How much does a custom web system cost
The question everyone asks
“How much does it cost to build a system?” is the first question in any sales meeting. And the most common answer is “it depends”. Which is true, but doesn’t help anyone.
I’m going to share real numbers. Investment ranges based on projects I’ve actually delivered, with concrete examples. This isn’t a generic blog pricing table — it’s what I charge and what the market charges for development with a senior dev.
Ranges by project type
Landing page or corporate website
R$3,000 to R$8,000. Static pages, optimized for SEO and performance. Includes design, development, responsiveness, and deployment.
In practice: GPM2 (corporate site for a tax consulting firm) and Tok Final (site for a door installation company) fall in this range. These are projects with 5-10 sections, professional design, local SEO configured, and Lighthouse 95+.
Timeline: 1 to 3 weeks.
What’s included: page design, development in Astro (or equivalent stack), responsiveness (mobile, tablet, desktop), basic SEO (meta tags, Schema.org, sitemap), deployment on Vercel/Netlify, and one round of post-delivery adjustments.
What’s not included: content (text and photos are the client’s or partner agency’s responsibility), branding/logo, and monthly maintenance.
What makes it vary within the range: number of pages, design complexity, whether there are elaborate animations, and whether it needs extra functionality like a form integrated with a CRM.
Simple webapp (1-2 panels, basic CRUD)
R$8,000 to R$20,000. System with authentication, registration, listing, filters, and one or two user roles.
Mariah (order management with a financial dashboard) fits here. One main panel with order CRUD, financial tracking with period filters, and an installable PWA on mobile.
Timeline: 3 to 6 weeks.
What’s included: authentication (login/registration), modeled database, responsive interface, deployment, and basic documentation.
Example projects in this range: appointment scheduling panel, simple inventory management system, metrics dashboard with a single data source, product catalog with an admin area.
Full webapp (multiple panels, integrations)
R$20,000 to R$50,000. System with 3+ user roles, external integrations, reports, PWA.
AutoPars fits here. Marketplace with 3 panels (buyer, seller, admin), 5 integrations (Asaas, Melhor Envio, Zapi, Resend, Cloudflare Stream), admin dashboard with GMV metrics, and PWA for sellers.
Timeline: 6 to 12 weeks.
What’s included: everything from the simple webapp + multiple roles with access control (RLS on Supabase), configured and tested integrations, dashboard with metrics, PWA, and an Astro capture landing page when needed.
FitPlan (gym platform with 6 panels) sits at the upper end of this range. 6 user roles, exercise library with video, integrated nutrition system, inactivity notifications, and an internal marketplace. This is the kind of project a software house would quote at R$100k+.
SaaS MVP
R$25,000 to R$60,000. Minimum viable product to validate in the market. Includes authentication, billing (payment gateway integration for recurring charges), onboarding, client dashboard, and admin panel.
Timeline: 8 to 16 weeks.
The difference between a full webapp and a SaaS MVP: the SaaS needs automated billing (plans, trials, upgrades, downgrades, cancellations), guided onboarding for new users, and infrastructure designed for multiple clients (multi-tenancy). These are layers of complexity that don’t exist in a webapp built for a single company.
These prices are for a senior dev working solo or in a pair. A large agency charges 2x to 3x for the same scope because of operational overhead (office, management team, account management, PM). More on the differences between hiring a freelancer, agency, and software house.
What drives the price up
Number of user roles
Each role is a different panel with its own access rules. AutoPars has 3 (buyer, seller, admin). FitPlan has 6 (owner, board, admin, trainer, nutritionist, member). Each role adds screens, permission logic, and tests.
A system with 1 role: the dev builds one interface. With 3 roles: it’s basically 3 applications sharing the same database. The work doesn’t triple (a lot of code is shared), but it easily doubles.
External integrations
Each third-party API adds real complexity. AutoPars had 5 integrations. Each one required: studying the documentation, setting up authentication (tokens, OAuth, API keys), implementing the necessary endpoints, handling webhooks, handling errors and edge cases, and testing in a sandbox environment.
