Bespoke software development in London typically costs between £20,000 and £250,000, with most small-business projects landing between £25,000 and £75,000. A minimum viable product (3 to 5 core features) runs £15,000 to £35,000 over 8 to 16 weeks. An internal operations tool costs £12,000 to £28,000 over 6 to 10 weeks. A customer-facing web app costs £30,000 to £75,000, and enterprise platforms reach £50,000 to £200,000 over 4 to 12 months. London day rates sit 20 to 40 per cent above the UK average: mid-level developers charge £400 to £550 per day and senior engineers £600 to £800. Budget a further 15 to 25 per cent of build cost per year for maintenance, plus £50 to £2,000 per month for hosting. The single biggest budget killer is a vague brief, so most of your savings come before a line of code is written.
Last updated: June 2026
Bespoke software is an application built specifically for your business, your data, and your way of working, rather than a product you rent and bend yourself to fit. The honest rule is simple: choose off-the-shelf SaaS when your process is standard, and choose bespoke when the software needs to match a workflow that gives you a competitive edge or that no product on the market handles well.
Most UK businesses already run a stack of off-the-shelf tools: accounting in Xero, email in Microsoft 365, a CRM, a couple of spreadsheets holding the whole thing together with string. That stack is cheap to start and fast to deploy. It becomes a problem when your team spends hours re-keying data between systems, when you pay per-seat fees for features you never use, or when the product roadmap is owned by a vendor in another country who will never build the one thing you actually need.
Our view, after twelve years of building for UK firms, is that most companies buy bespoke too late rather than too early. They wait until the manual workarounds are costing two full-time salaries a year before they commission a tool that would have paid for itself in eight months. The trigger to go bespoke is rarely a missing feature. It is the cumulative cost of glue work, double entry, and processes that exist only because the software cannot do them.
Here is the honest comparison so you can decide where you sit.
| Factor | Off-the-Shelf SaaS | Bespoke Software |
|---|---|---|
| Upfront cost | Low (often £0 to start) | £12,000 to £250,000 |
| Ongoing cost | Per-seat monthly fees that scale with headcount | Hosting plus maintenance you control |
| Time to value | Days to weeks | 6 weeks to 12 months |
| Fit to your process | You adapt to the tool | The tool adapts to you |
| Ownership | You rent; vendor owns the roadmap | You own the source code and IP |
| Competitive advantage | None: rivals use the same tool | Process becomes a moat |
| Lock-in risk | High: data hostage on exit | Low: you hold the code |
A practical middle path many of our clients take is to keep best-in-class SaaS for commodity functions (payroll, email, accounting) and build bespoke only for the one or two workflows that genuinely differentiate them. You do not need a custom CRM if a standard one works; but if your sales process has rules no product respects, a custom CRM built around your pipeline can pay back faster than another year of per-seat fees. The decision is never "build everything" versus "buy everything". It is knowing which is which.
A bespoke software project in London costs between £20,000 and £250,000 in 2026, and the price is driven almost entirely by scope rather than by your industry or postcode. The figure that matters is not the headline range, it is which category your project falls into, because each category has a tight, predictable band once requirements are clear.
The widest variable is feature count and integration depth. A tool that reads and writes to one database with five screens is a different animal to a multi-tenant platform with payment processing, third-party APIs, role-based permissions, and a public mobile app. Below are the categories we quote against, with London-realistic ranges and the typical delivery window.
| Project type | Typical London cost | Timeline | Good fit for |
|---|---|---|---|
| Minimum viable product (3 to 5 features) | £15,000 to £35,000 | 8 to 16 weeks | Startups validating an idea |
| Internal operations tool | £12,000 to £28,000 | 6 to 10 weeks | Replacing spreadsheets and manual work |
| Customer-facing web application | £30,000 to £75,000 | 3 to 6 months | SaaS, portals, booking platforms |
| Mobile app (iOS and Android) | £35,000 to £90,000 | 3 to 7 months | Consumer or field-team apps |
| Enterprise platform or system | £50,000 to £200,000+ | 4 to 12 months | Multi-department, high-integration |
Notice the overlap and the breadth of each band. An MVP at £15,000 and an MVP at £35,000 are both "an MVP", but the second has more screens, a cleaner design, real authentication, and one or two integrations the first skips. When you ask an agency for a number and they give you one without asking you forty questions first, be sceptical: they are either guessing or quoting a template they will later "discover" needs to cost more.
