What a custom web app costs in Europe depends on who builds it: the UK agencies that publish their prices put it at £10,000 to more than £150,000. Where yours lands comes down to six scope decisions, not the hourly rate. My own web and mobile app projects are from €6,000, depending on scope, with a fixed price in the proposal after discovery.
Neither number is your price. Your scope is.
What counts as a custom web app
A custom web app is browser software built around one organisation's process, with its own data model, user accounts, permissions and rules. It is not a subscription product configured to fit. A website publishes information. A web app takes input, stores it and changes what each person sees based on who they are.
In practice that means internal systems that replace a spreadsheet and three people retyping data, portals where clients or members log in, SaaS products and MVPs for a new idea. HoReCa PRO, a B2B portal for the hospitality industry, is one I built in 2023. It runs on a membership model, so login, registration and the newsletter are built into the reading flow, not a page off to the side.
How much does a custom web app cost? The price tiers
A custom web or mobile app is from €6,000, depending on scope. In my practice a web shop is from €3,000 and a website from €1,500. The figures are entry points, not a price list. The final price is fixed in the proposal after discovery, when we both know exactly what is being built.
| What you are building | Investment | Approx. in GBP | Timeline | What moves it most |
|---|---|---|---|---|
| Website | from €1,500 | ≈ £1,284 | 3–6 weeks | Number of pages, amount of content, photography |
| Web shop | from €3,000 | ≈ £2,568 | 6–12 weeks | Number of products, payment and ERP integrations, product migration |
| Web or mobile app | from €6,000 | ≈ £5,136 | from 3 months | Integrations, roles, user flows, data migration |
| Support after launch | from €50 / month | ≈ £43 / month | ongoing | Fixed hours a month, priority response |
The pound figures use the rate fixed in my proposals, 0.856 to the euro, set on 20 September 2026. The contract amount is always in euros, and a UK client sees the approximate pound figure next to it.
The web and mobile app service lists what an app project delivers. That includes a written specification, a clickable prototype before the first line of code, user accounts and roles, and integrations. At handover you get the repository, the documentation and every access credential.
Calculator
Scope and investment estimate
Seven questions, no contact details required. The result is not a proposal but a frame: which tier your project falls into, where it starts and what pushes it up within that tier.
Result
No result yet. Choose what you are building.
Choose what you are building
The result builds up here as you answer. Nothing is sent and nothing is stored.
Costs after launch
Independent of the tier. These keep running once the project is done.
- Hosting
- €90 a year≈ £77The first year is in the project price. After that it is paid yearly.
- Domain
- As it is nowIt stays with its registrar and you keep paying it. I do not take it over.
- Platform licence
- €0There is none. The build is custom, so there is no monthly subscription to the tool it was made in.
- Support
- from €50 / month≈ £43Price on 10 Sep 2026: from €100 / monthThe anchor price is what the same service cost on 10 Sep 2026, and from 1 Oct 2026 it has to be shown next to the current price (NN 101/2026). That way you can see at a glance whether the price has changed.Updates, security, backups and small changes, with priority response.
- Third-party licences
- Only if the scope needs themThe cost is quoted before anything is taken on.
Pound amounts are approximate, at the 20 September 2026 rate. The contract amount is always in euros.
This is a frame, not a proposal. You get the exact scope, timeline and price in writing, after a 30-minute call.
How the estimate works
The type of project sets the tier, and the tier carries the starting price and the indicative timeline. Scope drivers do not change the tier but your place within it: the more of them, the further the project sits from the starting price.
The exception: a website where users sign in and have several roles with different permissions is estimated as an app.
| Tier | Investment | Timeline |
|---|---|---|
| WebsiteStructure, design, build, legal pages and launch. | from €1,500≈ £1,284Price on 10 Sep 2026: from €1,500The anchor price is what the same service cost on 10 Sep 2026, and from 1 Oct 2026 it has to be shown next to the current price (NN 101/2026). That way you can see at a glance whether the price has changed. | 3–6 weeks |
| Web shopPayments, ERP, invoicing and product migration. | from €3,000≈ £2,568Price on 10 Sep 2026: from €3,000The anchor price is what the same service cost on 10 Sep 2026, and from 1 Oct 2026 it has to be shown next to the current price (NN 101/2026). That way you can see at a glance whether the price has changed. | 6–12 weeks |
| Web or mobile appUser accounts, roles and access permissions. | from €6,000≈ £5,136Price on 10 Sep 2026: from €6,000The anchor price is what the same service cost on 10 Sep 2026, and from 1 Oct 2026 it has to be shown next to the current price (NN 101/2026). That way you can see at a glance whether the price has changed. | From 3 months |
Scope drivers and why they move the price
- User sign-in: needs user accounts and sign-in
- Sign-in brings registration, account recovery, sessions and protection of each user's data, all of which is built and tested separately from the design.
