Skip to content

Glossary

Multi-tenant

One application serving many customers (tenants), where data and permissions are isolated per organisation — not one database per customer from day one.

What does it mean?

In a multi-tenant system, customers typically share the same app and database, but each tenant only sees its own data. Isolation can use org_id in queries, RLS, separate schemas — or in extreme cases separate databases.

It is the architecture choice that often decides whether a SaaS product can scale sales without rewriting auth and the data model mid-growth.

When it is the right focus

  • You sell to companies, not only individuals

    Organisations, roles, and invites show up quickly.

  • Multiple brands or white-label

    Same platform, different customers who must not see each other.

  • You outgrow the one-account MVP

    When the first enterprise customer asks for SSO and audit.

  • Billing per organisation

    Stripe Customer + subscription belong to the tenant — not a random user.

Short answers

  • Is multi-tenant the same as multi-user?

    No. Multi-user is several people. Multi-tenant is several isolated customers/organisations.

  • Do we need it in v1?

    If you know you sell B2B with org accounts: yes, at least the org model. If you are B2C single users: wait.

  • Where does RLS fit?

    RLS is a strong way to enforce tenant isolation in Postgres — see our RLS entry.

Outgrowing single-tenant?

Let’s design the tenant model you can own.

We cut scope to what you need now — and prepare what you know is coming.