The 3 Approaches

Shared everything (single DB, shared tables with tenant_id column).

Shared DB, separate schemas per tenant.

Isolated (database per tenant, or even server per tenant).

Shared Everything

Pros: cheapest to build and run. Easy to develop new features.

Cons: security risks (bug leaks data across tenants), noisy neighbors (one tenant slows others), harder compliance.

Best for: early stage, SMB customers, cost-sensitive.

Separate Schemas

Same database, but each tenant has own schema.

Pros: some isolation, still cost-efficient, easier backups per tenant.

Cons: schema migrations across all tenants, harder cross-tenant analytics.

Best for: mid-market, moderate compliance needs.

Isolated (DB per Tenant)

Each tenant has own database.

Pros: complete isolation, easy compliance, can run different versions per tenant.

Cons: expensive at scale, complex operations, harder to update.

Best for: enterprise, regulated industries, high-value customers.

Row-Level Security

PostgreSQL RLS: enforce tenant boundaries at DB level.

Even if application bug – DB won’t leak data.

Must-have for shared everything approach.

Tenant Context

Every request needs tenant identification.

Options: subdomain (acme.yourapp.com), URL path (/t/acme), header (X-Tenant-Id).

Best: JWT with tenant_id in claims.

Data Isolation

Beyond database: file storage, cache, search index.

Each layer needs tenant awareness.

Common bug: cache without tenant prefix leaks data.

Performance Isolation

Noisy neighbor problem in shared architectures.

One tenant queries millions of records = slow for all.

Solutions: query limits per tenant, resource quotas, priority queues.

Backup Strategy

Shared: whole DB backup. Restore = affect everyone.

Separate schema: can restore individual schemas.

Isolated: per-tenant backups. Enterprise expects this.

Migration Between Approaches

Shared → Isolated: hardest. Basically new architecture.

Isolated → Shared: rare but happens for cost reduction.

Plan for growth: start where you are, plan for the future.

Compliance Considerations

GDPR: right to delete requires tenant data isolation.

HIPAA: often requires isolated infrastructure.

SOC 2: shared can work with proper controls.

Our Recommendation

MVP: shared everything with RLS.

Growth: separate schemas for enterprise tenants.

Scale: isolated for top-tier customers.

Different customers, different approaches. Don’t over-engineer early.

Based on Real Projects

This guide is based on our work with:

• Platforma365
• ProperFlow

Further Reading

If this guide helped you, you might also want to read our comprehensive guide on Custom SaaS Development.