- Roles and permissions: needs roles and permissions
- Each role sees and may do something different, so every screen and every action is checked for every role, and the number of cases grows with each role.
- Card payments: adds a payment integration
- A payment has more outcomes than success and decline (an abandoned transaction, a refund, a double click), and each one has to be handled and run through the provider's test environment.
- ERP or accounting: adds an integration phase
- ERP and accounting systems have their own rules and data that do not match yours, so the mapping and syncing are agreed and tested on both sides.
- E-invoicing or fiscalisation: adds e-invoicing and tax reporting
- Tax rules are rigid: an invoice must follow the prescribed format, reach the tax authority or an access point, and stay valid when the external service does not respond.
- Another integration: adds an integration phase
- Every external system means documentation, access keys, limits and a plan for the moment it stops working.
- Data migration: needs data migration
- Old data is rarely clean: before the move it is mapped, cleaned and checked, and the transfer is repeated until the numbers match.
- Several languages: needs a multilingual structure
- Each language gets its own URLs, SEO tags and content that is maintained separately, and the interface is tested in every language because a longer translation breaks the layout. Translating the copy is a separate job.
- Accessibility and data protection: needs a stricter review and documentation
- Sensitive data calls for minimal collection, access logs and a written record of decisions, and accessibility means checking every screen by keyboard and with a screen reader.
Pound amounts are approximate, at the 20 September 2026 rate. The contract amount is always in euros.
What moves the price of a custom web app
Six things move the price of a web app more than anything else. They are integrations, roles and permissions, user flows, data migration, compliance and design depth. The number of screens is not on the list, and that is deliberate.
Integrations
Every system your app talks to is a small project of its own, with authentication, error handling, retries and the parts the documentation leaves out. Rate limits are the usual surprise. Stripe allows 100 operations per second in live mode but only 25 in its sandbox (Stripe, rate limits, accessed 24 September 2026). A bulk import tested against the sandbox has to be paced from the first day, not after it fails.
Smaller vendors are the harder case. Some publish no limits at all and return an error once you cross one. Some APIs have no endpoint for the records you need, so the only way in is an export from their interface. Before a proposal I ask which systems the app has to talk to. For each one I check for a documented API, readable limits and an endpoint for every record the app needs.
Signing in with Google, or reading a user's calendar, adds the OAuth consent screen. While it is in Testing mode, Google issues refresh tokens that expire after seven days (Google Identity, OAuth 2.0, accessed 24 September 2026). An integration that passes every test can stop syncing a week after launch, so I treat the consent screen's publishing status as scope, not as a detail for launch day.
Roles and permissions
Each role is a second app hiding inside the first. It has its own screens, its own rules about what it may see and its own tests proving it can't see the rest. Broken access control is number one in the OWASP Top 10:2025, the most widely used list of web application security risks (OWASP Top 10:2025, accessed 24 September 2026).
The cheap design puts customers and staff in one user table and tells them apart with a role flag. Keeping customer access fully separate from staff access costs more up front, and one wrong value can then never open your internal data to a customer. On a portal where your clients log in next to your team, I scope that separation as its own line.
Permission checks also run on the server for every action, not only in the interface, because a hidden button is not a permission.
The number of user flows
A user flow is one complete path through the app with a goal: sign up, place an order, approve a phase, invite a colleague. Pricing by screens misleads, because one screen can hold three flows and one flow can span six screens.
On HoReCa PRO, login, registration and the newsletter are built into the reading rather than sitting off to the side. That is three flows woven into the content, not three extra screens. Flows that have to hold up later cost more again. A client approving a deliverable, stored as a record that can't be edited, costs more than a button that flips a status.
Data migration
Migration is priced by how messy the data is, not by how much of it there is. Exports from invoicing, CRM and shop systems can carry fields that don't mean what their name says. Think of a customer type column that marks every company as a consumer, dates in two formats, or one client entered twice under slightly different names.
