Custom website development means building a site or web application from scratch rather than configuring a platform. It is the right answer less often than agencies selling it suggest — and when it is right, it is decisively right.
Here is the uncomfortable version first: most businesses asking for a custom website do not need one. WordPress, Webflow, Shopify and Squarespace handle an enormous share of what gets commissioned as bespoke work, at a tenth of the cost, and they do it well.
We build custom sites and web applications. We also turn this work down regularly.
When a platform is enough
Be honest about these before spending £10,000 or more.
It is mainly content. Pages, posts, images, a contact form. WordPress or Webflow will do this better than a custom build, because thousands of people have already found and fixed the problems you would hit.
It is standard e-commerce. Products, variants, cart, checkout, shipping rules. Shopify has solved payment edge cases, tax rules and fraud handling that would take months to rebuild badly.
You want to edit it yourself. Platforms have mature editing interfaces. A custom CMS good enough for non-technical staff to use comfortably is a substantial build in itself.
Budget is under £10,000. A custom build at that budget is a worse version of what a platform gives you immediately.
You need it in three weeks. Platforms win on speed by a wide margin.
The honest test: write down the five things your site must do. If a platform does four of them out of the box, use the platform and solve the fifth with an integration.
When custom is genuinely the right answer
It is an application, not a website. Users log in and do something — manage bookings, submit claims, track orders, run calculations, collaborate. Once there are user accounts and real business logic, platforms start costing more than they save.
The business logic is specific to you. Pricing rules, availability calculations, approval workflows, allocation logic. Anything where the rule is yours and no plugin implements it.
It must integrate deeply with your systems. Live stock from your ERP, pricing from your accounting system, availability from your scheduling tool. Shallow integrations are fine on platforms; deep two-way ones are where they break down.
Plugin sprawl has become the problem. A WordPress site held together by twenty-two plugins, four of which conflict and two of which are unmaintained, is no longer cheaper. It is a maintenance liability with a monthly fee.
Performance or scale genuinely matters. Measured, not assumed. If page speed demonstrably affects your revenue, a custom build gives you control a platform cannot.
You have outgrown the platform’s ceiling. You know this is you because you are paying for workarounds — three apps, a Zapier chain and someone doing twenty minutes of manual work daily.
Compliance or data residency requires it. Specific hosting, audit or access rules a platform cannot satisfy.
Custom website vs WordPress vs Webflow
| Platform (WordPress / Webflow) | Custom build | |
|---|---|---|
| Initial cost | £2,000 – £15,000 | £10,000 – £80,000+ |
| Time to launch | 2–8 weeks | 8–20 weeks |
| Editing content | Mature, familiar | Build it or use a headless CMS |
| Business logic | Plugins, with limits | Anything |
| Deep integration | Possible, fragile | Designed in |
| Ongoing cost | Hosting + plugin licences + updates | Hosting + maintenance |
| Security | Plugin surface area is the risk | Smaller surface, your responsibility |
| Ceiling | Real, and you will find it | Effectively none |
| Right when | Content, standard commerce, speed to market | Application behaviour, specific logic, deep integration |
A pattern worth knowing: marketing site on a platform, application on a custom build, same domain. Marketing team edits freely; the application does what only it can do. This is frequently the correct architecture and almost nobody proposes it, because it means a smaller invoice.
What we build
Web applications — users log in and do work. Portals, dashboards, booking systems, claims handling, internal tools with external access.
Customer portals — order status, documents, invoices, stock levels, self-service. Usually integrated with an ERP or accounting system, and usually justified by the call volume they remove.
Headless builds — custom front end for speed and control, proven CMS behind it so your team edits content without touching code. Best of both for content-heavy sites that also need application behaviour.
Platform rescue — a site that has outgrown its platform, migrated to something maintainable without losing search rankings. Redirect mapping and URL preservation planned before anything moves.
Performance rebuilds — where measured page speed is costing measurable revenue.
