Skip to content

Software development

You know what you want built. We build it properly.

Websites, ecommerce, portals, internal tools, booking systems, mobile apps and the design systems behind them, for the client who arrives with the brief already written.

Price
Fixed scope, fixed price
Timeframe
Weeks rather than months. Agreed with the scope
You get
The thing on your brief, built properly. Code and data owned outright
Best for
You already know what you want built

Websites and ecommerce. Anything with users, data and logic behind a login: a portal your customers sign into, a tool your team uses every day, a booking or ordering system. An app for the people who are never at a desk. And the design and brand systems that sit behind how all of it looks.

You've thought it through, you know what you need, and what you're short of is someone competent to build it and still be here afterwards. Twenty years of doing exactly that: standard technology, documented, quoted at a fixed price, looked after in year three by the people who wrote it.

What we build.

Eight kinds of job, and the same team builds all of them. If what you want isn't on this list it's usually a version of something that is. Ask.

Websites and ecommerce

Design and build as one job: we don't sell a design phase that produces pictures and then quote separately to build them. Marketing sites, landing pages, and shops, usually on an established platform rather than from scratch, because for most businesses that's cheaper to own.

Customer portals

Somewhere your clients sign in to see their own data, place an order, upload a document or check where something is, instead of emailing someone who then looks it up.

Internal tools

The application a team uses all day. Usually replacing a spreadsheet that four people edit and nobody trusts.

Booking and ordering

Availability, payments, confirmations, and the admin side that makes it survivable when it gets busy.

Mobile apps and PWAs

For work that happens away from a desk. We'll tell you honestly whether you need a native app or whether a progressive web app does the job for a fraction of the cost. Usually the latter.

Integrations

Connecting what you run to what you're buying: the accounts package, a payment provider, a supplier's API, or the specialist system your industry runs on that nobody outside it has heard of. Usually the fiddly part, and usually underestimated by whoever quoted you.

Design and brand systems

The machinery behind how things look, rather than the look itself: identity and tone at one end, design tokens and a component library at the other, so a value lives in one place and the tenth page matches the first. We built our own before we sold it to anyone.

Taking over someone else's code

Inheriting an application or a site whose original developer has gone. We audit it honestly first and tell you if the answer is to rebuild rather than rescue.

Why this costs less than it used to.

Building software costs less than it did five years ago, and we'd rather explain why than have you assume something is missing from the quote. It no longer takes a team of ten. It takes a lead architect who works out how the thing should be built, and AI doing a great deal of the typing under supervision.

That changes the economics, not the standard. The decisions that determine whether software survives — how it's structured, which technology it runs on, how the data is modelled — are still made by people who have made them before and had to live with being wrong. AI is very good at producing code quickly and has no opinion about whether it should exist.

Here's an example: the quoting, job management and invoicing system a small contractor would have been quoted for five years ago, against the same thing today. Same software, same standard. What changed is how it gets written.

Our rate hasn't changed. The number of days has. Experienced people cost roughly what experienced people have always cost; there are simply far fewer days of it, because the part that used to take months is the part a machine is now genuinely good at.

And it changes who can afford custom software at all. At sixty thousand a ten-person contractor couldn't, so the decision was made for them: take the nearest thing off the shelf and work around the parts that don't fit.

Which is the comparison worth making, and it isn't with the licence fee. Look at what happens around the software you already pay for. Someone exports it into a spreadsheet every week, because the report doesn't answer the question you actually ask. Someone types the job into the accounts package a second time. Someone keeps the real schedule in a notebook, because the one on screen is only accurate for half the team. That is a second system, running alongside the one on the invoice, and it is paid for in hours rather than in licences, which is exactly why it never gets costed next to them.

What you get back is room. Most businesses this size are not short of demand: they are short of the hours the admin eats, and that is the ceiling holding revenue where it is. Take the second system away and the ceiling lifts — more work through the same team, without more overhead. The monthly bill afterwards is smaller than a licence bill, and the money you put into building it buys something you own outright rather than something you rent: an asset on the balance sheet at the end of three years, instead of three years of receipts.

A landing page and a customer portal are still very different numbers, so we quote the job rather than publishing a rate. How long it takes is agreed with the scope, before anything starts, and it's weeks rather than months. It's invoiced monthly rather than as one lump at the end.

An example jobCustom quoting, job management and invoicing for a ten-person contractor

Five years ago

£60,000

Six months, a developer on it full time, every line written by hand.

Today

£10,000

A lead architect, AI doing much of the typing under supervision, and a working session with you every week.

Not sure what's right for you?

Take the 5-minute assessment. 12 quick questions, and we'll tell you honestly whether you need a partnership, a one-off build, or nothing from us at all.

Take the 5-minute assessment

Sound like the right size of job? Half an hour, no charge, and you'll know.

Book a 30-minute call

The quote is the price. A fixed number against a written scope, agreed before anything starts, and it does not move. If we estimate badly, that is our problem and not your invoice. The only thing you will ever pay extra for is something you asked for later, priced and agreed with you before we build it.

So what are you paying us for?

The stack is a decision, not a habit.
Most shops build everything in whatever they already know. We choose per project, and the right answer is usually boring: mainstream, well-supported technology that another developer could pick up in year three. Choosing badly here is the mistake nobody notices for two years and then can't undo cheaply.
Someone who's been wrong before reads it.
Twenty years of building this kind of thing, on every codebase we ship. That's the part a model can't supply: it will write whatever you ask for, confidently and quickly, including the thing that shouldn't exist. Experience is mostly knowing which request to argue with.
Security is continuous, not a phase at the end.
The risk with generated code isn't that it fails: it's that it looks right and is quietly wrong. Anything touching sign-in, payments or personal data is written or reviewed by a person. Dependencies are the standard maintained ones rather than whatever got suggested. Nothing ships on a machine's say-so.
It runs somewhere sensible afterwards.
A mainstream cloud platform, configured properly, backed up, and monitored, so a failure at three in the morning is our problem before it is yours. Plenty of builds are handed over at go-live and quietly rot. This is the part that decides whether year three is calm or expensive.

None of that is visible on the day you accept the work. All of it is visible eighteen months later, which is the only timescale worth judging software on: Touchline, built for Merthyr Town FC, has been taking payments on a phone in a matchday queue since 2024, and Rhydycar West has been live and looked after for years.

The questions this usually raises.

Usually a progressive web app does the job at a fraction of the cost, and it doesn't need app store approval every time you change something. Native earns its keep when you need the camera, offline work, or push notifications people actually act on. We'll tell you which before you spend anything.

Sound like what you need?

Half an hour, no charge. Tell us what the job is and we'll tell you what we'd do, what it'd cost, and whether it's worth doing at all.