A custom web application in the UK typically costs between £15,000 for a simple internal tool and £200,000-plus for a full SaaS product, with most small and medium business projects landing in the £30,000 to £80,000 range. A customer portal or workflow app usually costs £30,000 to £80,000, a bespoke CRM £35,000 to £100,000, and a minimum viable product (MVP) £15,000 to £35,000. The headline build price is only part of the story. Across a three to five year horizon, ongoing costs (hosting, maintenance, support and iteration) account for 60 to 70 per cent of total spend, so the build is often less than half of what you will actually pay. UK developer day rates run from £250 for a junior to £900 for a lead architect, and £800 to £1,200 for specialist AI work. This guide breaks down every tier and every hidden cost so you can budget honestly.
Last updated: June 2026
A custom web application in the UK costs between £10,000 and £500,000, but that range is so wide it is almost useless on its own. The honest answer is that price tracks complexity, and complexity falls into recognisable bands. A simple internal tool that replaces a spreadsheet sits at the bottom. A multi-tenant SaaS product with billing, role management and a public marketing surface sits at the top. Most UK SMEs commissioning their first serious application spend between £30,000 and £80,000, and London-based businesses pay a 10 to 20 per cent premium over the rest of the country for the same scope.
The single biggest mistake we see business owners make is comparing two quotes on price alone when the two proposals describe completely different products. A £20,000 quote and a £55,000 quote for "a customer portal" usually means one team scoped a login page with a list view, and the other scoped role-based access, document handling, audit trails and a payment integration. The number is not the product. The scope is the product.
Here is the realistic band structure we use when scoping projects for UK clients in 2026:
| Application type | Typical UK cost | Build time | Who it suits |
|---|---|---|---|
| MVP / proof of concept | £15,000 - £35,000 | 6 - 12 weeks | Startups validating an idea |
| Internal tool / workflow app | £15,000 - £30,000 | 6 - 10 weeks | Replacing spreadsheets or manual processes |
| Customer portal / web app | £30,000 - £80,000 | 3 - 6 months | Client-facing self-service |
| Bespoke CRM | £35,000 - £100,000 | 4 - 7 months | Sales and operations management |
| Custom e-commerce platform | £50,000 - £150,000 | 4 - 8 months | Non-standard retail logic |
| SaaS product / enterprise platform | £100,000 - £500,000+ | 9 - 18 months | Software businesses, large operations |
These figures assume a UK agency or a senior freelance team, not an offshore body shop selling hours at a quarter of the price. They also assume a clean build with no inherited legacy code, no broken data to migrate, and a client who can make decisions promptly. Each of those assumptions, when it fails, adds cost. If you take one thing from this guide, take this: the headline build figure is the deposit, not the bill. The real number is what you spend keeping the thing alive and useful over the years that follow, which we cover in detail below. For a deeper look at how scope translates into architecture, our web application development services in London page sets out how we structure each tier.
Five factors drive the majority of the cost variation between web application projects: business-logic complexity, the number of integrations, the number of distinct user roles, the design ambition, and any data migration. Get clear on these five and you can predict, within reason, which band your project will fall into before you speak to a single agency.
Business-logic complexity is the largest single driver and the hardest to see from outside. A booking app that simply records appointments is cheap. A booking app that handles overlapping resources, dynamic pricing, cancellation rules, deposits and automated reminders is several times more expensive, even though both are "booking apps" to a buyer. The visible interface can look identical while the engine underneath differs by an order of magnitude. This is precisely why screenshots are a terrible basis for estimating cost.
Integrations are the second driver. Every external system your application must talk to (a payment provider, an accounting package, an email platform, a courier API, an existing database) adds development and testing time. Each integration also adds a long-term liability, because third-party APIs change and your application has to keep up. A useful rule: budget for each non-trivial integration as a mini-project in its own right, not a free afterthought.
