"How long before we're online?"

It's the second most frequent question after budget. And the honest answer is: it depends on your scope, but we can outline realistic ranges.

This article describes the standard phasing of a web app project, milestone by milestone, with indicative durations and the concrete example of Panorama Pub (B2B marketplace delivered in 2 months).

The standard phasing

Phase 1 — Scoping (1 to 4 weeks)

What we do

  • User and business needs interviews (1-3 sessions)
  • Functional perimeter definition (MVP, future evolutions)
  • Data modeling (entities, relations, rules)
  • Wireframes of key screens (low fidelity, just to validate the journey)
  • Detailed estimate of planning and costs
  • Risk identification (complex integrations, external dependencies)

Duration by complexity

  • Small project (simple web app, 1-2 roles): 3-5 days
  • Medium project (marketplace, business platform): 1-2 weeks
  • Complex project (multi-role, multiple integrations): 2-4 weeks

The trap to avoid

Skipping scoping "to save time". It's the best way to lose 3-5 times more time in dev. A poorly scoped screen costs 3-10× more to redo than to think through up front.

Phase 2 — Design (2 to 6 weeks)

What we do

  • Visual identity (if not existing): palette, typography, tone
  • Design system: reusable components, inter-screen consistency
  • High-fidelity design of key screens
  • Interactive prototype (Figma) to validate the journey
  • Responsive (mobile, tablet, desktop)
  • Accessibility validation (contrasts, structure)

Duration by complexity

  • Small project: 1-2 weeks (with a simple design system)
  • Medium project: 2-3 weeks (design system + 15-25 screens)
  • Complex project: 4-6 weeks (multi-role, dashboards, varied states)

Optimization

  • If you already have a visual identity (charter, design system), we save 30-50% of design time
  • If you accept a "functional" design based on proven components (shadcn/ui, for example) rather than visual creation from scratch, we save 40-60%
  • If you want a very distinctive "signature" design, expect the opposite: +20-40%

Phase 3 — MVP development (3 to 16 weeks)

What we do

  • Project setup (Next.js, PostgreSQL database, Vercel hosting, CI/CD)
  • Authentication (accounts, sessions, roles)
  • Database schema
  • Front pages (public + logged-in screens)
  • API and business logic
  • Autonomous admin (data and user management)
  • Third-party integrations (payment, email, etc.)
  • Automated tests
  • Technical documentation

Duration by complexity

  • Small project (3-5 entities, 10-15 screens): 3-6 weeks
  • Medium project (5-10 entities, 25-40 screens, simple marketplace): 6-10 weeks
  • Complex project (10+ entities, 50+ screens, multi-role): 12-20 weeks

What slows down dev

  • Delayed decisions: "We'll see later if we want X or Y" → blocks progress
  • Third-party integrations: poorly documented APIs, capricious test environments
  • Mid-dev design back-and-forth: modifying an already-coded screen takes 3-5× more time
  • Slow reviews: if each demo takes 1 week before feedback, we lose 1 week out of 4

Phase 4 — Client QA (1 to 3 weeks)

What we do

  • MVP demo to client
  • Client tests, formulates remarks
  • Fixes and adjustments
  • User testing plan (if applicable)
  • Final scope validation

Duration

  • 1 week for a simple, well-prepared project
  • 2 weeks for a medium project
  • 3+ weeks for a complex project with pilot users

Tip

Preparing a test list during the dev phase makes QA much faster and more thorough. We systematically check each case rather than discovering at random.

Phase 5 — Launch (a few days)

What we do

  • Initial data migration (if applicable)
  • Domain switchover to production
  • Final checks (SEO, security, performance)
  • Client onboarding on the admin
  • User documentation
  • Monitoring and alerting plan

Duration

  • 2-3 days for a classic deployment
  • 1-2 weeks if complex data migration from an existing system

Phase 6 — Post-launch stabilization (4 to 8 weeks)

Often forgotten, it's the critical phase. Once online, real users surface bugs and friction that no scoping could anticipate.

What we do

  • Quick fixes for reported bugs
  • Small UX adjustments (labels, journeys)
  • Performance optimizations based on real traffic
  • Fine-grained monitoring setup

Budget

Plan 10-15% of initial budget for this stabilization phase.

Concrete example: Panorama Pub in 2 months

| Phase | Duration | Detail | |---|---|---| | Scoping | 1 week | Data model, MVP perimeter, wireframes | | Design | 1.5 weeks | Design system + main screens (directory, listing, search, admin) | | MVP dev | 5 weeks | Next.js + serverless PostgreSQL stack, autonomous admin, email integration | | QA | 1 week | Tests, fixes, client validation | | Launch | 2 days | Domain, SSL, monitoring |

Total: 8.5 weeks from concept to production.

What enabled this timeline:

  • Tight MVP perimeter (focus on directory and search, no secondary features)
  • Data model validated early
  • Responsive client back-and-forth (replies within 24-48h)
  • Proven design system (custom shadcn/ui) instead of design from scratch
  • Mastered stack (Next.js + serverless PostgreSQL, Vercel deployment)

Factors that lengthen a project

× 1.5

  • Multiple user roles with fine permissions
  • Non-trivial third-party integrations (1-2 max)
  • Complex business workflows

× 2

  • Multi-tenant (multiple organizations share the platform)
  • Internationalization from the start (i18n)
  • Mobile-first with enriched PWA
  • Data migration from an existing system

× 3 and beyond

  • Internal social networks (timeline, messaging, real-time notifications)
  • Large-scale moderation
  • Advanced analytics reporting (OLAP cubes, complex exports)
  • Multiple integrations with legacy systems

Acceleration levers

Reduce scope

By far the most powerful lever. Don't do 100% of what you imagine, do 60% with excellence and launch.

Panorama Pub could have had, in V1: a supplier area, a matchmaking system, editorial content. We did a directory with a solid admin. The other blocks come next, based on a product already live and real users.

Reuse rather than create

  • Existing design system (shadcn/ui, headlessui) rather than creation
  • Proven components for authentication, payment, email
  • Database schemas inspired by known patterns

Fast client decisions

A project moving at 100% speed on dev but 50% on client decisions goes at 50%.

Going further

To scope your project: run the diagnostic. Lead time estimate within 48h.