Looker: Google's Data Analytics Platform

THE MENTAL MODEL: Looker is not just a dashboard tool; it's a data modeling platform that sits on top of your existing database. Think of it as a 'source of truth' factory for your company's metrics. You define a metric like 'revenue' once in its modeling language, LookML, and everyone in the company uses that exact same definition for their analysis.

HOW IT WORKS: Looker connects directly to a company's live SQL database (like BigQuery or Snowflake). Developers write LookML code to define dimensions (attributes like 'customer signup date'), measures (aggregations like 'count of users'), and the relationships between data tables. This LookML model acts as a semantic layer. Business users then use a web interface to explore this modeled data, create visualizations, and build dashboards by pointing and clicking. Looker translates their actions into optimized SQL queries and runs them against the database in real time.

WHEN TO USE IT: Use Looker to empower non-technical users to self-serve their own data questions safely and consistently. It is ideal for establishing a single source of truth for key business metrics across an organization, ensuring everyone uses the same calculations. Its API also allows for embedding analytics and reports into other applications.

WHEN NOT TO USE IT: Looker is not a database; it does not store data. It queries your database live, so it's a poor fit if your underlying database is slow or not designed for analytical workloads. The upfront investment in building a robust LookML model can be significant, making it overkill for small teams or simple, one-off analyses where a direct SQL client would suffice.

ONE CANONICAL EXAMPLE: A company wants to track 'Monthly Active Users' (MAU). Instead of each analyst writing a slightly different SQL query, a data engineer defines the logic for MAU once in the central LookML model. Now, anyone on the marketing, product, or executive team can drag the 'Monthly Active Users' measure into a report, confident they are using the official, vetted definition.