Looker Turned Metrics into Code

Looker mattered because it turned metric definitions into software, not just dashboard settings. Instead of each analyst rebuilding revenue, retention, or conversion inside separate charts, teams could define joins, dimensions, and measures in LookML and reuse them across many reports. That was an early attempt to make business logic version controlled, reviewable, and consistent, which later became the foundation for today’s semantic layer category.

• Before this shift, BI tools mostly let users click together charts or write one off SQL. That made it easy for two dashboards to compute the same KPI differently. The core problem was not storing data, it was storing the meaning of the metric in one reusable place.
  2 sacra 3 google
• Looker’s key innovation was that the modeling layer lived in code, but it was still attached to Looker. As teams added notebooks, reverse ETL, planning tools, and other BI surfaces, that became the next bottleneck. The market then moved toward more portable semantic layers, especially around dbt.
  2 sacra 3 google 4 sacra
• The industry context also matters. Looker and Tableau were both absorbed into larger clouds in 2019, Google for Looker and Salesforce for Tableau. That pushed both products into broader platform strategies, while newer vendors like Omni emerged in the unbundled modern data stack with ex Looker DNA.
  5 google 6 salesforce 7 tableau

This is heading toward metrics becoming shared infrastructure across every data consuming app, not just a dashboard builder. The winning products will be the ones that let a company define business logic once, keep it in code, and serve the same answer into BI, planning, AI, and operational workflows without forcing everything through a single front end.