The real web app cost for a small business in 2026 runs from about $1,500 for one automated workflow to $10,000 and up for a custom application with logins and a database. Running it usually costs $50 to $150 a month. Four things set the price: users, data, connections and how much has to be built from scratch.
The cheapest web app is the one that solves one expensive problem well. This post explains what you actually get at each price, what drives the cost up, what it costs to keep running, and when AI tools are enough to do it yourself.
Is it a web app, or a website?
A website shows information to visitors. A web app does work for people who use it. If any of these are true, you need a web app, not a better website:
- People log in. Clients check their project, dealers place bids, staff see their own dashboard.
- It stores and changes data. Bookings, quotes, orders, inventory, leads.
- It follows rules. A quote is calculated, a bidding round closes at a set time, a lead is assigned to the right person.
- It connects tools. A form result lands in your CRM, triggers an email and updates a spreadsheet without anyone copying anything.
What does a web app cost?
There are three common shapes, and the price follows the shape. Our own web apps and automation menu shows typical prices:
- One automated workflow, from $1,500, 1 to 2 weeks. One manual process done properly by software: quote generation and follow-up, client onboarding, invoice chasing, weekly reporting. Usually built on tools you already pay for.
- A lead response system, $2,500, 1 to 2 weeks. Every inquiry from your form, phone and WhatsApp captured in one place, answered instantly, routed to the right person and followed up automatically.
- A custom application, from $10,000, 4 to 8 weeks. Client portals, dealer platforms, booking and quoting tools, dashboards. Anything with its own logins, database and business rules.
Every price is fixed before work starts. The full menu is on our pricing page.
The idea alone isn't enough to price an app. Before we quote, we often sit with the client to take it from idea to concept to a full blueprint: what the app does, who uses it, and what it really needs. That's why the range is so wide. One client needs a simple CRM. Another, like the Auvion dealer platform, needs a full portal with live bidding, a database behind it, and everything in between.
What makes a web app cost more?
Four things move the price more than anything else:
- Users and roles. One type of user is simple. Clients, staff and admins who each see different things take more work, because every screen and every rule has to respect who is looking.
- Data and rules. Storing a list is cheap. Calculating prices, closing timed rounds, keeping an audit trail of who did what, and handling the cases where things go wrong is where most of the real work lives.
- Connections. Every outside system, a CRM, a payment provider, WhatsApp, an accounting tool, adds setup, testing and things that can break.
- Custom versus ready-made. Logins, email sending and file storage are solved problems with good existing services. Paying a developer to rebuild them from scratch is the most common way a budget doubles.
What does it cost to run?
Running costs are lower than most people expect. A typical small business web app built on a modern stack runs on:
- Hosting, for example Vercel's Pro plan at $20 a month per team member.
- A database with logins and file storage, for example Supabase's Pro plan at $25 a month.
- Email sending and a domain, a few dollars to a few tens of dollars a month.
Together that is usually $50 to $150 a month. On top of that comes maintenance: fixing what breaks, updating what ages, adding what the business needs next. Budget for it from the start, either as a monthly retainer or as work quoted when it comes up.
- Hosting Vercel Pro
- $20 a month per team member
- Database, logins and file storage Supabase Pro
- $25 a month
- Email sending and a domain
- a few dollars to a few tens of dollars
- Together, usually
- $50 to $150 a month
Can you just build it with AI?
AI app builders such as Lovable, Bolt and Replit can now produce a working app from a description in an afternoon. For a prototype, or an internal tool used by 3 people with nothing sensitive in it, that is often enough, and it is worth trying before paying anyone.
The risk shows up when real customers and real data arrive. AI builders are good at the screens you can see. The expensive mistakes are in what you cannot see: who is allowed to read which data, what happens when 2 people book the same slot, how a password reset works, whether yesterday's records survive a bad update. Those are exactly the parts a customer never notices until they fail.
We use AI coding tools on every build. They are a big part of why our custom applications take weeks rather than months, and why the prices above are lower than they would have been a few years ago. The judgment about what to build, and the checking of what the AI produced, is still the job.
A good rule: if it has customer logins, payments or personal data, have a developer build it or at least review it before launch.
What does a real one look like?
A Dutch car brokerage used to run every round of dealer bidding by phone and WhatsApp. We built them a platform dealers log into, with timed bidding rounds that close themselves and a full record of every bid. It now runs the commercial core of the business. The Auvion dealer platform case study walks through how it was built.
How to keep the cost down.
- Start with the one process that costs you the most in hours or lost jobs, not the whole system at once.
- Use the tools you already pay for. Your CRM, email and calendar can usually do more than they are doing.
- Launch a small first version in weeks, then add to it based on how people actually use it.
- Own everything. Make sure the code, accounts and data are in your name, so you are never locked to one developer.
If you are not sure whether your problem needs a web app at all, send us your website for a free 5-minute audit. We will tell you whether it is software, a better page, or neither.
