I run a studio that builds custom software, and I’m going to open by talking you out of it. Most of the time you should just buy the off-the-shelf tool. If a £30-a-month app does ninety per cent of what you need, and thousands of businesses like yours already run on it happily, building your own version is usually a waste of money and a year of your life you won’t get back.
So this isn’t a “bespoke is always better” pitch. It’s the test we actually use to work out which side of the line a business sits on, because getting it wrong in either direction costs real money. Build when you should have bought, and you’ve burned a year on something a subscription did for £30. Keep buying when you should have built, and you bleed it slowly instead, a bit every month, until you’ve spent the price of a custom build three times over and have nothing to show for it.
The off-the-shelf trap nobody mentions
Off-the-shelf feels safe because each subscription is small. It stops being safe because you never buy just one.
Research into business software spend keeps turning up the same picture. CloudZero’s roundup of SaaS statistics puts the average company on well over a hundred separate apps, and Zylo’s work on unused software found that around half the licences a business buys go unused or barely touched, with roughly 30% of total spend wasted on tools nobody opens, straight duplicates, and quiet auto-renewals. Small firms are more exposed to this, not less, because there’s no IT team sitting between the card and the signup form. Nobody is watching the statement. The £19-a-month thing somebody trialled in March is still being billed in November.
On the ground it looks like a business running on six tools that don’t talk to each other. The quoting app. The booking app. The invoicing app. A spreadsheet stitching the three of them together. And the two people whose job has quietly become “be the integration”, copying a name and a number out of one screen and into the next, twice a day, forever. At that point you’re not really paying for software any more. You’re paying for the subscriptions, and then you’re paying people to cover the gaps between them. The second cost never shows up on the invoice, which is exactly why it grows. We pulled that hidden labour apart properly in what admin really costs, and it’s almost always bigger than people guess.
How to tell whether you’re a buy or a build
Bespoke earns its keep when the thing you do is the thing the off-the-shelf tool treats as an afterthought. A handful of honest questions usually settle it, and they’re worth sitting with rather than skimming.
Start with whether the process is core to how you make money or just generic plumbing. Payroll is generic. So is email, so is card processing, so is your accounting ledger. Go and buy all of it and don’t think about it again. But the specific way you turn an enquiry into a booked, paid, certified, signed-off job might be your actual edge, the thing customers feel and competitors can’t quite copy. Edges are worth building around. Plumbing isn’t.
Then ask who’s bending to fit whom. If you’ve changed how you work to suit an app’s assumptions, or you’re paying for two hundred features so you can use four, you’re paying a tax every month and calling it a subscription. A bit of bending is fine and normal. A lot of it, on the process that earns your living, is a signal.
Next, count the real cost honestly, because this is where most people fool themselves. You’re not weighing a custom build against one £30 subscription. You’re weighing it against the whole stack, every tool involved in that one job, plus the human glue holding the stack together, plus the mistakes that slip through when a name gets retyped wrong at handover. Add all of that up and the comparison changes shape completely.
Last, ask whether it’ll still be true in three years. Custom pays off on the load-bearing processes that aren’t going to change next quarter, the ones that are basically the shape of your business. For anything experimental, seasonal, or fast-moving, stay flexible and rent. If your process is still settling, renting buys you the freedom to change your mind cheaply.
If the answers land on “core, bending to fit, several tools plus a person, and stable”, you’re in the territory where a small, sharp custom build stops being an indulgence and starts being cheaper than carrying on as you are.
What the maths can look like
Numbers make this less abstract, so here’s a worked example. Say a small firm runs that classic six-tool stack: six off-the-shelf apps at roughly £35 a month each. On top of the subscriptions, one person spends about five hours a week stitching the tools together, rekeying data, chasing the bits that fall between systems. Cost that staff time at a modest £25 an hour and it adds up fast.
Now compare it against a one-off bespoke build that does the joining-up itself: a higher number on day one, then very little to run.
| Six-tool stack | One bespoke build | |
|---|---|---|
| Up-front | £0 | £8,000 |
| Per year | £2,520 subscriptions + £6,500 staff time = ~£9,000 | £600 to run |
| Payback | n/a | under a year |
Illustrative figures.
Walk the arithmetic. Six tools at £35 a month is £210 a month, which is £2,520 a year. Five hours a week at £25 an hour is £125 a week, and across roughly 52 weeks that’s about £6,500 a year. Together the stack costs in the region of £9,000 a year, every year, and it never gets cheaper. The build costs £8,000 once and around £50 a month to run, call it £600 a year. So you’re comparing roughly £9,000 a year forever against £8,000 once plus £600 a year. The up-front cost looks frightening until you notice the running-cost gap is more than £8,000 a year. It clears the build inside the first twelve months and saves you most of nine grand every year after that.
The figures are made up, deliberately. Yours will be different, and the staff-time line is the one people lowball hardest, because it doesn’t arrive as a bill. It arrives as a person who’s a bit too busy.
”Custom” doesn’t mean rebuilding everything
This is the part that scares people off, and it shouldn’t. Going custom almost never means throwing out your accounting package, your card terminal, or the tools your team already knows. The high-return version is usually a small, focused piece of software that does the one thing the market doesn’t do for you, your particular job-to-certificate flow, your sector’s compliance clock, your specific way of chasing money, and then quietly connects the tools you’re keeping. You’re not replacing the stack. You’re removing the person who was holding it together by hand. Chasing late invoices is a good example of a narrow, high-value job that’s easy to automate without rebuilding anything, and we walked through that one on its own in automating the chase for late payments.
If you want to see the shapes these builds usually take, the kinds of jobs that earn their keep, they’re laid out on our solutions page.
How we think about it
When someone asks us to build something, the first thing we try to do is talk them out of it. If a cheap app solves the problem, we’ll tell you to go and buy the cheap app, and we’ll mean it. There’s no version of our business where selling you software you didn’t need works out well for either of us.
The builds worth doing are the ones that pay for themselves, and they tend to look the same: a load-bearing process eating staff time, money leaking through the cracks between tools, or a job that’s capped because the off-the-shelf option simply won’t bend the way your business does. If that’s you, the build isn’t a luxury. It’s the cheaper option you haven’t priced up yet.
Not sure which side of the line you’re on? That’s genuinely the most useful conversation to have. Tell us what you’re trying to fix and we’ll give you a straight answer, including “honestly, just buy the app” if that’s the right one.