What Does Custom Software Cost? Prices and Examples for SMEs

“What does our own software cost?” is the first question in almost every initial conversation, and the one most vendors dodge. Here is the answer, as honest as we can make it: price ranges by project size, two transparent worked examples and a five-year comparison against rented software. We build custom software from Rotkreuz in the Canton of Zug, for SMEs across Switzerland, and the figures in this article are the same ones we use internally.

The rough map: price ranges by project size

Custom software in Switzerland costs between CHF 10’000 and CHF 500’000. That is an honest answer, but not a useful one. It only becomes useful once you know the scope. Three sizes cover almost every enquiry that reaches us:

The ranges overlap on purpose. A small tool with a demanding integration can cost more than a lean business application without any. The size of a project depends less on how important it is to your company and more on how many different things the software has to do.

Important for planning: with us, a fixed price follows the first assessment. The ranges above frame that assessment; they are not an open-ended hour counter.

How long real projects took

Figures from our own work say more than survey averages. Four examples where we can state the duration:

• The calculation software Easy-FM, a standalone web application that calculates cleaning and maintenance effort in facility management, was built in 10 weeks.
• The ordering platform plaene-scannen.ch, with a live price calculator and complete order handling in four languages, was up after one week.
• The CAFM software easyCAFM, which covers everything from space capture through calculation to staff planning and the quote PDF, took 8 weeks of development.
• The Cliantra membership management was in operation after 4 weeks, because we tailored an existing open-source base to the association instead of building from scratch.

Duration is not the same as price, but it shows how differently sized “software” can be: from one week to ten, with correspondingly different effort.

Two worked examples, step by step

Price ranges are good, calculations are better. Both examples use our hourly rate of CHF 180 and list every assumption. The hours are realistic estimates for a cleanly scoped project, not a quote. Your project will look different, the method stays the same.

Example 1: quoting calculator for a trade business

Starting point: a business with five employees prepares quotes in Excel today. Everyone calculates slightly differently. The wish is a browser-based calculator that derives a price from measurements, materials and labour hours and produces a quote PDF. The price lists should be maintained in-house.

Calculation: 86 hours x CHF 180 = CHF 15’480. Optionally add hosting with us for CHF 300 per year. That makes CHF 15’780 in the first year, then CHF 300 per year for as long as you want the hosting. The example sits right in the middle of the range for a focused tool.

Example 2: order management for twelve users

Starting point: a service business with twelve employees manages orders in three Excel files and a calendar. The wish is a business application with user roles (dispatch, field staff, management), order entry, planning, status tracking, reports and a connection to the existing accounting system. The legacy data from Excel should be migrated.

Calculation: 296 hours x CHF 180 = CHF 53’280. With optional hosting for CHF 300 per year that is CHF 53’580 in the first year. The example sits in the middle of the range for a full-grown business application. If the business had no accounting connection and no legacy data, it would be 64 hours less, which is CHF 11’520 less. Items like these decide the price, not the number of users.

The four cost drivers

Four things determine almost the entire price. Knowing them lets you estimate before the first conversation which range your project will land in.

1. The number of distinct workflows. Every screen, every process, every special rule is work: design, build, test, document. Ten similar forms are cheaper than five completely different ones. Ask yourself: how many genuinely different things does the software have to do?

2. The connections to existing programs. Accounting, inventory, online shop, email dispatch, payment provider: every connection is a small project of its own. Well-documented interfaces are connected in a few days. Old systems without an interface can be the most expensive item in the whole project.

3. The user structure. An application for three internal people with the same rights is simple. Roles, approvals, customer access, multiple tenants or multiple languages increase the effort, because every view of the data has to be built and tested. The plain number of users, by contrast, is rarely a driver: whether 10 or 100 people work with the software changes almost nothing about the development effort. Rented software is different, there every user costs every month.

4. The volume of data and migration from legacy systems. Starting with an empty database is cheap. Ten years of grown Excel files, duplicate customer entries and three spellings of the same address have to be cleaned before they may enter the new system. This is often more work than clients expect, which is why we raise it early.

