Why 90 days is a strategy, not a deadline

Ninety days is short enough that the market can't move under you and cheap enough that being wrong is survivable. The process below isn't theory — it's the refined average of 20+ MVPs, including our own SaaS products.

Days 1–14: Discovery sprint

We interview you (a lot), sketch the user journeys, and write the one-sentence test the MVP must pass: 'a [user] will pay for [outcome]'. Then the kill list — every feature that doesn't serve that sentence. A typical client wishlist loses 60% of features here, and the product gets better for it.

Output: clickable prototype in Figma, technical architecture, fixed quote and the 90-day calendar with weekly demo dates already booked.

Days 15–70: Build sprints (8 weeks)

One-week sprints, each ending in a demo of working software — not slides. Week 3 sets up CI/CD, staging and analytics while they're cheap. Design runs one sprint ahead of engineering; QA writes tests alongside features, not after.

• Weeks 3–4: skeleton walks — auth, data model, empty screens wired end-to-end.

• Weeks 5–6: the core loop — the one thing the product exists for, polished.

• Weeks 7–8: the money layer — billing/payments, onboarding, admin panel, emails.

• Week 9: hardening — load, security pass, bug bash, boring-and-vital work.

Days 71–83: Private beta

10–30 real users (never friends — friends lie kindly). We watch sessions, fix the top ten friction points, and confirm the signup-to-value funnel works before launch polish. Half our pivots-in-small happen here, at 1/10th the cost of post-launch.

Days 84–90: Launch

The checklist: app-store review buffer (iOS review times still surprise people), landing page with analytics, crash and uptime monitoring wired to alert a human, a support channel someone answers within hours, and a day-one metrics dashboard. Then we hold a 30-minute 'what did we learn' retro — the real product roadmap starts there.

What we refuse to build in an MVP

Custom animations nobody asked for. Admin panels with more than the 5 screens you'll use. 'Enterprise-ready' features before the first enterprise customer exists. Multiple user roles before the first role is validated. Social login variants — email first, always. Every one of these is a story from a real project that taught us the list.

After the 90

The MVP's job is learning velocity. The 3 months after launch decide everything: instrument, interview, iterate. That's why we hand over a metrics dashboard, not just a product.

Have an idea worth testing in 90 days? Book a free scoping call — worst case you leave with a kill list that saves you six months.