How much does a custom web app cost in the UK?
Use the planner to estimate a range for a custom web app, mobile app or MVP, then see which parts of the scope change the cost.
A custom app can cost a few thousand pounds or many times that. Before putting a range against it, I'd want to know who's using it and what they need to do. Then we can work out the smallest version they can rely on, and what has to work before release.
The planner below follows how I scope independent website and product work. It'll give you an early planning range, rather than a fixed quote or a survey of agency rates.
Shape a realistic first release
Choose the closest version of the product. The range is for planning an independent build, not a fixed quote or a claim about the whole market.
Indicative build range
£6,500 to £15,500
Roughly 4 to 7 delivery weeks
Assumes
- • A clear primary user and problem
- • One decision-maker and timely feedback
- • Existing brand foundations or a simple visual system
- • Hosting and third-party fees priced separately
Discovery can move the range in either direction. Regulated data, offline behaviour, complex migrations and native device features need a closer review.
The planning ranges I use
For a focused internal tool, I use a planning range of about £3,500 to £8,500. A customer-facing MVP with authentication, real data and one clear core workflow commonly starts around £8,000 to £22,000. A shared web and mobile product, or complex roles and integrations, can reach £18,000 to £45,000 or more.
These are the bands I use for planning, assuming a small group making decisions and clear feature boundaries. The release still needs to prove something useful. They won't cover every agency model, third-party subscription or long-term support arrangement.
Prototype, MVP or production release?
| Release | What it should prove | What it usually includes |
|---|---|---|
| Working prototype | People understand and can complete the core flow. | A believable interface, representative data and enough logic to test the idea. |
| Focused MVP | Real users will adopt the product for one meaningful job. | Accounts, real data, a primary workflow, analytics and a practical release process. |
| Production first release | The product can become part of a real operation. | Reliable permissions, edge cases, monitoring, support paths and stronger quality controls. |
Calling it an MVP won't remove the work. If it needs five user roles, offline behaviour, payments, a complex migration and native device access, those still need building.
What changes the cost most?
The number of real workflows
I'd rather count the workflows than the screens. Twelve screens with one clear journey may be simpler than four screens with approval states, exceptions and three kinds of user.
Users, roles and permissions
One account type is fairly straightforward. Once you add administrators, clients, staff and external partners, you need to work through what each can view, change, approve and export.
Data and integrations
With HubSpot, Stripe or another API, the connector is only part of the job. What happens with missing data, duplicates, failed requests or permissions? You also need to think about what happens when the other system changes.
Native and offline behaviour
There may be a good reason to build mobile: camera access, notifications, background tasks, location, app-store distribution or offline sync. Each brings more states to design and test. If the main job works in a browser, a responsive web app may cost less for the first release.
Design and content readiness
It's easier to estimate when the visual system and product language are agreed. If the proposition, onboarding and interface still need working out during development, I'd allow time for that too.
Three worked planning examples
Focused internal tool: £3,500 to £8,500
- One main workflow for a small known team.
- Simple sign-in or controlled access.
- One data source or lightweight integration.
- Responsive web interface with practical reporting.
Customer-facing MVP: £8,000 to £22,000
- Account creation and a useful onboarding path.
- A complete core workflow using real data.
- Transactional email, analytics and basic administration.
- A tested release with monitoring and a clear support route.
Web and mobile first release: £18,000 to £45,000+
- A shared product model across browser, iOS and Android.
- Native capabilities such as notifications or camera access.
- Several account types, integrations or payment flows.
- Store preparation, device testing and release operations.
What belongs in the first release?
The first release needs one outcome that helps you decide whether it's worth building the next one. Follow it all the way through: who arrives, what they do, what changes in the system and how you'll know it worked.
A useful first-release test is:
- Would the product still solve the main problem without this feature?
- Does it reduce a known risk, or only make the idea feel more complete?
- Can the team support it properly after launch?
- Can a manual step cover the need while demand is still uncertain?
- What event or customer behaviour will tell us whether to keep improving it?
What the cheapest quote might leave out
Two quotes can use the same feature list and still cover different work. One might include discovery, responsive states, data rules, analytics, testing and release support. Another might price the visible screens and leave everything else for later.
That's where I'd ask about the less visible parts. What happens if a request fails, someone lacks permission or the data already exists? What if the product needs changing after launch? The answers tell you whether you're buying a working product or a convincing demonstration.
Related product work
- BodyByCiara fitness subscription platform
- Listing Monster AI ecommerce platform
- Mirador Local business profile management product
Frequently asked questions
What determines the cost of a custom web app?
The main costs come from user journeys, permissions, the data model, integrations, payments, reporting and looking after the product. A short feature list can hide work in error handling, administration and release tests.
Is the MVP cost planner a fixed quote?
No. It's an estimate based on the scope assumptions you choose. You still need to check integration limits, migration, testing and the first release before agreeing a price.
How can I reduce the scope of an MVP?
Choose one audience and one complete useful workflow. Keep essential access, recovery and measurement. Leave secondary dashboards, extra platforms and some automation for later, where a named person can handle the work manually at first.
.jpg)