The remaining drivers compound. More user roles mean more permission logic and more testing. Higher design ambition (custom interactions, animation, a distinctive brand expression rather than a stock template) adds design and front-end engineering time. And data migration from a legacy system is frequently the most underestimated line in any quote, because the old data is almost always messier than anyone admits.
| Cost driver | Low end | High end | Why it matters |
|---|---|---|---|
| Business logic | Simple CRUD records | Multi-step rules, calculations, automation | Largest hidden multiplier |
| Integrations | None or one | Five-plus external systems | Each adds build and lifetime risk |
| User roles | Single user type | Admin, staff, client, partner tiers | Permission logic multiplies testing |
| Design ambition | Template-based | Bespoke, branded, animated | Front-end engineering hours |
| Data migration | Fresh start | Years of messy legacy data | Most underestimated line item |
Our honest stance: be sceptical of any agency that quotes a firm number before it understands these five factors. A responsible quote follows a discovery phase, not a sales call. If a developer gives you a fixed price within an hour of meeting you, they are either padding the number heavily to protect themselves, or they are about to discover problems mid-build and come back for more money. Neither is the position you want to be in. Much of the logic that drives this complexity can be reduced through smart business process automation, which is often cheaper than rebuilding the same manual workflow inside a custom app.
UK web developer day rates in 2026 range from £250 for a junior developer to £900 for a lead architect, with specialist AI and machine-learning engineers commanding £800 to £1,200 per day. Your final invoice is essentially a blend of these rates multiplied by the days each role works on your project, plus the agency's margin. Understanding the rate card helps you sense-check whether a quote is realistic or fantasy.
There is a persistent myth that you should always hire the cheapest available rate. In practice, a senior developer at £650 per day who solves a problem in two days is cheaper than a junior at £300 per day who takes a week and produces code that needs reworking. Rate is not cost. Output per pound is cost. We have rebuilt enough cheap-first projects to state this plainly: the lowest day rate is frequently the most expensive route once rework is counted.
| Role / seniority | UK day rate (2026) | Typical hourly equivalent |
|---|---|---|
| Junior developer | £250 - £350 | £35 - £50 |
| Mid-level developer | £400 - £550 | £55 - £75 |
| Senior developer | £550 - £750 | £75 - £105 |
| Lead / solution architect | £600 - £900 | £85 - £125 |
| AI / ML specialist | £800 - £1,200 | £110 - £170 |
The supplier model you choose also shifts the maths. There are three broad routes, each with a different risk profile:
One more cost most buyers never consider: IR35. If you engage an offshore contractor or an individual contractor inside a personal service company, the off-payroll working rules can apply, and getting the determination wrong creates a tax liability. For most SMEs, a fixed-scope agency engagement sidesteps this entirely, which is part of why it is often the cleaner commercial choice despite the higher rate. Our software development service in London works on a fixed-quote basis precisely so you are never exposed to open-ended day-rate drift.
The hidden costs of a web application are hosting, maintenance, support retainers, emergency fixes, third-party licences and the cost of iteration after launch, and together they routinely exceed the original build price within two to three years. Most agency cost guides bury these or omit them entirely, because they make the headline number look smaller and easier to sell. We think that is dishonest, so here is the full list.
Maintenance is the big one. A healthy application needs ongoing work: security patches, dependency updates, bug fixes, and adapting to changes in the browsers, devices and third-party APIs it depends on. The industry rule of thumb is that annual maintenance runs at 15 to 25 per cent of the original build cost. For a £50,000 build, that is £7,500 to £12,500 every single year, before you add a single new feature. Skip it and your application quietly rots: dependencies fall out of support, security holes open, and eventually you face a forced rebuild that costs more than the maintenance would have.
| Cost category | Typical UK cost | Frequency |
|---|---|---|
| Hosting and infrastructure | £50 - £500 per month | Ongoing |
| Maintenance and updates | 15 - 25% of build per year | Annual |
| Support retainer | £500 - £3,000 per month | Ongoing |
| Emergency / out-of-hours fixes | £80 - £150 per hour | As needed |
| Third-party licences and APIs | £20 - £2,000 per month | Ongoing |
| Iteration and new features | £5,000 - £30,000 per year | Annual |
| SSL, domain, monitoring | £100 - £600 per year | Annual |
Hosting scales with usage. A low-traffic internal tool might cost £50 a month. A consumer-facing app with thousands of daily users, image storage and background processing can easily reach £500 a month or more. Cloud platforms bill by consumption, so a viral spike or an inefficient query can produce a nasty surprise on the invoice. Good architecture controls this; lazy architecture lets it run.
