Bespoke software or off the shelf: what decides it for a UK business?

Buy off the shelf for the work every business does the same way. Build the part your customers touch that sets you apart, where no product works the way you sell. Between the two sits a third route that most guides skip: joining up the systems you already pay for. One question decides most of it. Does this part of the business win you customers, or does it just need to work? We build software for a living, and our advice is still to buy most of it.

Start with the result you’re paying for

Software is a means to an end, so start with the end. For most businesses that means more customers finishing a quote, booking or sign-up, and more of them coming back. You can measure all of it: how many people start a quote and how many finish, how long they wait for a first answer, and how many book again.

Time saved in the back office matters too, but it comes second. A faster invoice run is pleasant. A customer who books in two minutes instead of waiting for a call back is revenue.

So look first at the parts of your business a customer actually feels, and ask whether they’re good enough. Build or buy comes after that, one part at a time.

The test: does this part make you different?

Split your operation into its parts before you decide anything. Accounts, payroll, email and file storage work much the same in most businesses. Nobody picks a supplier because its payroll is unusually creative.

Other parts are where you compete. That means how a customer gets a price, and how a job is booked, tracked and followed up. It also means the thing your staff do that a rival’s staff can’t. Those are the parts worth owning.

The UK government runs the same test on its own services. Its guidance on choosing technology suggests mapping out your components to see which are mature enough to treat as stable and which will change frequently. Stable parts suit a product someone else maintains. Parts you’ll keep changing as you learn what customers want suit software you control.

Three routes, not two

Buy off the shelf

Buying gets you working software quickly, and someone else pays for the updates and security fixes. For the stable, standard parts of a business it is usually right. Building your own payroll is an expensive hobby.

The cost is fit. You bend your process to suit the tool. That’s fine for payroll. It hurts when the tool shapes the one thing customers judge you on, like a booking form they give up on halfway.

Before you sign, check:

• how you get all your data out, in what format and how often
• how long the contract ties you in, and what leaving costs
• whether it has an API, the connection point that lets other software read and write its data
• how the price changes as you add users, sites or customers

The same logic applies to websites, and we cover the trade-off in why not just use a website template or builder.

Connect what you already have

Plenty of businesses don’t need new software. They need the tools they already pay for to talk to each other. Staff copying details from the booking system into the accounts package is a connection problem. So is typing the same customer into three places.

Joining systems through their APIs keeps the products you’re happy with and removes the manual step between them. It’s often the quickest fix, and nobody has to learn a new system. The hours add up, too. When we automated billing and payroll for The Home Help Company, the CEO got 30 hours a month back.

This route only works if the tools let you in. A product with no API, or one that charges heavily for access, pushes you back towards replacing it.

Build your own

Build when three things are true: the part is where you compete, no product does it your way without a fight, and you’ll keep changing it as you learn.

It costs more up front, and you take on running it. In return it works the way you sell, changes when you decide and carries no per-user licence fee at renewal. Pristine Condition’s movement tracking is one we built. About 150,000 warehouse staff use it, and it works offline.

You rarely build everything. A sensible build buys the standard pieces, such as payments, login and email sending, and writes only what’s yours. That keeps the build smaller and the risky part in focus.

If you’re still testing whether customers want the thing at all, test before you build. How to validate an app idea covers cheap ways to do that.

What it really costs over the years

Compare the total cost of ownership, not the first invoice. That means everything you’ll pay while you use it: setup, licences or build, running costs, changes and, one day, leaving. The government’s Service Standard asks its own teams to do exactly this, and to minimise the cost of changing direction later. That’s sound advice well outside Whitehall.

Off the shelf usually looks cheaper in year one, and for standard work it often stays cheaper. The sums turn when per-user fees grow with your headcount, when workarounds cost staff hours every week, or when a clunky customer journey loses you sales.

For a sense of what a build costs, our LaunchPad prices are fixed and public. Starter is £25k over two months, Growth is £60k over three and Scale is £100k over four, all excluding VAT. The price covers everything up to launch, including developer accounts, hosting, licences and the third-party services the build needs. See LaunchPad pricing, get a rough figure from the app cost estimator, or read what an app costs to build in the UK for context.

Lock-in, data and who owns what

Whichever route you take, plan how you’d leave before you start.

With off the shelf, lock-in lives in the contract and the data export. Keep full control of your data, avoid long contracts where you can, and prefer tools that exchange data in standard formats any supplier can work with.

With a build, lock-in lives in who holds the code, the hosting account and the app store listings. A build you don’t own is an off-the-shelf product with exactly one customer.

So get ownership in writing. With Contic, you own the code and IP, and our contract assigns it to your company. The AWS account, the GitHub repositories and the App Store and Google Play apps transfer to you. We choose widely used, well-supported frameworks, so another team can hire for them. There’s more in who owns the code when a developer builds it and do I own the code and IP.

Where AI changes the sums

Customers now expect service inside the software they already use: an answer at 10pm, a quote without a phone call, a booking moved in a message. That wins more business than automating paperwork in the back office.

So the AI question is the same as the software one. Can it make you more competitive and bring in more money? And how far is your current product from offering it?

If the AI built into tools you already pay for does the job, use it. If it needs your prices, your rules and your customer journey, it usually has to be built on top of your own systems. That’s one more reason to own the customer-facing part. ChatGPT vs custom AI: when to buy, configure or build goes through the options.

An AI agent is software that answers questions and takes simple actions for a customer or member of staff. In our experience, a simple agent can cost as little as £500 to set up, but £5,000 is realistically the lower bound for a well-managed, organised project. Where yours lands depends mostly on the state of your data and what you want the agent to do. What an AI agent costs to build and run has the detail.

A worked example

Take a made-up regional removals firm with around 30 staff. Jobs come in by phone and email, and quotes go out the next day.

• Accounts and payroll: buy. Nothing about how the firm pays its staff wins it work.
• Rotas and van scheduling: buy, then connect. A standard scheduling tool does the job. Connecting it to the accounts package turns finished jobs into invoices without retyping.
• The quote: build. This is where customers decide. A customer who waits a day for a price may book whoever answered first.
• AI on top: add it to the part they own. An agent can answer questions and book survey slots out of hours, using the firm’s own prices and rules.

The firm ends up with one thing built, two things bought and one connection between them. It measures the build by quotes started against quotes finished, time to first answer and repeat bookings. The connection is measured in office hours saved, which is the smaller prize.

Questions to ask your team or supplier

Take these into the meeting, whoever you’re talking to.

• Which parts of our business do customers feel, and which just need to work?
• What will this cost over five years, including per-user fees, changes and leaving?
• If you build it, who owns the code, the hosting account and the app store listings?
• What will we see working, and when, before most of the budget is spent?
• How will we know it worked, in numbers we already track?

If you’re weighing up contract types, fixed price vs time and materials covers which one protects you.

Where Contic fits

If the answer is to build, LaunchPad is our fixed-price route. It runs in four steps: Discover, Design, Build and Launch. Discovery sets the exact scope within your tier, and if you aren’t confident in the plan after it, you get your money back.

We start with the hardest, most uncertain feature, not the login screen. You see working software at a weekly demo, and senior engineers review and own everything that ships. The team that built your product is the team that looks after it, if you want us to.

If you’re not sure which route fits, talk to us about what you’re weighing up. We’ll give you our honest view on whether to buy, connect or build. We’d rather tell you not to build than bill you for it.