Our blunt stance on cheap quotes: if one quote comes in at roughly half the others for the same brief, treat it as a warning, not a win. In nine cases out of ten it means one of three things. The team is junior or offshore with thin oversight, the quote silently omits scope you assumed was included (testing, deployment, project management, fixes), or the agency is buying the deal and will recover the margin through change requests once you are locked in. The cheapest quote is rarely the cheapest project.
For founders specifically, the smart play is to spend the smallest amount that proves the idea. A tightly-scoped MVP at £20,000 to £30,000 that gets real users beats a £120,000 "complete" build that nobody validated. If you are weighing a phased custom software build against a big-bang launch, phasing almost always wins on risk and cash flow.
Five things drive the cost of bespoke software: the number and complexity of features, the integrations to other systems, the seniority of the team, the design and user-experience bar, and the non-functional requirements like security, performance, and compliance. Of these, integrations and "invisible" requirements are the ones that blow budgets, because they are the parts buyers forget to mention in the brief.
Features are the obvious driver, but raw feature count is misleading. A "login" sounds like one feature; in reality it can mean email and password, social sign-in, two-factor authentication, password reset flows, session management, and account lockout. One word on a wishlist, a fortnight of work. The skill in scoping is breaking vague nouns into the real work they contain.
Integrations are where projects quietly double. Every connection to an external system (a payment gateway, an accounting package, a courier API, a CRM, an HMRC endpoint) carries authentication, error handling, rate limits, sandbox testing, and the joy of someone else's documentation. A single well-behaved integration might be two to four days. A poorly documented legacy system with no test environment can swallow weeks.
Here are the cost drivers ranked by how often they surprise our clients, from most to least underestimated.
The lever you control most is the brief. Every ambiguity you resolve before development starts is a change request you do not pay for later, at a markup, mid-build. We have seen identical projects come in £15,000 apart purely because one client arrived with a clear, prioritised requirements document and the other discovered what they wanted as they went. If you want to compress your automation or software budget, the cheapest hours are the ones you spend thinking before anyone codes.
London software development costs roughly 20 to 40 per cent more than the UK regional average, driven by higher salaries, office overheads, and the concentration of senior talent competing with fintech and big-tech employers. Whether that premium is worth paying depends on how much you value proximity, time-zone overlap, and the ability to sit across a table during discovery.
London day rates in 2026 break down roughly as follows. These are blended agency rates, which include project management and oversight, not bare freelancer rates.
| Role | London day rate (2026) | Regional UK day rate |
|---|---|---|
| Junior developer | £250 to £350 | £200 to £300 |
| Mid-level developer | £400 to £550 | £325 to £450 |
| Senior developer | £600 to £800 | £475 to £650 |
| Technical architect | £800 to £1,200 | £650 to £950 |
| UX or product designer | £450 to £700 | £375 to £550 |
The London premium buys you three things that genuinely matter on some projects and do not matter at all on others. It buys proximity, so you can run in-person discovery workshops and pin down requirements faster. It buys a deep senior talent pool, so when a project needs an architect who has built payment-critical systems before, that person exists nearby. And it buys accountability that comes from a registered local company you can meet, rather than a contract with a team you will never see.
Our honest take, and it is not the usual London-agency pitch: you do not always need a London team. If your project is a well-defined internal tool with a clear spec and few unknowns, a strong regional UK team will deliver the same quality for less. The London premium earns its keep on projects with high ambiguity, where the cost of misunderstanding the requirement dwarfs the difference in day rate. Discovery is the phase where being in the same room saves the most money, and that is the phase London teams do well.
