mvp-builder（MVP 建设者）

MVP Builder

Ship MVPs in 1-2 weeks, not months.

Core Principle

Start with the smallest thing that proves/disproves your riskiest assumption.

MVP = Minimum Viable Product, not Minimum Pretty Product.

MVP Feature Matrix

Categorize every feature idea using this matrix:

Example: Task Management MVP

P0 (Week 1):

• Create task with title/description
• Mark task complete
• View task list
• Basic authentication

P1 (Post-validation):

• Due dates, priorities, reminders

P2 (Future):

• Team collaboration, file attachments, mobile app

Out of Scope:

• Gantt charts, time tracking, integrations

5 MVP Patterns

1. Concierge MVP

Definition: Manually deliver the service before building automation

When to use: Core value is service delivery, not technology

Examples:

• Food delivery → Take orders via WhatsApp, deliver yourself
• AI copywriter → Manually write copy for customers
• Scheduling tool → Coordinate meetings via email

Time: 1-3 days | Validates: People want the service

2. Wizard of Oz MVP

Definition: Interface looks automated, but humans operate it behind the scenes

When to use: Automation is expensive/complex to build

Examples:

• Zapier early days → Manually created integrations
• AI chatbot → Human answers, customer thinks it's AI

Time: 1 week | Validates: Users engage with interface

3. Landing Page MVP

Definition: Explain the product + collect emails (no product yet)

When to use: Testing demand before building anything

Examples:

• Dropbox → Video showing concept
• Buffer → Landing page before code existed

Time: 1-2 days | Validates: People sign up for waitlist

4. Single-Feature MVP

Definition: One feature that solves one problem

When to use: One feature delivers 80% of value

Examples:

• Twitter → Just post 140-character updates
• Instagram → Just share photos with filters
• Stripe → Just accept credit card payments

Time: 1-2 weeks | Validates: People use the core feature

5. Piecemeal MVP

Definition: Combine existing tools instead of building custom

When to use: You can cobble together a solution with existing tools

Examples:

• Typeform + Airtable + Zapier + Stripe
• No-code tools: Webflow, Bubble, Retool

Time: 2-5 days | Validates: Workflow works end-to-end

Tech Stack Decision Tree

Is this a web app?
├─ YES
│  ├─ Need real-time?
│  │  ├─ YES → Next.js + Supabase + WebSockets
│  │  └─ NO → Remix + PostgreSQL
│  └─ Simple CRUD? → Supabase + React
└─ NO
   ├─ Mobile app? → React Native + Expo
   ├─ API only? → Express + PostgreSQL
   ├─ Chrome extension? → Vanilla JS + Chrome APIs
   └─ CLI tool? → Node.js or Python

Recommended MVP Stack:

• Frontend: React + Vite or Next.js
• Backend: Supabase (Postgres + Auth + Storage) or Express.js
• Styling: Tailwind CSS + shadcn/ui
• Hosting: Vercel/Netlify (frontend) + Railway/Fly.io (backend)
• Auth: Clerk or Supabase Auth
• Payments: Stripe Checkout

Speed-Focused Tools:

• Supabase (backend in 10 min)
• Vercel v0 (UI generation)
• Clerk (auth in 5 min)
• Stripe Checkout (payments in 20 min)

Anti-Patterns to Avoid

❌ Building features "just in case" → Build when 3+ users request ❌ Perfect design before launch → Ship functional, ugly is fine ❌ Over-engineering architecture → Monolith is fine for MVP ❌ Custom authentication → Use Clerk, Auth0, or Supabase Auth ❌ Building admin panel first → Use database GUI (Retool/Supabase) ❌ Mobile app before web → Web first, always ❌ Scaling for 1M users → Build for 10 users, scale when needed

MVP Quality Standards

Must Have ✅

• Core feature works end-to-end
• Basic auth (login/signup)
• Deploys without crashing
• Mobile-responsive (doesn't need to be beautiful)
• Basic error handling

Can Skip ⏸️

• Perfect UI/UX
• Email notifications
• Advanced features
• Analytics dashboard
• Error monitoring (add after validation)
• Tests (add after product-market fit)

Build Process

Week 1: Build Core Feature

Days 1-2: Design + Setup

• Sketch 3 screens (paper/Figma)
• Set up repo + database (3-5 tables max)
• Deploy infrastructure

Days 3-4: Build

• Authentication (use library, don't build)
• Core feature (1 user flow only)
• Basic UI (use component library)

Day 5: Polish + Deploy

• Fix critical bugs
• Deploy to production
• Share with 5 friends for feedback

Week 2: Validate

Days 1-2: Iterate

• Fix issues from Week 1 feedback
• Add 1-2 critical missing features

Days 3-4: Get Users

• Share on social media
• Post in relevant communities
• Email 20 people personally

Day 5: Analyze & Decide

• Did 10+ people sign up?
• Did 3+ people use it 3+ times?
• Are 1-2 people willing to pay?

Decision Point:

• ✅ Yes to above → Build P1 features
• ❌ No → Pivot or kill project

Launch Checklist

Pre-Launch (1 hour):

• Core user flow works end-to-end
• Sign up + login works
• Deployed with HTTPS
• Privacy policy + ToS (use generator)

Launch Day:

• Post on Twitter/LinkedIn with screenshot
• Share in 3-5 relevant communities
• Email 20 people personally

Week 1 Post-Launch:

• Reply to every piece of feedback
• Fix critical bugs within 24 hours
• User interview with 3-5 early users

Success Metrics

MVP validation criteria:

• 10+ signups in first week
• 3+ people use it 3+ times
• 1-2 people willing to pay

If you hit these: You might have something. Build P1 features.

If you don't: Pivot or move on. Don't invest months in something nobody wants.

Real MVP Examples

Successful minimal starts:

• Airbnb: Photos of their apartment + PayPal link
• Stripe: Just a form to collect card details
• Dropbox: Video demo before building product
• Uber: iPhone app in San Francisco only
• Facebook: Harvard students only, basic profiles

All started much smaller than you think.

Related Resources

Related Skills:

• product-strategist - For validating product-market fit
• frontend-builder - For building the UI
• api-designer - For backend API design
• deployment-advisor - For deployment decisions
• go-to-market-planner - For launch strategy

Related Patterns:

• META/DECISION-FRAMEWORK.md - Platform selection decision tree

Related Playbooks:

• PLAYBOOKS/build-mvp.md - Step-by-step MVP build procedure (when created)

Quick Reference

For experienced users:

• Identify riskiest assumption
• Choose MVP pattern (Concierge/Wizard/Landing/Single-Feature/Piecemeal)
• Categorize features (P0/P1/P2/Out of Scope)
• Build P0 only (1-2 weeks max)
• Ship ugly but functional
• Validate with 10+ users
• Decide: build P1 or pivot