A well-documented integration (Resend, Stripe) takes 3-5 days. One with bad docs or an outdated SDK can take 2 weeks. And all of them need maintenance: APIs change versions, deprecate endpoints, change response formats.
Business logic complexity
“Scheduling system” can be simple (date + time + client) or complex (multiple professionals with different schedules, automatic blocking, weekly recurrence, waitlists, automatic WhatsApp confirmation, rescheduling with advance notice rules).
The frontend screen looks similar in both cases. The backend changes completely. And it’s the backend that determines the price.
Another example: “shopping cart”. Sounds simple until you add discount coupons, shipping calculated by region, payment split between seller and platform, and variable taxes by product type. Each additional rule is logic that needs to be implemented, tested, and maintained.
Custom design vs template
Design built from scratch with the company’s visual identity costs more than adapting an existing design system. Both work. The difference is whether the system needs to look on-brand or whether functionality matters more than aesthetics.
For an MVP, I usually recommend UI with Tailwind and standard components. Clean, functional, and fast to implement. Custom design goes to v2, once the product has validated and it makes sense to invest in an elaborate visual identity. In practice, the difference between custom design and well-applied standard UI can be 2 to 3 weeks of work — time that in an MVP is better spent on features that generate value for the user.
What brings the price down
Lean scope
The biggest impact on price. An MVP that does 3 things well costs less than a system that tries to do 15 in the first version. And the MVP that validates quickly lets you invest with more confidence in the full version.
I’ve had a client who came in with an R$80k scope. After talking about what was essential to validate the business, the MVP came to R$25k. Launched in 6 weeks, validated the hypothesis, and now they’re evolving with real data.
A stack the dev knows well
When the developer already has deep experience with the stack, development is faster. React + Supabase + TypeScript is what I use every day. Every component I build and every integration I implement is faster because I’ve done it dozens of times before.
A dev learning a new technology on your project = delays, bugs, and hidden costs.
Fast client decisions
A project that stalls for 2 weeks waiting for layout approval costs more than a project with a client who responds the same day. This isn’t an exaggeration: I’ve had a 6-week project take 10 because the client took 4-5 days to approve each weekly delivery.
The dev’s time is reserved for your project. If they’re stuck waiting for approval, the cost of that idle time is real — whether they charge for idle hours or the entire project timeline stretches.
Pricing model
I work with fixed-price projects. The client knows how much they’ll pay before we start. The scope is defined in a document, deliverables are clear, and the price doesn’t change unless the scope changes.
If scope changes during development (and sometimes it does — businesses evolve, priorities shift), I handle it transparently. “This change adds X to the project. Do you want to include it now or save it for post-launch?” No surprises on the invoice.
Hourly billing works well for maintenance and post-launch evolution. Bug fixes, new features, UI adjustments. But for a project from scratch, fixed price protects both sides: the client knows how much they’ll invest, and I know the scope I need to deliver.
How to evaluate whether the investment is worth it
More useful than asking “how much does the system cost” is asking “how much does it cost to not have the system”.
If the team spends 10 hours per week on a manual process that a system automates, that’s 40 hours per month. At R$30/hour in employee cost, that’s R$1,200/month. In 3-4 months, the simplest system (R$8k) has paid for itself.
If data errors cause rework, lost sales, or client issues, the cost is even harder to calculate — but it’s real. Webapp vs spreadsheet covers the signs that it’s time to migrate.
If your competitor has an online system and you’re still sending PDFs over WhatsApp, the cost is missed opportunity. Every client who searched on Google and didn’t find you went to your competitor.
The investment is in operational efficiency, in how clients perceive your business, and in being able to grow without hiring at the same rate. A web system scales without proportional cost: serving 50 or 500 clients on the same system has practically identical infrastructure cost. Hiring 10 people to manually do what the system does, doesn’t.