What hardly influences the price: the industry, the design (as long as it should be clean and functional) and whether the software runs in the browser or as an app on the phone. As long as it maps the same workflows, the effort stays similar.

Rent or own: the calculation over five years

Rented software (SaaS) looks cheap because the number is small: CHF 30 to 80 per user per month. The calculation only becomes interesting over the years. We compare three options for a business with 15 employees: a cheap rented solution, an expensive rented solution and the custom business application from example 2 (CHF 53’280 once plus CHF 300 hosting per year). All amounts are cumulative, meaning paid by the end of the respective year.

Assumptions: 15 users x CHF 30 x 12 months = CHF 5’400 per year, or 15 users x CHF 80 x 12 months = CHF 14’400 per year. No price increases, no additional users, no paid add-on modules. That is the friendliest possible assumption for renting.

What the table shows honestly: against the cheap rented solution, your own software does not win over five years if you only look at the price. Against the expensive rented solution it is cheaper from the fourth year on, and from then on by CHF 14’100 every year. After ten years the difference is more than CHF 85’000.

What the table does not show is usually more decisive: rent never stops, and in the end you own nothing. Prices rise, users are added, and rented software maps your workflow only as well as the vendor designed it for all of its customers. Your own software is paid once and belongs to you, with source code. Extensions only cost when you want them. The full comparison of both models, including the cases where renting is the better choice, is in the article Custom software vs. SaaS.

Keeping the first version small

The most effective lever against high costs is not a lower hourly rate but a smaller scope for the first version. Almost every enquiry that reaches us contains wishes for three years. That is good, the direction has to be clear. But what gets built first should only be what saves time from day one.

This is how we go about it:

• Find the one workflow that hurts the most. Usually it is the one where data is copied by hand from one program into another today. That workflow is the first version.
• Postpone reports. Reports are easy to wish for and laborious to build. As long as the data is captured cleanly, they can be added at any time.
• Integrations only once the core runs. If a file export is enough for the first few months, that saves days. The automatic connection comes in phase two.
• Leave out edge cases deliberately. The one customer with the special rule is handled by hand in the first version. That is cheaper than building the rule in right away.
• Settle open questions before the start. Changes during development cost the most, because finished work gets touched again. One hour of clarification often saves a day of rework.

easyCAFM shows that this works: after four weeks a first usable version with data capture and calculation was in use, with real buildings. Planning, work orders and defect management followed in the second four weeks. At plaene-scannen.ch, pricing logic and ordering flow were defined on day one, and the calculator ran internally on day three. Neither project was thought smaller, they were built in the right order.

For you this means: the first version of a business application does not have to cost CHF 53’000. If you start with the core and move integrations, data migration and reports to a second phase, you often land in the range of a focused tool and decide after a few months of real use what is actually missing.

When you do not need custom software

There are vendors who always recommend a new build, because they earn from it. We prefer to say up front when custom software is the wrong answer:

• Your workflow is common. Accounting, payroll, email, appointment booking, simple time tracking: mature off-the-shelf programs cover these cases well. Buy the ready-made program. Custom software pays off where your workflow differs from your competitors’, not where it is the same.
• Your workflow is not stable yet. If the process changes every few weeks because the company is still working it out, software is too early. Stabilise it in Excel or with a standard tool first, then have it built.
• The budget is clearly below CHF 10’000. Below that, no clean custom software can be built that includes handover, tests and documentation. A well-configured standard solution is the more honest recommendation in that case.
• A proven base fits 80 percent. Then adapting is smarter than building new. That is exactly what we did with the Cliantra membership management: freely available open-source software tailored to the association’s workflows, without a licence per user, deliberately cheaper for the client than a new build would have been.

The seven questions that help with this decision are in the article Off-the-shelf or custom software?. If you conclude after reading that a ready-made program is enough, you have saved a lot of money. If not, you now know why.

Your ballpark in two minutes

Everything in this article is ranges and examples. Your figure depends on your workflows, integrations, user structure and data. The fastest way to get there: our price calculator. A few questions about your starting point and scope, and within two minutes you receive an honest initial assessment with a fixed price, before you commit. After that we talk about what belongs in the first version and what can wait.

Frequently asked questions