Insights
What ships in a 3-week SaaS MVP
A SaaS MVP is finished when a customer you have never spoken to can sign up, do the thing they pay for, and receive an invoice. Three weeks is enough for that if the first release is kept to one paid workflow.
What does a 3-week plan look like?
The schedule below is an example, not a commitment. Reorigin plans a working SaaS MVP in 1 week and the production release in 1–3 weeks, and three weeks is a common shape when there is one paid workflow and no legacy data to migrate. The proposal after discovery sets the real dates.
| When | Focus | What exists at the end |
|---|---|---|
| Week 1, days 1–2 | Discovery and tenancy | Written scope, data model, tenant isolation with a test that proves it |
| Week 1, days 3–5 | Accounts, roles, and the paid workflow | A working MVP: sign-up, invitations, owner and member roles, one end-to-end path a customer pays for |
| Week 2 | Billing and onboarding | Plans, subscription, invoice, self-serve first run |
| Week 3 | Hardening and launch | Production deploy, monitoring, handover notes |
What has to be in the first release?
Tenant isolation comes first because it is the most expensive thing to retrofit. Every row belongs to a tenant, every query is scoped to one, and a test proves a user from one tenant cannot read another's data. The options are compared in multi-tenant SaaS architecture decisions.
Accounts and roles follow: sign-up, invitation, and at least an owner and a member role. Then the one workflow customers pay for, built end to end. Billing is part of the MVP, not a later phase, because a product that cannot charge has not tested its own business model. Self-serve onboarding closes the loop: a second user should reach first value without a call.
What waits for version two?
A white-label domain for every tenant, an integration marketplace, enterprise single sign-on, usage-based pricing with metering, and a second paid workflow are all reasonable features. None of them is needed to learn whether customers will pay for the first one. Each is cheaper to add to a product with real tenants than to guess at before launch.
What moves the schedule?
Three things push work past the three-week release into a later phase: an integration with a system that already owns the record, a compliance requirement that fixes the hosting region and the audit trail, and migrating data from a tool the first customers already use. All three are found in discovery. The band and what moves it are on SaaS platform development.
What does the customer have to bring?
A three-week schedule assumes decisions arrive within a day. That needs one person on your side who can say yes or no to scope, and access to at least one real future customer who will try the paid workflow in week one. It also needs the accounts the product depends on opened in your company’s name before the build starts: the cloud provider, the payment processor, the email sender, and the domain. Waiting on a payment provider’s verification is the most common reason a launch slips, and nobody on the engineering side can speed it up.
Which stack fits a schedule this short?
One the team has shipped before. A three-week plan has no room to learn a framework. For a product with a mobile client we use Flutter, Go services, and PostgreSQL, the stack behind RecycleOrigin, described in Flutter, Go, and PostgreSQL: the stack behind a multi-role product. For a browser-only product the mobile layer drops away and the rest stays the same. The database decision is the one that is expensive to reverse, which is why tenancy is settled in the first two days.
How do you know the MVP worked?
Decide the measure before the build. For a first SaaS release it is usually one of three: a customer completed the paid workflow without help, a second user from the same tenant was invited and got in, or an invoice was issued and paid. If none of those happens in the first weeks after launch, the next release should fix that. It should not add features.
See SaaS platform development for the band and startups if this is your first product.
Sources and further reading
- SaaS Lens, AWS Well-Architected Framework — tenant isolation and onboarding guidance.
- Row security policies — PostgreSQL documentation, for shared-schema tenant isolation.
- Stripe Billing documentation — one common way to implement plans, subscriptions, and invoices.
- Flutter architectural overview — how one codebase targets mobile and web clients.
FAQ
What is included in a 3-week SaaS MVP?
Tenant isolation, accounts and roles, one paid workflow, subscription billing, self-serve onboarding, and a monitored production deploy with handover notes.
How much does a SaaS MVP cost?
Reorigin plans a production SaaS MVP at $1,000–$30,000 in 1–3 weeks. That is a planning band, not a quote. The proposal comes after discovery.
Should billing be part of the MVP?
Yes. A product that cannot charge has not tested its business model. Plans, a subscription, and an invoice belong in the first release.
What should be left out of a SaaS MVP?
White-label domains per tenant, an integration marketplace, enterprise single sign-on, metered usage pricing, and a second paid workflow. Each is cheaper to add once real tenants exist.