Then there are the costs that only appear when something goes wrong. Emergency out-of-hours fixes are charged at a premium, typically £80 to £150 per hour, and you only need them when the application is down and revenue is bleeding, which is exactly when you have least leverage to negotiate. This is why a sensible monthly support retainer, despite feeling like an expense you would rather avoid, is usually cheaper than paying emergency rates twice a year. Our blunt advice: if a proposal does not have a clear, written line for ongoing maintenance and support, that cost has not disappeared. It has simply been hidden from the quote and will land on you later.
The true total cost of ownership of a web application over five years is typically two to three times the original build price, because post-launch costs (hosting, maintenance, support and iteration) account for 60 to 70 per cent of lifetime spend. The build itself is only 40 to 60 per cent of what you will pay across the application's working life. This is the single most important number in the whole guide, and it is the one competitors consistently fail to put in front of you.
Let us work a real example. Imagine a UK SME commissions a £50,000 customer portal. The temptation is to budget £50,000 and move on. Here is what the five-year reality looks like:
| Year | Activity | Cost |
|---|---|---|
| Year 0 | Discovery, design and build | £50,000 |
| Year 1 | Hosting (£200/mo) + maintenance + minor features | £14,400 |
| Year 2 | Hosting + maintenance + new module | £18,600 |
| Year 3 | Hosting + maintenance + integration added | £16,200 |
| Year 4 | Hosting + maintenance + iteration | £15,400 |
| Year 5 | Hosting + maintenance + refresh | £17,000 |
| Total five-year cost of ownership | £131,600 |
The £50,000 build became a £131,600 commitment, and the post-launch portion (£81,600) is well over half. None of this is waste. Every pound after launch is what keeps the application secure, current and growing in value to your business. The mistake is not spending it; the mistake is failing to plan for it, then being ambushed by it.
This is also where the return on investment argument becomes serious. A well-built application that removes manual data entry and process friction can deliver around a 90 per cent reduction in data-entry errors and, on Forrester-cited figures, a 200 to 300 per cent return over a five-year period. The way to read the TCO table is not as a frightening bill, but as an investment schedule. If a £131,600 five-year spend saves two full-time-equivalent salaries of manual work and removes costly errors, the payback period can be under two years. The honest framing is this:
Where a chunk of that ongoing spend exists purely to handle repetitive admin, it is often worth routing those tasks through an AI automation agency approach rather than building and maintaining the logic by hand. Automation can compress the iteration line in the table above considerably.
Compliance can add anywhere from a few hundred pounds for basic UK GDPR alignment to tens of thousands for a regulated finance, healthcare or legal application, and almost no competing cost guide mentions it at all. If your application stores or processes personal data, you have legal obligations under UK GDPR, you may need to register with the Information Commissioner's Office, and you must build in appropriate technical and organisational measures. These are not optional extras; they are part of the build, and pretending otherwise is how projects fail their first security review.
For an ordinary SME application, the baseline compliance cost is modest: ICO registration is a small annual fee, and building privacy-by-design into the application (encryption, access controls, data retention, a clear way to handle subject access and deletion requests) adds engineering time rather than a separate licence. The cost rises sharply in regulated sectors, where the application itself becomes part of a regulated process and must satisfy a regulator, not just good practice.
| Sector | Key regime | Compliance cost impact |
|---|---|---|
| General SME | UK GDPR, ICO registration | Low: privacy-by-design plus small annual fee |
| Financial services | FCA rules, strong customer authentication | High: audit trails, security review, reporting |
| Healthcare | Data protection, clinical safety standards | High: governance, access control, assurance |
| Legal | SRA expectations, client confidentiality | Medium to high: data segregation, retention |
| Public sector | Accessibility regulations (WCAG) | Medium: accessibility build and audit |
Accessibility deserves a specific mention because it is widely ignored until it is too late. Public-sector bodies in the UK are legally required to meet accessibility standards, and even private businesses face growing pressure to make digital services usable by people with disabilities. Retrofitting accessibility into a finished application is expensive; building it in from the start costs little. Our stance is that accessibility should be a default, not a regulated afterthought, and any agency that treats it as an optional bolt-on is quietly storing up cost and risk for you.
