Custom software or SaaS?
When an off-the-shelf product is enough, and when building your own starts to make commercial sense.
For most businesses, most of the time, the answer is SaaS. Subscribing to something that exists is faster and cheaper than building, and the product improves without you paying for it. If you are choosing an accounting package, a payroll system or an email tool, stop reading and subscribe.
The question gets interesting when the software is close to how your business actually makes money. That is where off-the-shelf starts to cost more than it appears to.
The real cost of SaaS
The monthly fee is the visible part. The less visible costs tend to be:
- Per-seat pricing. €40 a user a month is cheap at five users and €24,000 a year at fifty.
- Workarounds. The spreadsheet that sits beside the system because it cannot do the one thing you need. Someone maintains that.
- Multiple tools. One product for quotes, another for scheduling, a third for invoicing, and staff copying data between them.
- Changing your process to fit the software. Sometimes healthy. Often it means doing things a slower way because the tool insists.
Add these up honestly and a €300-a-month subscription can be costing €2,000 a month in time.
The real cost of custom
Custom software has an up-front cost and a smaller ongoing cost for hosting and maintenance. Traditional agencies price that up-front cost high because of the layers and hand-offs behind them. Working directly with Donworth Studio, it is significantly lower, which changes the comparison for many businesses. You also carry the risk: if the scope is wrong, that is your money.
In return, the system does exactly what your business does, it costs the same whether ten people or two hundred use it, and it belongs to you.
A simple way to decide
Ask four questions about the process you want software for.
Is this process how we compete?
If the way you quote, schedule, fulfil or serve customers is part of why customers choose you, software that flattens it to the industry standard is working against you. If the process is generic, buy.
Have we already tried the obvious products?
If you have trialled two or three SaaS options and each one needed a workaround for something central, that is evidence, not bad luck. If you have not looked, look first.
How many people will use it, for how long?
Per-seat pricing over five years is a useful comparison. Twenty users at €50 a month is €60,000 over five years, which can be more than building something that fits.
How much is spent moving data between tools?
Count the hours a week staff spend re-keying, exporting, reconciling or chasing information between systems. Multiply by a loaded hourly cost. Multiply by 48. That number is the annual budget you already have for fixing it.
| Signal | Points to SaaS | Points to custom |
|---|---|---|
| The process | Generic, same as competitors | Distinctive, part of how you win |
| Existing products | One fits well enough | Every option needs workarounds |
| Users | Few, or shrinking | Many, or growing |
| Data movement | Lives in one tool | Spread across tools, re-keyed by hand |
| Change | You can adopt the tool’s way | The tool’s way costs you customers or time |
Buy the software that runs your back office. Consider building the software that runs your business.
The middle path
It is rarely all or nothing. The most common shape we see working well is SaaS for the generic parts, accounting, payroll, email, and one custom system for the part that is specific to the business, connected to those tools through their APIs. You get the fit where it matters and the low cost where it does not.
The custom piece can also start small. A system that replaces the two worst spreadsheets and connects to your accounting package is a real, contained first project. It does not need to replace everything on day one.
What to do next
Write down the process that causes the most friction. List what you have tried. Count the hours lost each week. If the numbers point to building, that document is most of a brief already, and it is exactly the kind of starting point we work from on custom software projects.