Where we are genuinely cautious is fully offshore development chosen purely on price. The day rate looks irresistible until you account for time-zone friction, the cost of writing specifications precise enough to survive a 10-hour communication gap, and the rework when something is built to the letter of a brief rather than its intent. A blended model (UK-based product and architecture leadership, with vetted nearshore or offshore delivery for well-specified work) is often the best value. The mistake is outsourcing the thinking, not the typing. We base our own team in Stanmore, north-west London, precisely so the requirements-shaping work happens face to face, where it pays off.
Most bespoke software projects take between 6 weeks and 12 months, with the typical small-business custom application landing at 3 to 5 months from kick-off to launch. The single biggest determinant is not how fast the team codes, it is how clearly the requirements are defined before development begins, which is why the discovery phase deserves its own budget line.
Discovery is the 2-to-6-week phase (most projects need 2 to 4 weeks) where you turn "we want a portal" into a concrete, prioritised set of screens, user roles, data structures, and integration points. Skipping it feels like saving time. It is the most reliable way to lose money. A project that starts coding without discovery typically spends that "saved" time later as rework, scope arguments, and change requests, often at a worse exchange rate than if the thinking had been done up front.
Here is a realistic phase-by-phase breakdown for a mid-sized customer-facing application.
| Phase | Typical duration | What happens |
|---|---|---|
| Discovery and scoping | 2 to 4 weeks | Requirements, wireframes, architecture, fixed quote |
| Design (UX and UI) | 2 to 4 weeks | User flows, visual design, prototype sign-off |
| Development | 8 to 20 weeks | Built in 2-week sprints with demos |
| Testing and QA | 2 to 4 weeks | Functional, security, accessibility, user acceptance |
| Launch and handover | 1 to 2 weeks | Deployment, training, documentation, source-code transfer |
Some of these phases overlap. Testing should not be a block at the end, it should run continuously inside each sprint. Design and discovery can blend. But the structure holds: roughly a fifth of the calendar on thinking, the bulk on building, and a meaningful chunk on making sure it actually works before real users touch it.
Three things reliably stretch a timeline, and you can control all of them. First, slow decisions on your side: if sign-off on a design takes three weeks because the right person is on holiday, the clock keeps running. Second, scope creep: every "while we are at it" feature pushes the date. Third, integration surprises, especially with legacy systems whose owners are slow to grant access. Our advice is to nominate one decision-maker on your side with the authority to sign off quickly, and to front-load every integration so the nasty ones surface in week two, not week twelve. If you are building a web application or a mobile app, the discovery phase is where you buy down almost all of the schedule risk.
The right pricing model depends on how well-defined your requirements are: fixed price suits tightly-scoped projects, time-and-materials suits exploratory work where requirements will evolve, and a dedicated team suits ongoing, long-term product development. The dangerous mistake is forcing a fixed price onto a project nobody has scoped, because the agency prices in a fat risk margin and you pay for uncertainty either way.
Each model shifts risk differently. Understanding who carries the risk tells you what each model really costs.
| Model | Best for | Who carries risk | Watch out for |
|---|---|---|---|
| Fixed price | Clear, stable requirements | The agency | Padded margins; rigid change-request process |
| Time and materials | Evolving or exploratory scope | The client | Open-ended cost; needs a cap and tight oversight |
| Dedicated team | Long-term, ongoing product | Shared | Only economical at scale; needs your product input |
Our recommended approach, and the one we use ourselves, splits the difference. Run discovery as a small, fixed-price engagement first. Discovery is well-defined work (workshops, wireframes, a specification, an architecture), so it prices cleanly. The output is a detailed scope that then lets the agency quote the build itself as a genuine fixed price, because by that point the unknowns are known. You pay a few thousand pounds to remove the uncertainty that would otherwise be priced into the whole build as risk.
Here is the honest rule on fixed price: never accept a fixed-price quote for a full build that was produced without a detailed requirements document. A real fixed price requires the agency to understand exactly what they are committing to. If they hand you a fixed number off the back of a one-hour call and a wishlist, one of two things will happen. Either they have padded the price heavily to protect themselves, or they will hit you with change requests the moment reality diverges from their guesses. Both cost you more than honest time-and-materials with a cap.