The practical takeaway: tell your agency about your regulatory environment in the very first conversation. Compliance requirements shape the architecture, and architecture decided in week two cannot be unwound in month six without expensive rework. A custom build for a regulated sector should bake these requirements into the foundation, which is exactly how we approach custom CRM development for finance and professional-services clients who handle sensitive client data.
You evaluate a web application quote by comparing scope rather than headline price, demanding that ongoing costs are itemised, and watching for specific red flags that signal a number designed to win the deal rather than deliver the project. The lowest quote is the most dangerous one if it wins by leaving things out, because everything it omits comes back as a change request at a worse price later. A good quote is detailed, slightly uncomfortable in its honesty, and includes the costs that a sales-led quote conveniently forgets.
Here is the red-flag checklist we give to clients before they sign anything:
| Red flag in a quote | What it really means |
|---|---|
| Fixed price given before any discovery | Either padded heavily or about to change |
| No line for maintenance or support | Cost hidden, will land on you post-launch |
| Hosting not mentioned | Ongoing cost omitted to shrink the headline |
| Vague "design included" with no detail | Template work dressed up as bespoke |
| No mention of who owns the code | You may not own what you paid for |
| Unusually low day rate | Junior or offshore team, expect rework |
| No data migration line | The messiest, costliest task ignored |
Beyond the red flags, there are positive signals to look for. A trustworthy proposal includes a discovery or planning phase as a distinct, paid line item. This frequently surprises buyers who expect scoping to be free, but discovery is where the real money is saved, because a few thousand pounds spent understanding the problem prevents tens of thousands wasted building the wrong thing. An agency that wants to charge for discovery is an agency that takes the work seriously.
Use this build-versus-buy filter before you commission anything custom at all:
Our honest rule: do not commission custom software to feel modern. Commission it because a specific, measurable process is costing you real money or capping your growth, and no existing product solves it. If you cannot name the number you are trying to improve, you are not ready to spend on custom development yet.
Softomate Solutions builds custom web applications through a five-stage, fixed-quote process that takes most SME projects from discovery to launch in three to six months, with starting prices from £15,000 for an internal tool and from £35,000 for a customer-facing application. We work on a fixed quote agreed after discovery, not an open-ended day rate, so you know the number before we write a line of production code. That structure exists specifically to remove the budget-overrun risk that this entire guide warns against.
Our process is deliberately front-loaded with thinking, because every problem caught in discovery costs a fraction of the same problem caught in testing:
| Stage | Typical duration | What you receive |
|---|---|---|
| Discovery and scoping | 2 - 4 weeks | Specification and fixed quote |
| Design and architecture | 2 - 4 weeks | Designs and technical plan |
| Build | 6 - 16 weeks | Working software in cycles |
| Testing and compliance | 2 - 4 weeks | Tested, secure, compliant app |
| Launch and support | 1 - 2 weeks, then ongoing | Live application plus support plan |
Because we are a Stanmore-based team serving London and the wider UK, you get a senior engineer who answers the phone, works in your time zone, and signs a contract under UK law. That continuity is the difference between a support retainer that actually fixes things and a cheaper arrangement where nobody knows your codebase when it breaks. Where parts of your requirement are better served by automation than bespoke build, we will tell you, and route that work through our AI automation and GoHighLevel automation services rather than charging you to rebuild functionality you can configure. The goal is the lowest sensible total cost of ownership, not the biggest invoice.
A basic web application or internal tool in the UK typically costs £15,000 to £30,000 and takes six to ten weeks to build. This assumes a single user type, limited integrations and straightforward business logic. Add roles, integrations or complex rules and the figure rises quickly into the £30,000-plus customer-portal band.
