their client-facing, embedded Looker environment is tested with Spectacles daily as a safeguard to ensure customers get clean, accurate reports.

They leverage Spectacles daily on their internal Looker instance to ensure that accidental errors aren’t introduced via PR’s.

About Spectacles

Spectacles helps data-driven organizations reduce development costs, improve velocity, and increase stakeholder trust by testing Looker instances and LookML code throughout the development cycle. Data analysts and engineers at 80+ industry-leading companies rely on Spectacles daily to check their work and uphold their teams’ quality standards and reputation. Spectacles’ customers include an S&P 500 consumer goods company with over 20,000 employees, one of the world’s top-ten largest retailers, and high-growth startups including Canva, Casper, Cargurus, and more.