We build bespoke software for a living, so it is worth saying this first: most of the time, you should buy the off-the-shelf product. We tell people that on calls, including people who arrived ready to commission a build.
Off-the-shelf software is cheaper to start, live in days rather than months, and someone else worries about keeping it running. For accounts, email, payroll, a CRM for a normal sales process, there is rarely a good reason to build your own.
The question is how to tell when you are in the minority of cases where that stops being true. Here is how we work it out.
What each one actually costs
Off-the-shelf costs a licence, usually per user per month, plus the time spent setting it up and bending your process to fit it. The licence is visible. The bending usually isn't, and that is where the real cost hides.
Bespoke costs a build, then a running cost. At 628 a focused tool or a set of integrations starts around £5,000, and a system the business runs on starts around £30,000. On top of that, budget roughly 15 to 20% of the build a year for hosting, updates and the occasional integration breaking because someone else changed their API. Anyone who tells you bespoke software is a one-off payment is hoping you won't ask.
The 90% test
Most decisions come down to one question: what does the product not do, and what does that gap cost?
If a product does 90% of what you need and the missing 10% is cosmetic, buy the product. A screen laid out differently than you would like, a report you have to filter yourself, a field with the wrong name. Live with it. That gap is not worth £5,000, let alone £30,000.
If the missing 10% is the part that makes you money, it is a different conversation. The pricing rules that win you trade customers. The way your jobs move through the factory. The term structure your club actually runs on. When the product forces you to work around the thing that sets you apart, you are paying a licence to be made more ordinary.
And if the missing part is small but costs someone a day a week, forever, do the arithmetic. A day a week is around fifty days a year. That adds up quickly, and it never stops.
The licence arithmetic
Per-seat pricing is the other thing that tips the balance, because it grows with headcount whether or not the software is doing anything more for you.
Roughly: a £20,000 build with £2,500 a year of upkeep beats a £900 a month product inside three years, and unlike the product it doesn't get more expensive every time you hire. Below about £300 a month, keep paying the licence. The build will not pay for itself in any reasonable time, and you lose the updates someone else was doing for free.
Between those two numbers is a judgement call, and it mostly depends on how settled your process is.
When to stay off the shelf
Even when a product fits badly, there are times bespoke is the wrong move.
Your process is still changing. If the way you sell or deliver shifts every few months because the business is still working out what it is, wait. Software is expensive to build and more expensive to keep rewriting. A year more on a spreadsheet is often the cheaper choice.
It's a solved problem. Accounts, payroll, email, file storage. Thousands of businesses have the same needs and someone has built a good product for them. Your accounts are not special, and that is a compliment.
Nobody owns the decision. Bespoke software needs someone in the business who can say what it should do and sign off what it does. If that person doesn't exist, a product with sensible defaults is safer than a system built from guesses.
When bespoke wins
The workaround is permanent. People export, retype, reconcile and re-check every week, and there is no product that removes the step.
The process is your advantage. Pennine Gymnastics is a good example. Gym software assumes monthly memberships. A gymnastics club runs on terms and blocks, rec through to squad on the same timetable, registers taken on the floor and waiting lists that matter. Built around how the club actually runs, class bookings rose 14%.
The licences have outgrown the value. See the arithmetic above.
One person is the system. If a process only works because one individual understands it, you are carrying a risk the business probably hasn't priced. Software won't stop them leaving, but it stops their leaving from being a crisis.
The option most people miss: both
This is where most of our work actually sits, and it is cheaper than either extreme.
Keep the off-the-shelf products that work: Xero, Sage, your CRM, your ERP, the courier. Then build only the bit that is missing, and connect it to them. Most business software offers a connection point for other systems, called an API, so a small bespoke tool can read from and write to what you already have.
That might be:
- An integration that moves orders from the website into stock and accounts, so nobody retypes them.
- A quoting tool that knows your pricing rules and pushes the accepted quote into the system you already use for invoicing.
- A customer portal that sits on top of your existing data and answers the "where's my order" calls.
- A dashboard that pulls figures from three products into one screen.
None of those replace anything. They remove the gap between products, which is where the hours usually go. Integrations are often the cheapest bespoke project there is and the one with the fastest payback, which is why we tend to suggest starting there. Replacing everything at once is the most expensive and the riskiest version of this work. Connecting what you have and building only the missing part is cheaper and reversible.
There are more examples of what businesses actually get built in a separate post.
A quick way to decide
Answer these honestly:
- What does the product not do? Write it down specifically.
- What does that cost a year? Hours times rate, plus errors, plus anything lost.
- Is the process settled? Will it look roughly the same in two years?
- What is the licence heading towards? At today's headcount, and at the one you are planning for.
- Can the gap be filled without replacing the product?
If the answer to the second question doesn't clear a build inside about two years, buy the product. If the fifth answer is yes, build the small thing rather than the big one. Only when the gap is expensive, the process is settled and nothing can be bolted on does a full bespoke system make sense.
Getting a second opinion
We do this arithmetic with people before quoting, and a fair number of those conversations end with us recommending a product and not building anything. What custom software costs has the detail on pricing, and our custom software development page explains how we scope a build.
If you're local, we have studios in Leicester and Huddersfield, and the first step is usually us coming to watch the work happen. It's the quickest way to find out whether you need software at all.