What it costs in the UK
| Scope | Typical range | Duration |
|---|---|---|
| Discovery & technical approach | £2,000 – £6,000 | 1–2 weeks |
| Marketing site, custom front end + headless CMS | £10,000 – £30,000 | 6–12 weeks |
| Web application, single user type | £15,000 – £45,000 | 8–16 weeks |
| Web application, multiple roles + integrations | £45,000 – £120,000 | 16–30 weeks |
| Platform migration with SEO preservation | £8,000 – £35,000 | 6–14 weeks |
| Maintenance retainer | From £800/month | Ongoing |
Our minimum build engagement is £10,000. Below that, a platform will serve you better and we will say which.
For context: ITJobsWatch reports a median advertised UK contractor day rate of £525 for software developers in the six months to 1 October 2026, London 15–30% higher. Divide any quote by weeks, then by five, to see what it implies.
What we include as standard
Not upsells. These are part of the build.
Performance. Core Web Vitals measured and met, not promised. Images optimised, scripts minimal, static where static works.
Accessibility. Keyboard navigation, contrast, labels, semantic markup, screen reader testing. Beyond being correct, UK public sector procurement requires it and many enterprise buyers now ask.
Technical SEO. Semantic markup, structured data, clean URLs, correct canonicals, XML sitemap, sensible internal linking. A fast site with broken markup ranks badly.
Security. Validation at trust boundaries, proper authentication and authorisation, current dependencies, no secrets in source control, HTTPS enforced.
Analytics that respect UK GDPR. Configured properly, with a lawful basis and a cookie approach that is not theatre.
Documentation and handover. Source code, deployment instructions, infrastructure in your accounts. All IP assigns to you on final payment.
Performance — what we actually mean
“Fast” is a claim. These are the measurements, and you can hold any supplier to them.
Core Web Vitals are Google’s user-experience metrics and they affect rankings:
| Metric | What it measures | Target |
|---|---|---|
| LCP (Largest Contentful Paint) | How long until the main content appears | Under 2.5s |
| INP (Interaction to Next Paint) | How quickly the page responds to a click or tap | Under 200ms |
| CLS (Cumulative Layout Shift) | How much the page jumps about while loading | Under 0.1 |
How we hit them: static HTML where the content is static, images sized and served in modern formats, minimal JavaScript, fonts loaded without blocking rendering, and anything below the fold deferred.
How platform sites usually miss them: a theme loading five stylesheets and twelve scripts, a slider plugin, three tracking tags and unoptimised images. Each addition is small; the aggregate is a four-second load.
Measure on a real phone on 4G, not on your office fibre on a desktop. That is where your customers are, and that is what Google’s field data records.
Headless, explained without jargon
“Headless” sounds technical and the idea is simple.
A traditional website has the content and the presentation welded together — WordPress stores your text and also decides how it looks. Headless separates them. The content lives in a content management system; a separate front end fetches it and decides how to present it.
Why it matters to you:
- Your team edits content in a familiar interface, without touching code
- The site is fast, because the front end can be plain static HTML rather than a page assembled on every request
- The same content can feed several things — website, mobile app, customer portal — without being written twice
- Security surface is smaller. There is no admin login on the public site to attack.
The trade-off: it is two systems rather than one, so the initial build costs more than a theme. For a content site that also needs application behaviour, it is almost always the right architecture.
Migrating without losing your rankings
The most expensive mistake in this work is a migration that loses search visibility. It is preventable and it is routinely not prevented.
What we do before anything moves:
- Crawl and record the existing site — every URL, its current ranking, its traffic
- Map every old URL to a new one. Every single one, not the top fifty
- 301 redirects, never 302s. Permanent redirects pass ranking signals; temporary ones do not
- Preserve or improve metadata — titles, descriptions, headings, structured data
- Launch and monitor daily for the first fortnight in Search Console, with a rollback ready
A migration that drops traffic 40% has cost more than the build, and that outcome is a planning failure rather than bad luck.
