What Is Looker? Google's Data Platform and Business Intelligence Tool

Looker is Google's business intelligence and data platform, acquired by Google in 2020 and now part of Google Cloud. What distinguishes Looker from Tableau, Power BI, and other BI tools is LookML — a code-based semantic layer in which data teams define metrics, dimensions, and relationships centrally. BI users access those definitions through Looker's UI; every visualization uses the same underlying metric definitions regardless of who built it or where it appears.

LookML: The Semantic Layer

LookML is a YAML-based modeling language for describing the data model that Looker users query. It defines:

**Views:** LookML representations of database tables or SQL-derived queries. A view file describes the table's fields — their names, types, SQL expressions, labels, and grouping behavior.

**Dimensions:** Non-aggregated attributes — customer region, product category, order status, date fields. A dimension in LookML is a SQL expression that Looker includes in the GROUP BY clause of the generated query.

**Measures:** Aggregated calculations — revenue, count of orders, average session duration. A measure in LookML is a SQL expression applied to an aggregation function — COUNT, SUM, AVG, MAX, MIN — that Looker wraps in the SELECT clause.

**Explores:** Join configurations that define which views can be queried together and how they join. An explore exposes a set of related dimensions and measures to business users as a coherent exploration surface.

**Derived Tables:** SQL or LookML-based virtual tables built within Looker's semantic layer, materializable to the database for performance. Persistent derived tables (PDTs) are materialized to database tables on schedule; non-persistent derived tables are generated as subqueries.

The semantic layer is the core of Looker's value proposition. When a data engineer defines revenue in LookML as SUM(order_amount) - SUM(return_amount) with the fiscal calendar logic applied, every report that uses the revenue measure applies that exact calculation. There is no divergence between the finance dashboard and the sales dashboard on what "revenue" means. Changing the definition in LookML propagates everywhere instantly.

Looker vs Tableau vs Power BI

**Governance model:** Looker's metric governance is code-first and centrally enforced — it is architecturally difficult for a business user to bypass the LookML definitions and calculate their own version of revenue. Tableau and Power BI allow users to build calculated fields in workbooks, which is powerful for self-service but creates the risk of diverging definitions. Organizations with strong metric governance requirements favor Looker's model.

**Self-service authoring:** Tableau's drag-and-drop authoring on any field in any visual is significantly more flexible for exploratory analysis. Looker's explore interface is constrained to the dimensions and measures the data team has defined in LookML — business users cannot query fields that are not in LookML. This constraint is a governance feature, not a bug, but it means exploratory analysis outside the LookML model requires data team involvement.

**Visualization richness:** Tableau produces more sophisticated analytical visualizations. Looker's visualization library covers standard business reporting but is not designed for advanced analytical storytelling.

**Embedded analytics:** Looker is a strong choice for embedding analytics in products and customer-facing applications. Looker's API and SSO embedding capabilities are mature; organizations building analytics into SaaS products use Looker's embedded analytics more commonly than Tableau or Power BI for this use case.

**Google Cloud integration:** For GCP-native organizations, Looker integrates directly with BigQuery without additional connectivity. Looker Studio (formerly Data Studio) is Google's simpler free BI tool; Looker is the full-featured enterprise platform.

**Price:** Looker pricing is at the higher end of BI tool market — Creator licenses are comparable to Tableau. The LookML investment (data engineer time to build and maintain the model) is a real ongoing cost that is lower with self-service tools but paid instead in governance debt.

Looker Architecture

Looker generates SQL from LookML definitions and sends it to the data warehouse. It does not import data or maintain its own data engine — it is a pure semantic and query generation layer on top of the warehouse. This architecture means Looker's performance is warehouse performance. Organizations with well-optimized BigQuery or Snowflake environments get fast Looker queries; organizations with poorly structured warehouse tables get slow ones.

**Looker API:** Looker's REST API exposes the full functionality of the platform programmatically — running queries, embedding, managing users, accessing LookML metadata. For organizations building data products on top of Looker, the API is a key capability.

**Scheduler and alerts:** Looker supports scheduled delivery of dashboard content to email and Slack, and threshold-based alerts when metrics cross defined values. These capabilities are standard across enterprise BI tools.

When to Choose Looker

Looker is the right choice when:

- Metric consistency and governance are the primary requirements — the organization has historically had diverging metric definitions across tools and teams

- The engineering team is capable of building and maintaining LookML models — the governance model requires ongoing data engineering investment

- Embedded analytics is a requirement — Looker's embedding capabilities are strong

- The organization is GCP-native or BigQuery-primary

Tableau is the right choice when:

- Visualization depth and flexibility are primary requirements

- Self-service exploration by business users across arbitrary data is important

- The organization is not GCP-native and does not have specific metric governance requirements driving toward LookML

Our BI strategy services helps organizations evaluate Looker, Tableau, and Power BI against their specific requirements and build the analytics stack that fits. Contact us to discuss your BI evaluation.

A former Microsoft data architect audits your data foundation, identifies your top priorities, and sends you a written plan. Free. No pitch.