Quotes differ because they often describe different products under the same name. One team may scope a simple login and list view, another a full role-based system with integrations and audit trails. Compare detailed scope and ongoing costs, not headline price. Wildly low quotes usually win by omitting work that returns later as change requests.
A freelancer has a lower headline rate, £40 to £100 per hour versus an agency's £80 to £150, and suits small, well-defined jobs. For anything business-critical, an agency is usually better value because you get a team, project management, quality control and continuity, which prevents the costly rework that single-person engagements often produce.
Budget 15 to 25 per cent of the original build cost per year for maintenance. For a £50,000 application that is £7,500 to £12,500 annually, covering security patches, dependency updates, bug fixes and adapting to third-party API changes. Skipping maintenance leads to a forced rebuild later that costs far more than the maintenance would have.
Total cost of ownership is the full cost across the application's life, not just the build. Over five years, post-launch costs (hosting, maintenance, support and iteration) make up 60 to 70 per cent of the total. A £50,000 build commonly becomes a £130,000-plus five-year commitment, so budget the full figure, not the build price.
Build an MVP first if you are validating a new idea or unsure of demand. An MVP costs £15,000 to £35,000, launches in six to twelve weeks, and lets real users prove the concept before you commit to the full spend. If the process is already proven and business-critical, building the full product directly can be more efficient.
Offshore teams are 40 to 60 per cent cheaper on paper, but the real saving is usually smaller once you account for communication overhead, time-zone lag, rework and the cost of a UK manager to oversee the work. For regulated or business-critical projects, the apparent saving frequently disappears. Treat the headline discount with caution.
You should own the source code outright once you have paid for a custom build, but this is not automatic. Make intellectual property ownership and source-code handover explicit in the contract. If a quote is silent on ownership, ask directly, because some arrangements leave you dependent on the original developer to make any future change.
Most UK SME web applications take three to six months from discovery to launch. A simple internal tool can be six to ten weeks; a customer portal three to six months; a SaaS or enterprise platform nine to eighteen months. Discovery quality and prompt client decisions affect the timeline as much as the technical scope.
Yes. Start with a tight MVP, choose off-the-shelf products for generic functions and reserve custom build for your genuine differentiators, and automate repetitive admin rather than building bespoke logic for it. A clear specification before build begins is the single biggest cost saver, because it prevents expensive mid-project changes of direction.
A custom web application in the UK costs £15,000 for a simple internal tool, £30,000 to £80,000 for a typical SME customer portal, and £200,000-plus for a full SaaS product, but the build price is the smaller half of the story. Over five years, ongoing hosting, maintenance, support and iteration make up 60 to 70 per cent of total spend, turning a £50,000 build into a £130,000-plus commitment. The businesses that get the best value are the ones that budget the full five-year figure, compare quotes on scope rather than headline price, watch for the red flags that signal hidden costs, and reserve custom development for processes that genuinely cost them money or cap their growth. Match the right tier to a real, measurable problem, insist on a discovery phase and itemised ongoing costs, and a custom application becomes an investment with a payback period rather than an open-ended liability. The next move is to get a fixed quote against a written specification.
If you are weighing up a custom build and want a clear, fixed-quote scope rather than an open-ended day rate, talk to our team through our web application development services in London and we will model the full five-year cost with you before you commit.
Written by Deen Dayal Yadav, Founder of Softomate Solutions, a London-based software development and AI automation agency in Stanmore (HA7). With over 12 years building software and automation systems for UK businesses, I have scoped, costed and delivered web applications across finance, healthcare, professional services and retail. Softomate Solutions is a UK company registered with Companies House, and we work on fixed quotes agreed after discovery so clients always know the number before the build begins. Learn more about our team and approach or get in touch via our contact page.
We protect the real names of all clients featured in examples and case studies. Every testimonial is from a real client.
More from the blog
Work with us
Book a free 30-minute discovery call with DD and get a personalised automation roadmap.
Softomate Solutions Limited
Deen Dayal Yadav
Online
We use essential cookies to keep the site running. With your permission, we also use analytics cookies to understand how visitors use our site so we can improve it. No data is sold. Privacy Policy