Time-and-materials is not the enemy it is made out to be, provided it comes with guardrails: a not-to-exceed cap, weekly burn reports, sprint demos so you see progress, and the right to stop at any sprint boundary. For genuinely exploratory work, that transparency beats a fixed price built on a guess. The model matters less than the discipline around it.
The build price is rarely the full cost of bespoke software. Budget a further 15 to 25 per cent of the build cost per year for maintenance, £50 to £2,000 per month for hosting and infrastructure, plus often-forgotten line items for third-party licences, support, and future enhancements. Over three years, total cost of ownership commonly runs 40 to 75 per cent above the original build figure, and the buyers who get burned are the ones who only budgeted the build.
Maintenance is not optional and it is not a sign of poor work. Software lives in a moving world: browsers update, operating systems change, security vulnerabilities are discovered, third-party APIs deprecate old versions, and your business needs evolve. A maintenance retainer covers security patching, dependency updates, bug fixes, monitoring, and minor tweaks. Skipping it does not save money, it defers and compounds the cost, because a system left unmaintained for two years needs a small fortune to bring back to safe and current.
Here is what a realistic three-year total cost of ownership looks like for a £50,000 customer-facing web application.
| Cost line | Year 1 | Years 2 to 3 (each) |
|---|---|---|
| Initial build | £50,000 | £0 |
| Hosting and infrastructure | £1,800 to £6,000 | £1,800 to £6,000 |
| Maintenance and support (20%) | £10,000 | £10,000 |
| Third-party licences and APIs | £600 to £4,000 | £600 to £4,000 |
| Enhancements and new features | £0 to £8,000 | £5,000 to £15,000 |
| Approximate yearly total | £62,400 to £78,000 | £17,400 to £35,000 |
The line buyers forget most often is third-party costs that scale with usage. A payment gateway takes a percentage of every transaction. A mapping API, an email-sending service, an SMS provider, or an AI model charge per call. At launch these are trivial. At scale they can become your largest software cost, so it is worth modelling them at your projected volumes, not your launch-day volumes.
The other quiet cost is your own time. Bespoke software needs your input during the build, your people for user-acceptance testing, and ownership after launch. A good agency reduces this load with clear documentation and training, but it never reaches zero. Our stance is that the agencies who pretend maintenance and total cost of ownership do not exist are the ones to avoid, because the cost does not disappear, it just lands on you unannounced in year two. Honest total-cost planning up front is a feature, not a sales obstacle.
The right questions separate a partner who will deliver from a vendor who will disappoint, and the most important ones are about ownership, process, and people, not technology. Ask who owns the source code, who actually writes it, and what happens when something goes wrong, then watch how comfortably they answer. Evasion on any of these is the reddest flag there is.
Here are the questions we tell our own clients to ask any agency they are evaluating, including us.
Now the red flags. Some are obvious; the dangerous ones are subtle.
| Red flag | Why it matters |
|---|---|
| A fixed price quoted without detailed requirements | Either padded or set up for change requests later |
| A quote roughly half the others | Usually junior, offshore-without-oversight, or missing scope |
| Vague or evasive on source-code ownership | You may not own what you paid for |
| No discovery phase, straight to building | They will discover the cost on your budget |
| Cannot name the people who will do the work | Likely subcontracted to an unknown team |
| No mention of testing or it costs extra | You inherit the bugs in production |
| Pressure to sign quickly | Good agencies have a pipeline; they can wait for a good fit |
Our strongest single piece of advice: insist that source-code and IP ownership transfer to you on final payment, in writing, before you start. Under UK law the default position is that, in the absence of a written agreement, the developer who created the code can retain copyright even though you paid for it. That is not a quirk, it is the standard rule, and it catches buyers out constantly. A reputable agency will hand over a clean assignment clause without blinking. If you get resistance, walk away, because everything downstream (your ability to switch suppliers, raise investment, or sell the business) depends on actually owning the asset.
