Business Intelligence

Metabase

Metabase is self-service BI for teams that want broad access, a real AI assistant, and the flexibility to self-host — without enterprise-tier spend.

What It Is

Metabase is open-source BI with a full dashboarding, metric, and question-building surface. Business users explore data through a visual interface; analysts build reusable Models and Metrics that enforce consistent definitions; power users drop into SQL when they need it. Metabot — Metabase's AI assistant, open-sourced in version 60 — generates charts, writes SQL, analyzes results, and answers questions in natural language.

Why We Chose It

Metabase is the pragmatic middle of the BI market. It's cheaper than Looker, broader than Steep, and more flexible than Lightdash — you can run it fully self-hosted on Cloud Run, use Metabase Cloud, or start free on the Community Edition. The semantic layer (Models + Metrics + Data Studio) matured fast in 2025–2026, and Metabot's AI features — including self-hosted AI with your own Anthropic API key — make it credible as an AI-native BI tool without the cloud-only lock-in of Looker.

How We Use It

Stand up Metabase — Metabase Cloud for managed, or self-hosted on Cloud Run / GKE for full control

Define Models and Metrics so business users get consistent definitions across dashboards and AI queries

Configure Metabot with the client's own LLM provider (Anthropic) for natural-language exploration

Build operational and executive dashboards with email/Slack subscription delivery

Set up row-level and column-level permissions for multi-team access

Integrate dbt metadata via the community dbt-metabase package when the client is on dbt

When Metabase is the right BI tool — and when it isn't

Choose Metabase when:

  • You want self-service BI with a real AI assistant, at a reasonable price
  • You want the option to self-host (Community Edition free, Pro self-host paid)
  • You want Metabot AI features with BYO Anthropic key — available even in self-hosted
  • Your team mixes technical and non-technical users

Choose Looker instead when:

  • You need enterprise LookML governance and embedded analytics at scale
  • Budget is 5k+/month and you are scaling across 50+ users

Choose Lightdash instead when:

  • You're already on dbt and want metrics defined in code
  • You want flat-rate pricing that does not scale with headcount

Choose Steep instead when:

  • The team is entirely non-technical and you want the simplest metric-first UX
  • Budget is under €500/month for a small team