Open source business intelligence for people who don't have time to become analysts.
Above: the famously dotted “M.” Metabase calls it tasteful and inspired-yet-not-threateningly-avant-garde. We call it the logo that launched a million dashboards.
Somewhere right now, a marketing manager who has never written a line of SQL is dragging a few fields into a box and getting back a chart that, two decades ago, would have required a ticket, an analyst, and a week of waiting. That box is Metabase. It runs quietly inside more than 50,000 companies - some self-hosting the free open source edition, others on Metabase Cloud - and it has one stubborn job: get the answer to the person who has the question.
Metabase is not the flashiest name in analytics. It is, however, one of the most used open source projects in the category, and it got there by refusing to make data feel like a privilege. Connect a database, point at a table, ask a question. The tool handles the translation.
“Metabase is open source analytics that makes it easy for your teams and customers to explore and work with data.” — Metabase, on what it does for a living
For most of the 2010s, the deal was lopsided. Companies sat on warehouses full of data and a tiny number of people who could actually query it. Everyone else filed requests and waited. The expensive enterprise BI tools promised to fix this, then mostly priced themselves into the hands of - you guessed it - the same small group of specialists.
Co-founder Sameer Al-Sakran had felt the friction directly. At his earlier startup, scaling data-driven decisions kept running into the same wall: the gap between abundant data and tools that were either too technical or too costly for the people who needed answers. The irony was hard to miss. The more data a company collected, the further most of its employees drifted from it.
The more data a company hoarded, the fewer of its people could actually use it. Metabase exists to flip that ratio. — The central tension
Metabase began in 2014, not as a grand plan but as a side project inside Expa, the venture studio. Sameer Al-Sakran and Allen Gilliland were building an internal way for non-technical people to poke at datasets without bothering an engineer. It worked a little too well. People wanted it. So in 2016 they open-sourced it, and the project did what good open source tends to do - it spread on its own.
The bet underneath was contrarian for the time: that the path to a serious BI business ran through giving the core product away. Make it genuinely easy, make it free to self-host, and trust that adoption would create demand for hosting, support, and enterprise features. It was the opposite of the locked-down, sales-gated model that ruled the category.
“I'm super excited to announce that Metabase has raised a $30M Series B.” — Sameer Al-Sakran, Co-Founder & CEO, 2021
Sameer Al-Sakran and Allen Gilliland start building an easy way for non-technical people to query data inside the Expa venture studio.
Metabase is released as an open source project. Community adoption and contributions take off.
Insight Partners leads the round, with Expa and NEA participating, bringing total funding to roughly $42.5M.
Metabase deepens embedded analytics, ships an Embedded Analytics SDK, adds Data Studio, and introduces AI-assisted querying with Metabot.
Metabase's design trick is that it serves two very different people without making either feel out of place. There's a visual query builder for the person who thinks in questions, and a full SQL editor for the person who thinks in joins. Both end up at the same place: a chart, a dashboard, an answer worth sharing.
It connects to more than 20 data sources - Snowflake and Databricks among them - and it does the unglamorous governance work too. Data sandboxing enforces row and column-level security, so a sales rep sees only their own deals, not the whole company's revenue. For software companies, white-label embedding drops Metabase analytics straight into their product, no Metabase branding in sight, with the SDK and Modular Embedding handling the integration.
Drag fields, filter, group, summarize - build a question without writing SQL.
For analysts who want full control, with parameters, snippets, and drill-throughs.
White-label dashboards inside your own app, scoped per customer in multi-tenant setups.
Row and column-level security, SSO (SAML/JWT), and audit logs for compliance.
Two doors, one room. The analyst writes SQL; the marketer drags and drops. They meet at the same dashboard. — On Metabase's design
Open source projects can claim popularity; Metabase can show it. Tens of thousands of companies run it, a $30M Series B validated the open-core strategy, and an active GitHub community keeps the core moving. The estimates below come from public reporting and the company's own figures.
Bars scaled to ~$42.5M cumulative. Investors: Insight Partners (lead), Expa, NEA. Reported ARR sits around $13.4M.
“Metabase connects to 20+ data sources, with embedding to fit any framework.” — From the Metabase product docs
Strip away the feature list and Metabase is arguing one thing: that data access should not require a job title. Its open-core model is the mission made operational. The core is free because the point is reach. The paid tiers - Cloud, Pro, Enterprise - exist for the teams that need managed hosting, deeper governance, and white-label embedding, not as a tollbooth on the basics.
It is a quietly democratic stance in an industry that historically charged the most to the people who could afford analysts anyway. Metabase would rather a thousand non-technical employees ask their own questions than one specialist answer all of them.
Every company now wants to be data-driven and AI-powered, which mostly means the gap between people and their data matters more than ever. Metabase's response is consistent with its origins: Metabot brings AI-assisted querying into the same friendly box, and Data Studio adds transformation and a semantic layer so the answers stay trustworthy. The bet is that natural-language questions only help if the plumbing underneath is governed and the access is broad.
Back to that marketing manager with the chart. A decade ago she would have filed a ticket and waited a week. Today she dragged three fields and had her answer before the coffee cooled - and tomorrow she'll probably just ask in plain English. The data didn't change. The door did. That's the whole point of Metabase: it keeps widening the door, and it would rather you forget there was ever a wall.