dbt Cloud's free tier gives you a web-based IDE, scheduled runs, and documentation generation. Write SQL models that transform raw data into clean, business-ready tables. Start with simple models: daily revenue, active users, conversion rates.

Before building dashboards, agree on metric definitions. What counts as an 'active user'? How do you calculate MRR? How do you handle refunds in revenue? Write these as dbt metrics or in a shared document. Inconsistent metrics are the #1 source of analytics distrust in every company.

Connect Metabase to your data warehouse and build dashboards on top of your dbt models. Metabase's question builder lets non-technical team members explore data without writing SQL. Create a company dashboard, a sales dashboard, and a product dashboard to start.

Use dbt tests to validate your analytics: check that revenue sums match Stripe, that user counts match your app database, and that no critical columns are null. Failing tests alert you before bad data reaches dashboards — because nothing kills trust in analytics faster than wrong numbers.

The goal of analytics engineering is self-serve: your team gets answers without asking you. Use dbt docs to create a searchable data catalog and add descriptions to every Metabase dashboard explaining what it shows, what it doesn't, and how to interpret it.