Decision Guide · Custom Software
One of the first questions every growing business faces: buy an off-the-shelf product, pay for a SaaS subscription, or build something custom? Here is how to decide, what it really costs each way, and where a mix of the two is the honest answer.
Last reviewed: October 2026
The decision in one viewMost businesses sit somewhere in the middle, and that is normal. The useful question is not "build or buy?" in the abstract, but "for this process, which one?" Answer it process by process and the picture usually becomes clear.
We talk plenty of people out of building bespoke software. If an off-the-shelf product already does the job, buy it — you'll be live next week and it'll cost a fraction of a build. The time to build is when the tool is fighting your business instead of helping it, and you can see the workarounds quietly turning into someone's full-time job.
Off-the-shelf and SaaS almost always win on day one. The comparison that matters is total cost of ownership over the years you will actually run the software — and who owns it at the end.
For most of the last twenty years, the honest answer to "should a small business build its own system?" was no. Custom software meant a five or six-figure project, a year of meetings, and a fair chance it would be late. That has changed quickly. But "build everything" is not the new answer either.
AI coding tools now do a lot of the typing. Work that took a developer months can take weeks, sometimes days. A custom CRM or job system that used to be out of reach for a small business can now be built by a small team using AI properly.
What did not change: someone still has to decide what to build, check that the code is secure, host it somewhere sensible, back it up, patch it, and change it when the business changes. AI is very good at writing code. It is terrible at knowing whether it should, and it does not answer the phone when the system goes down on a Friday afternoon.
So the cost of building has fallen a long way. The cost of owning has not fallen nearly as much. That is the whole decision in a sentence.
There is a third option people forget. For a lot of the software small businesses rent (file sharing, password managers, help desks, CRMs, e-signing) there is a free, open-source alternative. No per-user fees, nobody can hold your data hostage, and because the code is open it can be changed to fit. The catch is the same as custom software: someone has to host, update and back it up. We have listed the good ones in open-source alternatives to the software small businesses rent.
| If this is true | Lean towards |
|---|---|
| Every business does this job the same way | Off the shelf |
| It's regulated or changes with the law (payroll, tax, card payments) | Off the shelf |
| A product does nearly everything and you can live with the gaps | Off the shelf |
| You're paying per-user fees for something standard, for 10 or more people | Open source |
| You want to own your data and stop the price rises | Open source |
| The job is how your business is different | Custom |
| Staff copy data between systems by hand | Custom integration |
| A spreadsheet runs a core part of the business | Custom |
| Nobody will look after it once it's built | Don't build it |
That last row matters most.
Custom software is not a purchase, it is a pet. It needs hosting. It needs security updates, because the libraries it is built on get patched every few weeks. It needs backups that are tested. It needs changing when your business changes, and fixing when something it connects to changes without asking.
The classic way this goes wrong: a freelancer or agency builds something good, gets paid and moves on. Two years later it is running on an out-of-date server, nobody has the passwords, and the person who understood it cannot be reached. If that is happening to you right now, see our developer has left, what now?
The newer way: someone in the business builds an app with an AI tool over a weekend. It works, which is impressive. It also has the database password in the code and lets any logged-in user see everyone else's data. Before anything AI-built touches customer data, have someone experienced check it before it goes live.
So before you build, budget for owning it every year, not just for the build, and decide who will host, patch and change it. If you cannot name that person or company, do not build it yet. If you would rather it was all one team, that is what Your tech department is for.
If it is a differentiator — the thing you do better than anyone else — owning it as bespoke software can be worth it. If it is a commodity like payroll, email or accounting, buy a product. Nobody wins by building their own helpdesk.
An off-the-shelf tool you bend out of shape with add-ons, integrations and workarounds can end up costing more than a clean bespoke build — and you are still renting it. Heavy customisation is often a signal that the fit was never there.
SaaS pricing is per-seat or per-tier. Model the cost at the size you are aiming for, not the size you are now. A bargain at ten users can be punishing at fifty, and the switching cost only grows the longer you wait.
With SaaS you are a tenant. Check what happens to your data if you leave and how hard export really is. With bespoke, the data and the code are yours. If owning your data matters to the business, that alone can decide it.
If people export CSVs every morning to move data between systems, you are already paying for software that does not quite fit. A bespoke layer that integrates the tools you keep is often the cheapest fix of all — no full rebuild required.
The temporary spreadsheet that became a process. The manual rekeying. The daily reconciliation between systems. That time is a real, recurring cost. Put a number on it, then weigh it against the one-off cost of building the thing properly.
The build-or-buy question almost never has a single answer for a whole business. The right approach for most companies is a deliberate mix, decided process by process.
This is how we usually advise clients: buy first, build deliberately, integrate everything. The goal is not bespoke software for its own sake — it is the lowest-friction way to run your specific business. If that means we talk you out of a build, we will.
Not upfront. Off-the-shelf and SaaS tools are almost always cheaper to start with because the build cost is shared across thousands of customers. Bespoke software costs more to build. The comparison changes over time: SaaS is a per-seat or per-tier subscription that grows as you grow and never stops, while bespoke is a larger one-off build you then own, with only hosting and maintenance to run. For a standard process used by a few people, off-the-shelf usually wins on total cost. For a core process at scale, or one no product fits well, bespoke often becomes cheaper over a few years, and you own the result.
When the process is common and well understood, when an existing product already does most of what you need, when you need it running quickly, and when the subscription cost is comfortable at the scale you expect to reach. There is no point building a bespoke email platform, accounting package or helpdesk when mature products already exist and fit. Buy first. Build only where buying leaves a real gap.
When the software is a competitive differentiator rather than a commodity, when no product fits your actual process without heavy workarounds, when per-seat subscription costs have grown painful at your scale, when you need systems that do not talk to each other to work as one, or when owning your data and your roadmap matters. If your team spends more time working around a tool than working in it, that is usually the signal.
Yes, and most of the time that is the right answer. You rarely need to rebuild everything. The common pattern is to keep the off-the-shelf tools that work, such as accounting, email or a good SaaS product you already rely on, and build a bespoke layer that fills the gaps and integrates them so data flows automatically. A lot of our work is exactly this: bespoke software sitting alongside and connecting systems a business already owns.
With Dev Partners, yes. The code we write for you is yours. There are no licence fees, no vendor lock-in, and no requirement to keep paying us to use your own system. That is one of the structural differences from SaaS: a subscription is access you rent for as long as you keep paying, while bespoke software is an asset you own outright.
Yes, a lot cheaper to build. AI coding tools now do much of the typing, so work that took months can take weeks. What has not got much cheaper is owning it: hosting, security updates, backups and changes still need an experienced person every year. Budget for owning it, not just building it.
For a prototype or an internal tool with no sensitive data, go for it. Before it handles customer data or goes on the internet, have someone experienced check logins, permissions, secrets, backups and hosting, because AI tools routinely miss those.
We have no incentive to push you toward a build you do not need — we would rather give you the honest answer and earn the work where it genuinely pays off. If an off-the-shelf product is the right call, we will say so.
Or call directly on 01474 639 089. We reply within one working day.
Describe the process or the tool you have outgrown and we will tell you where it sits on the build-or-buy line.