Each of those needs a rule: ignore the column, decide from another field, merge on the tax number. None of that is visible in a feature list, and all of it is work. I scope migration from a real sample export, not from a record count.
Compliance: the European Accessibility Act and GDPR
The European Accessibility Act applies from 28 June 2025 to services such as e-commerce and consumer banking. Microenterprises that provide services are exempt (Directive (EU) 2019/882, EUR-Lex, accessed 24 September 2026). A microenterprise has fewer than 10 staff and a turnover or balance sheet of no more than €2 million. If your app sells to consumers in the EU and you are above that line, accessibility is a requirement.
Built in from the prototype, accessibility is part of every screen. Added after launch, it means reopening every screen.
GDPR moves the price through features, not paperwork. A processor that handles personal data on your behalf needs a contract under Article 28. Your users have rights of access and erasure under Articles 15 and 17 (Regulation (EU) 2016/679, EUR-Lex, accessed 24 September 2026). Access and deletion on request, retention periods and a record of who changed what are real scope.
Security sits on the same line. Article 32 of the same Regulation names encryption of personal data among the appropriate security measures. An app where clients upload documents or send access credentials needs encryption and a permission check on every download, and that belongs in the proposal as scope.
Design depth
Design depth is the gap between a styled admin template and an interface designed around your users' flows. Design depth is the easiest driver to scale up or down.
Every project gets a clickable prototype before the build. Changing your mind at the prototype costs an hour. Three weeks later it costs three weeks.
My contracts include two rounds of design revisions, and further rounds are billed hourly. An internal tool used by six people in your office doesn't need the design depth of a product you sell to thousands. Saying that in the brief takes a real piece out of the scope.
How to brief a web app so the quote holds
A web app quote holds when the brief names the price drivers, not the screens. Five steps get you there before the first call:
- List every system the app must talk to, and note whether each has an API or only a CSV export. That answer moves the price more than any screen.
- Name the roles and write one line on what each may see and do.
- List the flows, not the pages. Five flows and fifteen flows are different projects, even when both are called a portal.
- Get a sample export of the data you will migrate. A real file in the first week of discovery tells you more than any estimate.
- Say which rules apply to you: whether you are above the EAA microenterprise line, what personal data the app will hold, and how much design depth your users need.
The common mistakes mirror those steps: pricing by screens, comparing headline figures instead of included lists, and budgeting for the build but not for the years of running it.
What UK agencies publish for a custom web app
UK agencies that publish prices put a standard custom web app at £30,000–£80,000 in their 2026 guides. Tulip Tech lists £10,000–£30,000 for a simple web app, £30,000–£80,000 for a standard one and £80,000–£150,000+ for a complex one (Tulip Tech, web app development cost UK, accessed 24 September 2026). Wakeflow gives £20,000–£80,000 and more for web apps such as dashboards, portals and SaaS (Wakeflow, custom web development cost UK, accessed 24 September 2026).
Published agency ranges are a map of the market, not a benchmark for a quote. Each one bundles different things. Tulip Tech, for example, budgets £5,000–£12,000 for discovery on a standard project and £1,000–£5,000 per integration, and adds 10 to 15 percent for project management.
Two quotes for the same app can differ by a factor of two on what's in them alone. Compare the included list, not the headline figure.
Mine is on the pricing page. It covers discovery and a written specification, a clickable prototype, build, testing and launch. It also covers the repository and documentation at handover, the first version of the copy, and 30 days of support after launch. Photography, a visual identity from scratch and third-party licences are billed separately.
My entry points sit below the agency ranges because one person does the design, the build and the communication, with no layer between you and the person writing the code. That is a description of the structure, not the reason to hire me. If the scope, the contract and the person don't fit, the price is irrelevant.
How payment and contracts work for a UK or EU client
Projects are a fixed price agreed before work starts. Payment runs in phases tied to deliverables: the first part on signing, the rest by the phases set out in the contract. Each phase closes with a written sign-off, so on a larger project you never pay much more than has been delivered. Only consulting and support are billed hourly, and consulting hours are deducted from the proposal if we carry on into a project.
Discovery ends with a written specification, a fixed proposal, the Project plan, the Contract and an NDA, all in English. What to check before you hire someone in another country is in my guide to hiring a developer in Croatia. My English service agreement gives jurisdiction to the courts where the defendant is registered, not to a court in Zagreb. Defendant-based jurisdiction is fair to both sides of a cross-border contract.