Any software that handles personal data must comply with UK GDPR and the Data Protection Act 2018, and any public-facing service should meet WCAG accessibility standards, while the contract must explicitly assign intellectual property and source-code ownership to you. These three areas (data protection, accessibility, and IP) are where buyers get hurt, and they are the areas almost every cost-focused competitor guide ignores entirely.
Start with data. If your software stores, processes, or even just touches the personal data of UK individuals (names, emails, anything that identifies a person) it falls under UK GDPR and the Data Protection Act 2018, regulated by the Information Commissioner's Office. The build needs to bake in data-protection-by-design: collecting only what you need, securing it properly, providing a lawful basis for processing, and supporting individual rights like access and deletion. Retrofitting compliance after launch is expensive and risky. A penalty under UK GDPR can reach the higher of £17.5 million or 4 per cent of global turnover, so this is not a corner to cut. If your software sends marketing messages, the Privacy and Electronic Communications Regulations (PECR) add consent rules on top.
Accessibility is the second area buyers underweight. The Equality Act 2010 requires that services do not discriminate against disabled users, and the practical standard for digital services is WCAG (the Web Content Accessibility Guidelines). For public-sector bodies and their suppliers, meeting WCAG 2.2 to level AA is a legal requirement under the public-sector accessibility regulations. For private businesses it is strongly advisable both to avoid discrimination claims and because accessible software is simply better software that reaches more customers. Building accessibility in from the design phase costs a fraction of bolting it on afterwards.
Here is a compliance checklist to run against any bespoke project before you sign.
| Compliance area | What to confirm | Who governs it |
|---|---|---|
| Data protection | UK GDPR by design; lawful basis; data minimisation; encryption; UK or adequate data residency | ICO |
| Marketing consent | PECR-compliant opt-in for email and SMS | ICO |
| Accessibility | WCAG 2.2 AA target; tested with assistive tech | Equality Act 2010 |
| IP and source code | Written assignment of copyright and IP to you | Copyright, Designs and Patents Act 1988 |
| Security | Secure coding, audit logging, breach process; Cyber Essentials a plus | NCSC guidance |
Our firm stance: compliance is not a feature you add at the end, it is a constraint you design within from day one. An agency that treats UK GDPR and accessibility as afterthoughts is showing you how it treats quality generally. When we scope a build, data protection and accessibility are line items in discovery, not surprises in user-acceptance testing. It costs a little more up front and saves a great deal in penalties, rework, and reputation. If your project involves customer data at any scale, make sure your agency can talk fluently about the ICO's data-protection-by-design expectations before you hand over a brief.
The Softomate process runs in five stages, from a fixed-price discovery that removes uncertainty to a clean handover where you receive full source-code ownership, documentation, and training. We deliberately front-load the thinking so that the build itself can be quoted as a genuine fixed price, with no surprise change requests once we agree scope. Most projects start at £18,000 for a focused internal tool and £30,000 for a customer-facing application.
We are a London-based software and automation agency in Stanmore (HA7), and we build for UK businesses that want a partner, not a black box. Here is how an engagement works.
Here is the indicative timeline and what each stage costs as a share of a typical mid-sized build.
| Stage | Typical duration | Share of budget |
|---|---|---|
| Discovery and scoping | 2 to 4 weeks | 5 to 10% |
| Design and prototype | 2 to 4 weeks | 10 to 15% |
| Agile build | 8 to 20 weeks | 55 to 65% |
| Testing and UAT | 2 to 4 weeks | 10 to 15% |
| Launch and handover | 1 to 2 weeks | 5% |
What makes us different is not a secret technology, it is the commercial honesty: a fixed quote after real discovery, clear ownership transfer written into the contract, no scope games, and a single point of contact who actually understands your business. We also build a great deal of AI automation into custom systems where it earns its place, from AI chatbots that handle customer queries to back-office process automation that removes manual data entry. If your project would benefit from an off-the-shelf ERP foundation rather than a from-scratch build, we also deliver Odoo ERP implementation, because the right answer is sometimes "configure, do not code".
