The exact process we use to ship web apps in 2-8 weeks. From idea to production.
By David Friedman, Founder of AppBrewers
I have shipped 80+ web apps in the last five years. Some took 2 weeks. Some took 3 months. The difference was never the idea -- it was the process. Here is the exact process we use at AppBrewers to ship production web apps in 2-8 weeks.
Step 1: Validate Before You Build (Week 1)
The 100 Euro Validation Test
Before writing code, spend 100 Euro and one week to answer three questions:
- Does anyone want this? -- Create a landing page with a waitlist. Run 50 Euro of ads. If fewer than 10 people sign up, kill the idea.
- Will they pay? -- Email the waitlist. Offer a pre-order at 50% off. If fewer than 3 people pay, the problem is not painful enough.
- Can I build this? -- Map the core feature. If it requires more than 3 screens, it is too big for v1.
Tools we use:
- Carrd or Typedream for landing pages (0-19 Euro)
- Brevo or Mailchimp for email (free tier)
- Meta Ads or Google Ads for traffic (50 Euro test budget)
Step 2: Design the Core Loop (Week 1)
Every successful web app has a core loop -- the 3-step action users repeat:
| App | Core Loop |
|---|---|
| Airbnb | Search -> Book -> Review |
| Uber | Request -> Ride -> Pay |
| Slack | Join -> Message -> Notify |
| Notion | Create -> Edit -> Share |
Your job: Define your core loop in one sentence. Everything else is a distraction.
Example for a booking app:
Browse services -> Select time -> Confirm booking
If your v1 does not include this loop, you are building too much.
Step 3: Choose Your Stack (Week 1)
For Non-Technical Founders
| Tool | Best For | Limitation |
|---|---|---|
| Webflow | Marketing sites, simple apps | 29 Euro/month, limited backend |
| Bubble | MVPs, marketplaces | Performance ceiling at scale |
| Softr | Internal tools, portals | Limited customization |
| Framer | Design-focused landing pages | Not for complex apps |
For Technical Founders
| Stack | Speed | Scalability | Cost |
|---|---|---|---|
| Next.js + Firebase | Fast | High | 50-200 Euro/month |
| Next.js + Supabase | Fast | High | 50-300 Euro/month |
| Laravel + Vue | Medium | High | 30-150 Euro/month |
| Django + React | Medium | High | 40-200 Euro/month |
Our recommendation for 2026: Next.js 15 + Firebase. Server components, real-time database, auth, and hosting in one ecosystem.
Step 4: Build the MVP (Weeks 2-5)
The 3-Screen Rule
Your MVP should have exactly 3 screens:
- Landing/Auth -- Sign up, log in, password reset
- Core Feature -- The one thing users pay for
- Settings/Billing -- Profile, subscription, logout
Everything else is v2.
Week-by-Week Breakdown
| Week | Task | Deliverable |
|---|---|---|
| Week 2 | Auth + Database | Login works, data stores |
| Week 3 | Core Feature | The main user action works end-to-end |
| Week 4 | Billing + Admin | Stripe subscriptions, basic dashboard |
| Week 5 | Polish + Deploy | Bug fixes, performance, production deploy |
Step 5: Launch to 10 Users (Week 6)
The Soft Launch Checklist
- [ ] Core loop works without bugs
- [ ] Stripe charges successfully
- [ ] Password reset email arrives
- [ ] Mobile responsive
- [ ] Google Analytics installed
- [ ] Error tracking (Sentry) active
- [ ] Privacy policy and ToS pages live
Where to Find First Users
- Your waitlist -- Email everyone who signed up during validation
- Reddit -- Post in relevant subreddits
- Twitter/X -- Share your launch thread
- Product Hunt -- Schedule for a Tuesday at midnight PST
- Indie Hackers -- Post a milestone story
Step 6: Iterate Based on Data (Weeks 7-8)
Metrics to Track
| Metric | Target | What It Tells You |
|---|---|---|
| Signup conversion | >20% | Is your landing page clear? |
| Activation rate | >40% | Do users complete the core loop? |
| Retention (Day 7) | >20% | Is the product sticky? |
| NPS | >40 | Do users love it enough to recommend? |
| Churn (monthly) | <5% | Is pricing right? |
What to Build Next
Based on data, not gut:
- Low activation -> Simplify onboarding
- Low retention -> Add notifications or reminders
- High churn -> Interview churned users
- High NPS -> Build referral program
Cost Breakdown
| Stage | DIY (No-Code) | DIY (Code) | Agency (AppBrewers) |
|---|---|---|---|
| Validation | 100 Euro | 100 Euro | 500 Euro |
| Design | 0 (template) | 0 (wireframe) | 1,500 Euro |
| MVP Build | 50/month (Bubble) | 3,000-8,000 Euro | 8,000-15,000 Euro |
| Hosting | 29/month | 50-200/month | Included for 60 days |
| Launch | 0 | 0 | Included |
| Total v1 | 200-500 Euro | 3,500-9,000 Euro | 10,000-16,000 Euro |
When to choose each:
- No-code -- Testing an idea, non-technical founder, budget <1,000 Euro
- DIY code -- Technical founder, time to learn, budget 3,000-10,000 Euro
- Agency -- Need it fast, want it right, budget 10,000+ Euro
Common Mistakes
Building Too Much in v1
Founders want AI, analytics, and integrations in the MVP. The result: 6-month build, 40,000 Euro spent, zero customers.
Fix: Strip to 3 screens. Launch in 4 weeks.
Ignoring Mobile
60% of web traffic is mobile. If your app is not mobile-responsive, you lose 60% of users.
Fix: Design mobile-first. Test on actual phones.
No Billing in v1
Founders say we will add billing later. Later never comes. If you do not charge from day one, you do not know if anyone will pay.
Fix: Add Stripe in week 4. Even if it is just a 5 Euro/month plan.
Perfectionism
Just one more feature before launch. This kills more startups than bad ideas.
Fix: Launch when it is embarrassing. Fix in production.
Need Help Building?
We have shipped 80+ web apps. If you want it done right, fast:
Originally published on the AppBrewers Blog.
Top comments (0)