I work from Zagreb on Central European Time, Monday to Friday, 09:00–20:00 CET. Central European Time is the same clock as Brussels, Amsterdam and Copenhagen, and one hour ahead of London. The response time written into the project plan is 24–48 hours, excluding weekends.
In 2025 I delivered the website for FESCA, the Federation of European Scleroderma Associations. It is a Belgium-based federation of 23 associations from 18 countries. FESCA's project was a website, not an app, but the proposal, the NDA and the contract ran in English across borders, the same way they run for an app.
What a custom web app costs to run after launch
After launch you keep paying for hosting, the domain and any third-party services. Each is listed in the proposal with an approximate yearly figure before you sign. Nothing is charged without your prior agreement.
There is no platform licence, because a custom app isn't built inside a subscription tool. For a website, hosting in my proposals is €90 a year, and the first year is inside the project price. The domain stays with your registrar and you keep paying it as you do now.
An app usually adds lines a website doesn't have, such as a service for sending email or storage for uploaded files. Each one is quoted before it is taken, with its yearly figure, so the running cost is known on the day you sign rather than discovered on the first invoice.
The first 30 days of support after launch are included. After that there is a support package from €50 a month, with a fixed number of hours and priority response, and no year-long contract. An app nobody maintains isn't finished. It is waiting to break.
When you don't need a custom web app
You don't need a custom web app if an off-the-shelf tool covers your process. The test is whether its compromises cost you money or customers. If an off-the-shelf platform solves your problem more cheaply than custom, I'll tell you, even though a custom project suits me better.
Concretely, a custom build is the wrong answer in these cases:
- You sell a standard catalogue with a standard checkout. Shopify or a similar platform does that well, and a custom shop pays off only when integrations, B2B pricing or logistics outgrow it.
- Your process is the same as everyone else's. A standard CRM, ticketing or booking tool covers it. Custom pays off when the process is your advantage, not when it is a formality.
- You are still testing whether anyone wants the product. Talk to customers and test with a landing page or a no-code prototype first. Build the MVP when you know what it has to prove.
- Your budget is below the entry point. Under €6,000 for an app, a custom build doesn't make sense, and I'd rather say so than cut the scope until the work stops working.
- You want a mobile app you don't need. Without push notifications, offline use, the camera or location, a good web app is enough.
I build only custom projects. If a template or a SaaS subscription does the job, that is the better choice.
FAQ
Why can't I get a fixed price for a web app by email?
Because it would be invented. A price without understanding the scope is either too high, so the developer can protect themselves, or too low and later corrected through extra invoices. An hour of conversation and a discovery phase give you a number that holds to the end of the project, written into the proposal and the contract.
How long does a custom web app take to build?
A custom web app takes from three months, depending on scope. An MVP with a clearly bounded scope usually fits into three or four months. The deadline is agreed before the build starts and is part of the proposal. If the scope changes, you see what that does to the deadline at the moment it changes, not at the end.
Do you charge VAT to a business in the UK or another EU country?
For a business in another EU country, the place of supply is generally where the customer is established. The customer accounts for VAT under the reverse charge (Articles 44 and 196 of Directive 2006/112/EC, EUR-Lex, accessed 24 September 2026). For a UK business the general rule is the same, with the reverse charge applied in the UK (HMRC, VAT Notice 741A, accessed 24 September 2026). Your proposal states the VAT treatment, and this is not tax advice.
Who owns the code after launch?
You do. At handover you get the repository, the database, every access credential, the hosting and the documentation. Nothing stays with me to hold over future negotiations. If you carry on with another developer or your own team tomorrow, they pick up where the work stopped, with documentation written for exactly that situation.
Do I need a mobile app, or is a web app enough?
For most businesses a good web app that also works on a phone is enough. A mobile app pays off when you need push notifications, offline use, the camera or location. When I build one, it is React Native, so one codebase serves iOS and Android. If your case genuinely needs fully native apps, I'll tell you and point you on rather than charge for it.
If you have a process that has outgrown spreadsheets, or a product you want in front of real users, book a 30 minute intro call. You'll leave knowing whether it is work for me, roughly what it costs and what the next step is, whether or not we work together.
About the author
Jurica Stublić
Web designer and developer · 4 years · 50+ projects
I build custom web and mobile apps, web shops and websites, and I stay on the project after launch. Alongside client work I develop my own product: ai.webica.hr.