What we hand over
Specified, because this is where the industry is weakest.
- Source code, in a repository you own, from week one rather than at the end
- Documentation — architecture, deployment, configuration, known limitations
- Deployment instructions a competent engineer who has never seen the project can follow
- Hosting in your accounts, on your billing, on your keys
- Domain and DNS under your control
- A dependency list with licences, so you know what the site is built on
- All intellectual property, assigning on final payment
What we do not do: host your site on our platform, retain deployment access as leverage, or licence back to you code you paid for. Why this matters more than it sounds.
Who we build for
Good fit: businesses whose website is doing real work — taking bookings, serving logged-in customers, exposing data from an internal system, replacing phone calls with self-service. Build budget from £10,000.
Sectors where a custom web application usually pays: wholesale and distribution needing trade customer portals with live stock and account pricing; professional services needing client document portals; healthcare and clinics needing booking with real availability rules; construction needing subcontractor document submission; and logistics needing customer-facing order tracking.
The common thread: every one of these replaces phone calls and emails with self-service, and the saving is measurable in hours before it is measurable in anything else.
Frequently asked questions
What is custom website development?
Building a website or web application from scratch rather than configuring an existing platform. It makes sense when the behaviour you need is specific to your business, and it does not when a platform already does the job.
How much does a custom website cost in the UK?
A custom marketing site with a headless CMS runs £10,000–£30,000. A web application runs £15,000–£45,000 for a single user type and £45,000–£120,000 with multiple roles and integrations. £10,000 is our minimum build engagement.
Is a custom website better than WordPress?
Not inherently. WordPress is better for content sites and standard requirements. Custom is better once you need real business logic, deep integration or application behaviour. A frequent best answer is both — platform for marketing, custom for the application.
How long does it take?
A custom marketing site takes six to twelve weeks. A web application takes eight to thirty weeks depending on scope. Discovery takes one to two weeks first.
Will I be able to edit the content myself?
Yes, if you want that — we use a headless CMS so your team edits content without touching code. If content never changes, we can skip the CMS and reduce the cost.
Will I lose my Google rankings if I rebuild?
Not if the migration is planned. Every URL mapped, 301 redirects, metadata preserved, Search Console monitored daily after launch. Rankings are lost when this is treated as an afterthought.
Do I own the website?
Yes. All IP assigns on final payment, including source code and documentation, hosted in your accounts on your billing.
What ongoing costs are there?
Hosting (typically £10–£150/month depending on scale), domain, any third-party services, and optionally a maintenance retainer from £800/month. No licence fees, because there is nothing to licence.
Can you work with our designer?
Yes. We can build from supplied designs, or handle design ourselves. If you have a brand and a designer already, keeping them is usually the better outcome.
What if we need an app later?
Build the web application with that in mind and the API serving the site can serve a mobile app too. Worth deciding early, because retrofitting it is more expensive than planning it.
Common questions about rebuilding
“Will the new site look the same?” Only if you want it to. A rebuild is the natural moment to fix design problems, but it does not require it — we can rebuild on the same design if the problem is purely technical.
“Can we do it in phases?” Usually yes. A common approach is to build the application first, keep the existing marketing site running, and replace that later. Two smaller projects are easier to fund and easier to recover from than one large one.
“What happens to our existing content?” Migrated, with URLs preserved. Content migration is quoted separately for large sites because, exactly as with data migration, the volume matters far less than the consistency.
“How long will our team be involved?” Three to five hours a week during discovery, one to two during the build, more around content migration and training.
Next step
A 45-minute call. We will scope what a custom build would cost you and what it would be worth — with the five-year comparison against platform licence fees.
Written by Jas Singh, Founder. Last updated 8 October 2026. Day rate figures from ITJobsWatch UK contract data for the six months to 1 October 2026.
