Tableau → Metabase
What it saves, what actually moves, what you rebuild — and the thing that catches people.
What it costs, and what it saves
Tableau Cloud, as listed on its pricing page on 2026-10-03: an annual contract, billed annually, with at least one Creator licence. Standard is $75 per Creator, $42 per Explorer and $15 per Viewer per user/month; Enterprise is $115, $70 and $35. Tableau Server and the Cloud+ edition are quoted by sales. Metabase, as listed the same day: the self-hosted Open Source edition is free for unlimited users under the AGPL. Cloud Starter is $100/month with 5 users included and $6 per extra user. Pro is $575/month with 10 users included and $12 per extra user, on cloud or self-hosted. Enterprise starts at $20,000/year, and yearly billing saves 10%. Worked example from those prices: 2 Creators, 3 Explorers and 20 Viewers cost $576/month ($6,912/year) on Tableau Standard. The same 25 users cost $220/month on Metabase Cloud Starter, or $0 in licences self-hosted, plus a server and a PostgreSQL application database.
Moves cleanly
Data that already sits in a database Metabase supports stays where it is, and Metabase queries the same tables live. Supported databases include PostgreSQL, MySQL, MariaDB, SQL Server, Oracle, Snowflake, BigQuery, Redshift, Databricks, ClickHouse and Athena. SQL you have already written against those databases can be pasted into Metabase's SQL editor, and saved questions can then back new dashboards.
You rebuild
You rebuild workbooks, dashboards, calculated fields, Level of Detail expressions (FIXED, INCLUDE, EXCLUDE), parameters, data blends, extracts and permissions. Metabase has no Tableau importer, and its own YAML import and export (serialization) is a Pro and Enterprise feature. Tableau's visual layouts carry over only as screenshots to copy from.
What Tableau costs you today
Per user per month, billed annually, by role: Creator about $75, Explorer about $42, Viewer about $15. A department with 5 creators, 20 explorers and 100 viewers lands near $30,000 a year. Tableau Server or Cloud is required for sharing, and the viewer tier — the one you need most of — is the one that multiplies. Figures checked 2026-08-04.
What actually holds you in
Moderate, and concentrated in the workbooks. Data connections are configuration and move easily. What does not move are the .twb workbooks themselves — calculated fields, table calculations, level-of-detail expressions and parameter logic are Tableau-specific and have no export path into another tool. A migration rebuilds every dashboard by hand. Organisations that keep their business logic in dbt models and use Tableau only for presentation migrate far more cheaply than those with years of calculation embedded in workbooks.
What you are moving to
Metabase's defining quality is that people who do not write SQL get answers from it unaided — a question builder that reads like English, sensible defaults, and dashboards that assemble in minutes. AGPL-3.0 open-source edition, self-hostable in a single JAR, with paid cloud and enterprise tiers above it. Around 48k stars. For the very common case where a company needs a hundred people to read dashboards and five to build them, this removes the per-viewer bill entirely and is often better received than the tool it replaces.
Open-source edition free and self-hostable. Cloud and enterprise tiers are paid.
Metabase strengths
- Non-technical staff genuinely use it without training
- No per-viewer cost when self-hosted — the main saving
- Single JAR deployment; very simple to run
- Question builder plus full SQL when you need it
What you give up
- Weaker than Tableau for complex statistical and geospatial work
- Some governance features sit in the paid enterprise tier
- Performance depends heavily on the underlying warehouse
- AGPL-3.0 needs reading if you embed it in a product
The migration, step by step
- 1Take an inventory of your workbooks on Tableau Cloud or Server. List the dashboards people still open, and for each one note its data sources, whether it uses a live connection or an extract, and its calculated fields.
- 2Download each workbook as a .twbx and unzip it. The .twb inside is XML, so you can read and list the calculated-field formulas. Any .hyper extracts can be read with Tableau's Hyper API (Python, C++ or Java) and loaded into a database.
- 3Deploy Metabase with PostgreSQL as its application database instead of the default H2, which Metabase's docs say to avoid in production. Connect your warehouse with a read-only user.
- 4Rebuild the shared metrics first. Rewrite each calculated field as a Metabase custom expression or as SQL, turn LOD expressions into SQL subqueries or window functions, and save shared logic as models so every dashboard reuses one definition.
- 5Run both tools in parallel for one reporting cycle. Publish each Metabase dashboard next to its Tableau version and reconcile the headline numbers (totals, counts, period-over-period changes) until they match.
- 6Cut over one dashboard at a time. Recreate scheduled deliveries as Metabase dashboard subscriptions by email or Slack, set the Tableau workbook to read-only, and reduce Creator, Explorer and Viewer licences at the annual renewal.
The gotcha
Tableau can combine data sources in ways Metabase cannot. Tableau's docs describe data blending as 'a method for combining data from multiple sources', and packaged workbooks often carry .hyper extracts or local Excel and CSV files. Metabase's join docs say 'You can only pick tables and saved questions that are from the same database as your starting data'. Hyper is not one of Metabase's supported databases. CSV uploads go only to PostgreSQL, MySQL, Snowflake, Redshift or ClickHouse Cloud, with a 50 MB limit per file. So for any dashboard built on a blend, an extract or a spreadsheet, you first have to load its data into one supported database. Calculations need care as well. Metabase's Offset window function is unavailable on MySQL/MariaDB, ClickHouse, MongoDB and Druid, so period-over-period logic carried over from Tableau may have to be rewritten as SQL.
When to stay on Tableau
Stay if your analysts depend on Tableau Desktop and Prep Builder, on LOD expressions and blends across many sources, on Tableau Pulse, or on the Data Management and Advanced Management features in the Enterprise edition. Metabase is built around SQL and click-to-explore questions, so teams that do most of their analysis in heavy desktop authoring will notice what is missing.
More Layer 5 migrations
Figures verified 2026-08-04 against vendor pricing pages. Prices change and migrations differ by estate — treat the cost delta as a starting model, not a quote. Rankings and recommendations here are merit-only; affiliate income never changes a verdict. See our methodology.