Buy or Build: When Off-the-Shelf SaaS Stops Being Cheaper Than Custom Software

The question is almost always framed wrongly. It is not which one is cheaper, it is which cost grows with what. The cost of an off-the-shelf tool grows with headcount. The cost of a custom system grows with change. If you have not drawn those two curves, you are deciding on the basis of one figure from a quote, and that figure is the smallest part of the cost.

Three costs that are not in the licence price

Integrations. Every additional system means one more data flow that somebody maintains. By the time there are five, maintaining the connections costs more than the licences do.

The process bending around the tool. The most expensive item, and the only one nobody invoices. It is paid in working hours, every day, and it is never booked as a software cost.

What goes unused. According to Zylo's 2026 SaaS Management Index, organisations use roughly 54% of the licences they pay for; the remaining 46% is waste. For companies under 500 employees the average annual spend on unused licences is estimated at 3.8 million dollars. Those are US market figures, but the ratio of roughly half holds broadly.

Five questions that produce an answer rather than an opinion

1. Is this process a source of your advantage, or does it merely have to exist?

Payroll has to exist and it works the same way in every company. Buy it. The way you price a made-to-order product, or schedule crews, or run a quote through a configurator, is probably part of why customers choose you. That is a candidate for building.

2. How many people use it, and how does that number grow?

Licence times headcount times years. If the number of users grows with revenue while the price per user does not fall, the two curves cross. The only question is when.

3. How many systems have to talk to each other?

Up to two, buy. More than four, and an integration layer is already a project. The only question left is whether you build it deliberately or by accident.

4. How many times in the past year have you said "the tool cannot do that, we do it manually"?

Count those processes and estimate the hours. It is the one number in this whole calculation you already have. You just have not added it up.

5. What happens if the vendor doubles the price, drops a feature, or gets acquired?

This is not paranoia. It is a question about where your data lives and what it costs to leave.

What the calculation looks like when it is done honestly

Take a company with 40 users and a tool at 30 euros per user per month. That is 14,400 euros a year, or about 43,000 euros over three years.

Now add what usually gets left out. Say the tool does not cover one process, so two days a month are done by hand. Two days times twelve months times three years is 72 working days. At 150 euros a day, that is another 10,800 euros. Add the maintenance of two integrations and you are close to 60,000 euros over three years.

A custom system covering the same scope typically sits somewhere between 35,000 and 60,000 euros one-off, plus 15 to 20 percent a year for maintenance and development.

The point is not that custom is always cheaper. The point is that the crossover moves earlier every time you add a user, an integration, or a process the tool does not cover, and later every time your process turns out to be standard after all.

A hybrid is usually the right answer

The best outcome rarely looks like a choice between two options. You buy the core - accounting, payroll, mail, a standard CRM - and you build the layer that describes you specifically, plus the integration between them.

Custom does not mean "the whole system". It means the part no product on the market understands, because nobody else does it that way.

Three signs it is time to build

• There is a master spreadsheet the company cannot function without, and one person maintains it. When that person goes on holiday, the business slows down.
• The same data is entered into two systems. That is not an inconvenience, it is where the errors come from that somebody later spends days hunting.
• You bought a tool and then hired a person to do what the tool cannot. That is the most expensive combination there is.

Three signs it is not

• The process is not stable yet. If it changes every month, you will automate chaos and then pay to change it.
• There is a global standard for this job and everyone does it the same way. Bookkeeping and payroll are solved problems.
• There is nobody inside to own the product. A custom system without someone deciding what goes into it turns into an expensive collection of wishes.

One test before you decide

Describe the process in five sentences without naming a single tool. Then ask whether those same five sentences would apply to any other company in your industry.

If they would, buy something and adapt. If they would not, and the reason they would not is exactly what you put in front of customers, then that process is your product rather than somebody else's.

If you want to see the calculation for your own case

APPARO builds custom software, but the first conversation is always about what should not be built. We go through the tools you already pay for, the processes still done by hand, and the integrations you already maintain, then give you the calculation with all three items in it rather than just the licence.

Often the conclusion is that you need one module, not a system. That is a perfectly good outcome for us too.

Write to office@apparo.rs or book an intro call at apparo.rs/en/start-project.