Engineering
  • MVP
  • next.js
  • startups

A 30-Day MVP Checklist (That Survived Real Founders)

Not a sales deck — the week-by-week cuts we make when scope is screaming and the runway isn't.

Quezt Labs

Quezt Labs team

  • 13 min read
Contents· 8 sections

What this is

A cut list, not a feature buffet. We run 30-day sprints for founders; this is the schedule that survives contact with reality when someone asks for "Uber + Notion + AI" in one month.

Week 0 — Scope or suffer

InOut (for now)
One user loopSecond persona
Email auth or OAuth pick oneSocial + magic + password same week
PostgresMicroservices
Vercel deployMulti-region

Week 1 — Skeleton

  • Next.js App Router + Tailwind + component kit you already know
  • DB schema: minimum tables for the loop
  • Preview deploy day 3 — forces env discipline early
  • loading.tsx / error boundary on the loop route only

Stack default:

LayerPick
FrameworkNext.js 15+
DBPostgres + Drizzle or Prisma
AuthClerk or magic link — not both
EmailResend
HostVercel

Week 2 — Core loop

  • Server Actions or Route Handlers — avoid extra API layer unless mobile client needs it
  • Payments only if money is week-1 critical (Stripe/Razorpay)
  • Admin? A single protected /admin route, not a second app

Week 3 — Ops & trust

  • Error boundaries on user-facing routes
  • Basic analytics (Vercel + one product event)
  • SEO: metadata, sitemap, OG for marketing page
  • Staging vs prod env documented

Week 4 — Launch

  • Lighthouse on marketing + app shell (not obsession, sanity)
  • Runbook: env vars, rollback, who gets paged
  • Handoff: README, demo script, known limitations list

Kill list (say no out loud)

  • Real-time everything
  • Custom design system before users
  • Mobile app and web unless budget doubles
  • "We'll add tests later" on payment code

TL;DR

MVPs die from scope creep, not missing dark mode. One loop, one deploy, one honest demo.