Metabase - A tech bet case study

Engineering May 24, 2021

Last week, we decided to suspend a Saas vendor of Business Intelligence Tooling. While the main reason is we can't afford the cost for existing high number of users, we also found that we already built a great ecosystem around our self-host BI tool - Metabase, which became a so successful technology bet for us since we started up that we can accept the risk to rely on Metabase solely.

We started using Metabase very soon, almost at the time this open source announced on Hacker News 5 years ago. At that time, Metabase is not mature enough so we employed a hybrid tactic with Metabase is the default tool for employees and another SaaS tool to backup and compensate Metabase weakness. Along the time, Metabase UX and performance improved and simply made us loving it more, that attract more and more of us learning SQL and working with Metabase, and we also gained experiences to monitor and stabilize metabase operating as well. The positive effect caused by Metabase is, our company populated SQL to a lot of staff and more than 50% employees using Metabase daily. This impressive penetration have a lot of reasonings: free OSS, great features and UX, fast feedback loop and release cycle, etc... But there is a huge win of Metabase is its openness, and I believe our team has utilized it to unexpected use-cases that not really aimed for Metabase in its core.

  1. We consider Metabase as our SQL brain to automate our daily operations, without much interference from technical side. Business team can do it with metabase and a workflow tool.
  2. We made Metabase as an automated reporting tool to Telegram that can produce Metabase's native chart on Telegram as an image, which means we can keep the original visualization of Metabase to any user chat platform, help them gain in-time, expressive and true insights. More than that, it just take less than five minutes to setup this reporting process.
  3. We built Metabase connector to Google Data Studio, help our data/business analysts can get the raw data quickly and seamlessly from Metabase to visualize them on Google Data Studio.
  4. We developed and open sourced a Metabase driver called Http Driver, to help us and others can get data from other sources to Metabase in convenient way via Http apis. In this case, Metabase can visualize coming data or become a nice proxy/playground to test some data api.

Metabase is a BI tool, but at AhaMove, we built a great ecosystem around it, and this made us love and stick more to it, gradually. I think I can say it is a succesful tech bet.