What to Build First in a SaaS MVP (and What to Leave Out)

Billing, teams, roles, an admin panel, onboarding. You cannot build them all first. Here is the order that keeps you solvent.

September 2, 2026 · 2 min read · Business

In short: A SaaS product has roughly ten things that feel mandatory before launch. Building all of them takes a year and most will be wrong. The skill is knowing which three are genuinely load-bearing.

Build these three

The thing people would pay for. One workflow, done properly. Not five, done adequately. If a user cannot complete that one job end to end, nothing else matters.
Sign-up and sign-in. Unglamorous, non-negotiable, and much easier to use a well-tested service than to write yourself.
A way to take money. Even if it is a payment link you send manually. "We will add billing later" is how products accumulate users who will never convert.

Fake these

Onboarding. A personal walkthrough on a call teaches you more than an onboarding flow built on guesses, and it costs nothing to change.
The admin panel. Query the database directly. Build the panel when the manual work becomes a daily annoyance — by then you know exactly what it needs.
Analytics dashboards for users. Almost everyone asks for these. Almost nobody opens them twice. Wait for the second request.
Integrations. Build the one a paying customer has asked for by name. Not a list of logos.

Leave these out entirely

Teams, roles and permissions. A large body of work that only matters once you have customers big enough to need it. Single-user until proven otherwise.
Multiple pricing tiers. One price. You do not know enough yet to design tiers, and the wrong tiers teach you nothing.
A mobile app. A responsive web app covers it until someone tells you otherwise with their wallet.
White-labelling, custom domains, an API. Each is a product in itself. None belong in a first version.

The test to apply

For each feature, ask: if this were missing, would a paying customer refuse to pay? If the honest answer is no, it is not MVP. Most features fail this test, including several you feel strongly about.

A reasonable first scope

One workflow. Sign-up. Payment. A manual way for you to see what is happening. Launched to ten real users who gave you a reason to believe they want it. That is an achievable build, and it will teach you more than six months of adding features.

We build SaaS products and the most useful thing we do early is argue for a smaller first version. Tell us what you are building.