Custom software in London costs between £20,000 and £250,000, with most small-business projects between £25,000 and £75,000. A minimum viable product runs £15,000 to £35,000, an internal tool £12,000 to £28,000, and a customer-facing web app £30,000 to £75,000. Scope and integrations drive the price far more than your industry does.
Most bespoke software takes 3 to 5 months, ranging from 6 weeks for a simple internal tool to 12 months for an enterprise platform. The biggest factor is how clearly requirements are defined before coding starts. A 2-to-4-week discovery phase up front is the most reliable way to keep the overall timeline on track.
Bespoke software is worth it when your process gives you a competitive edge or when no off-the-shelf product fits your workflow, and when per-seat SaaS fees and manual workarounds cost more than a custom build over three years. For standard, commodity functions like payroll or email, off-the-shelf is cheaper and faster. Most businesses should use both strategically.
You should own it, but only if your contract says so. Under UK law, without a written assignment, the developer can retain copyright even though you paid. Always insist on a written clause transferring source code and intellectual property to you on final payment. A reputable agency provides this without resistance; hesitation is a serious red flag.
Budget 15 to 25 per cent of the build cost per year for maintenance, £50 to £2,000 per month for hosting, plus third-party licences and usage-based API fees. Over three years, total cost of ownership commonly runs 40 to 75 per cent above the original build figure. Maintenance is not optional; deferring it compounds future cost.
London costs roughly 20 to 40 per cent more due to higher salaries, office overheads, and competition for senior talent with fintech and big-tech employers. The premium buys proximity for in-person discovery, a deeper senior talent pool, and local accountability. For well-defined projects with low ambiguity, a strong regional team can deliver the same quality for less.
Use fixed price for tightly-scoped work and time-and-materials for exploratory projects where requirements will evolve. The best approach is to run discovery as a small fixed-price engagement first, then quote the build as a genuine fixed price once unknowns are resolved. Never accept a fixed-price full build quoted without a detailed requirements document.
Discovery is a 2-to-6-week phase that turns a vague idea into a concrete specification with screens, user roles, data structures, and integration points. Yes, you need it. Skipping discovery does not save money; it moves the cost later into rework and change requests, usually at a worse rate than if the thinking had been done up front.
Any software handling personal data must follow UK GDPR and the Data Protection Act 2018: collect only what you need, secure it properly, have a lawful basis, and support data-subject rights like access and deletion. Build data-protection-by-design from the start. The ICO regulates this, and penalties reach the higher of £17.5 million or 4 per cent of global turnover.
List the problem you are solving, who the users are, the must-have features versus nice-to-haves, the systems it must integrate with, your budget range, and your deadline. Prioritise ruthlessly. The clearer your brief, the more accurate the quote and the fewer change requests later. A good agency will then run discovery to fill the gaps before quoting a fixed price.
Bespoke software in London costs £20,000 to £250,000, with most small-business projects between £25,000 and £75,000, delivered over 3 to 5 months. The numbers that decide your budget are scope, integrations, and team seniority, not your postcode. Plan for the full picture: a further 15 to 25 per cent of build cost per year for maintenance, hosting from £50 to £2,000 per month, and a three-year total cost of ownership commonly 40 to 75 per cent above the build figure. Protect yourself by insisting on a real discovery phase, written source-code and IP ownership, and UK GDPR and accessibility built in from day one. Be sceptical of any fixed price quoted without detailed requirements, and of any quote that comes in at half the others. The cheapest project is almost never the cheapest quote. Get the brief right, choose the right model, and bespoke software becomes one of the best investments your business can make.
If you are planning a custom software project and want a fixed-price quote built on real discovery rather than guesswork, explore our bespoke software development service in London or get in touch for a no-obligation scoping conversation.
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, Deen has led custom builds across CRM, ERP, mobile, and web platforms for companies that needed software to fit their process rather than the other way around. Softomate Solutions is a UK company registered at Companies House and works with founders, operations leaders, and finance directors across London and the wider UK. Learn more about Softomate Solutions and our approach.
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
