Looker review

Looker is Google’s business intelligence platform for exploring data, building dashboards, and embedding analytics in applications. Its main distinction is LookML: a semantic layer where teams define governed business metrics so people, dashboards, AI agents, and connected tools can use shared definitions.

The product also offers natural-language analytics with Gemini-powered agents, dashboard summaries and data explorations, and self-service tools for creating visualizations, calculations, and explorations. Users can combine local files with LookML models. For product teams, Looker supports embedding through APIs, iframes, and SDKs. Google also lists a FinOps view that unifies billing data across Google Cloud, AWS, and Azure.

This is the paid Looker BI product, not Looker Studio. Several Reddit discussions in the published material are about Looker Studio, formerly Data Studio; their complaints about its report-building interface or data limits should not be treated as evidence about Looker’s features or plan limits.

Looker’s clearest strength is governed metrics. LookML gives teams a shared place to define business logic, which can help keep dashboard results and AI-driven answers tied to the same definitions. That is a better fit for organizations that want controlled, reusable data models than for teams looking only to assemble quick charts.

The feature set also covers more than standard dashboards. The pricing page lists an Embed Edition with higher API allowances, while product materials describe APIs, iframe embedding, and SDKs for putting analytics in applications. Conversational analytics, Gemini-powered agents, and dashboard-level summaries add natural-language ways to explore or explain data. The self-service tools let users create visualizations and calculations and combine local files with modeled data.

There is a plausible fit for teams already comfortable treating analytics logic as code. In one Reddit thread, a commenter said their organization chose Looker for Git versioning and pull-request reviews, and because it fit an existing dbt code culture. That is one user’s account, not a guarantee that Looker will suit every analyst or workflow.

The main purchasing problem is price opacity. The pricing page lists no monthly or annual prices; annual commitments of one, two, or three years are “Call sales,” and monthly pricing is not published. Standard Edition is described for organizations or teams under 50 users, but its listed package includes only 10 Standard Users and 2 Developer Users. Enterprise keeps those user counts but raises query-based API calls from 1,000 to 100,000 per month and administrative API calls from 1,000 to 10,000. Embed lists up to 500,000 query-based and 100,000 administrative API calls monthly. API overage prices are not published.

Conversational Analytics includes 60 million input and 1.2 million output tokens per production instance each month; listed overages are $3 per million input tokens and $20 per million output tokens. Additional user-seat prices are also not published.

LookML’s code-centered approach can be a drawback for teams expecting mostly drag-and-drop modeling. A Reddit commenter moving from Power BI described Looker as more code-heavy, but that is an individual report. Complaints about Looker Studio in other threads concern a separate product and are not a fair basis for judging Looker.

Looker does not publish plan prices, so there is no reliable dollar figure for the jump between tiers. The pricing page lists Standard, Enterprise, and Embed as “Call sales”; monthly pricing is not published. It also lists annual commitments of one, two, or three years, so confirm the commitment and billing terms before comparing quotes.

The tiers mainly differ in API allowances. Standard includes up to 1,000 query-based and 1,000 administrative API calls per month. Enterprise raises those caps to 100,000 and 10,000; Embed raises them to 500,000 and 100,000. Each plan lists one production instance, 10 Standard Users, and two Developer Users. Additional Standard, Developer, and Viewer licenses cost extra, but seat prices are not published.

Conversational Analytics includes 60 million input and 1.2 million output tokens per production instance monthly. Overage rates are $3 per million input tokens and $20 per million output tokens. API overage pricing is not published. Get a quote that spells out seats, API use, token overages, and term length.

Switching to Looker likely takes more than recreating charts: its LookML semantic layer defines governed business metrics, so teams may need to translate existing metric definitions and rebuild dashboards around that model. That is an inference from the product’s documented structure, not a published migration estimate. The published materials do not specify migration tools, export formats, or a data-export process; confirm what can be taken with you before signing.

The same semantic layer is the main switching cost in the other direction. If dashboards, connected tools, and AI workflows rely on LookML definitions, replacing Looker means deciding where those definitions will live and how to keep metrics consistent. Teams with code-led BI workflows may accept that work; teams seeking low-code dashboard building may find the fit less appealing. The published materials include no Reddit posts or comments, so we can’t substantiate claims about why users leave or how difficult migrations have been for them. Ask vendors to map your LookML models, dashboards, and embedded analytics to their equivalents before committing.