Every founder we meet wants the same three things: launch fast, launch cheap, and launch something users love. The uncomfortable truth is that most MVPs miss all three — not because the team is slow, but because the scope was decided by excitement rather than evidence.
After launching ten of our own SaaS products and dozens more for clients, we’ve distilled MVP planning into a repeatable 90-day framework. This is it, end to end.
Step 1: Define the One Job Your Product Does
An MVP isn’t a smaller version of your product — it’s the smallest version of your promise. Write one sentence: “This product helps [who] do [what] without [pain].” If a feature doesn’t serve that sentence, it waits.
Version one should be embarrassing in breadth and impressive in depth. One workflow, done brilliantly, beats ten done adequately.
Step 2: Sort Features With the 90-Day Filter
We run every proposed feature through three questions:
- Does launch fail without it? If users can’t complete the core job, it’s in.
- Can we fake it? Manual onboarding, a shared inbox instead of a ticketing module — concierge versions count.
- Will usage data change the design? If yes, build it after launch, when the data exists.
Typically this cuts a 40-item wishlist down to 8–12 buildable features. That’s a healthy MVP.
Step 3: Architect for Change, Not Scale
You don’t need heavyweight orchestration on day one. You need a codebase you can change quickly when the first 50 users tell you what’s wrong. Boring stacks are fast to hire for, cheap to run, and easy to refactor. Scale problems are good problems — solve them when you have them.
Step 4: The 12-Week Calendar
- Weeks 1–2: Discovery workshops, user flows, clickable prototype.
- Weeks 3–4: Design system, data model, project scaffolding.
- Weeks 5–9: Build in two-week sprints — core workflow first, billing second, settings last.
- Weeks 10–11: Private beta with 10–20 real users, instrumentation, bug triage.
- Week 12: Hardening, launch checklist, go live.
What to Measure After Launch
Ship with analytics or you’ve shipped blind. The four numbers we watch in month one: activation rate, time-to-first-value, week-two retention, and support tickets per active user. Everything on the version-two roadmap should trace back to one of these.
Plan this way and 90 days is not aggressive — it’s comfortable. The discipline isn’t in building fast; it’s in deciding small.
Part of the team building our SaaS products and client platforms.