Custom software development in the Netherlands typically costs €75–€150 per hour, or roughly €10,000–€30,000 for a small tool, €30,000–€80,000 for a mid-sized application, and €80,000–€200,000+ for a larger platform. The exact figure depends on scope, how many systems it has to integrate with, and how much is genuinely custom versus assembled from things that already exist. Almost all of the cost is people-time, so the real question is not the hourly rate, it is how many hours the work honestly needs.
The short answer: typical price ranges
As a rule of thumb in the Dutch market:
A proof of concept or MVP, one workflow, working on real data: €15,000–€40,000.
A small internal tool or integration: €10,000–€30,000.
A mid-sized application (portal, SaaS module, custom workflow): €30,000–€80,000.
A larger platform or multi-team product: €80,000–€200,000 and up.
Freelance developers bill roughly €75–€110/hour; specialist agencies €100–€150+; the large consultancies more. Those ranges are a starting point, not a quote. The same brief can differ 3× depending on the decisions below.
Custom software costs per phase
With a fixed price per phase you know what each step costs before it starts. At Crux Digits there are three phases, each with a deliverable you keep:
- Audit and strategy, €2,500. In one to two weeks: the process, systems and data mapped, with a scope and a projected return.
- Production-ready MVP, €20,000. In four to six weeks: working software on your own data and in your own systems, tested against the current way of working.
- Production, from €50,000. From eight to twelve weeks: integrated, secured, monitored and handed over with documentation and all source code.
Prices exclude VAT; the audit is credited if you continue to the MVP within 60 days. How the work runs: custom software development in the Netherlands at a fixed price.
What actually drives the cost
Five things move the number far more than the hourly rate:
Scope and complexity. A single-purpose tool is inexpensive; a product with roles, permissions, edge cases and a real UI is not. Vague scope is the most expensive thing you can buy.
Integrations. Connecting to your ERP, CRM and other APIs (AFAS, Exact, Salesforce and the like) is often half the work. Every integration is a small project of its own.
Data and AI. Machine-learning and LLM features add data preparation, evaluation and guardrails. They pay back, but they are not free. See what an AI proof of concept costs.
Design, security and compliance. A usable interface, proper authentication, GDPR and, where relevant, the EU AI Act all take real hours that a low-cost quote quietly leaves out.
Team seniority. A senior team costs more per hour and far less per outcome, because they build the right thing once instead of rebuilding it three times.
Hourly, day-rate or fixed price, and why it matters
Open-ended hourly and day-rate work shifts all the budget risk onto you: if the estimate slips, you pay for it. A fixed price per stage moves that risk to the builder and forces a clear scope up front. It is the single biggest lever on what a project actually ends up costing. See how we structure fixed-price software and AI projects.
Where AI changes the cost equation
The least expensive custom software is the software you do not build. A good partner buys the boring 80%, such as auth, hosting, standard components and off-the-shelf models, and only builds the 20% that is genuinely yours. That is the difference between a €200,000 project and an €80,000 one for the same outcome. It is also why we work AI-first: AI, ML and LLM building blocks now cover work that used to be bespoke. For the build side of that, see our role as an AI development agency in the Netherlands.
What custom software costs after go-live
Every quote you collect prices the build. None of them prices the years after it, and that is where the lowest quote turns into the expensive one. Four lines make up the running cost, and only the first is predictable from the quote.
- Hosting and infrastructure. Modest for most business applications, and genuinely small next to the build. It only becomes a real number when you are processing documents or running models at volume.
- Maintenance and dependencies. Frameworks, libraries and operating systems move whether you touch the software or not. A system nobody updates for two years is not stable, it is accumulating a rebuild.
- Change. Your process changes, a supplier changes a format, the law changes. Budget for change from the start instead of treating each one as an incident.
- Model and API costs, when AI is in it. These scale with use rather than sitting flat, so they need a ceiling and an alert, not an estimate in a spreadsheet.
Ask every supplier the same two questions before you compare prices at all. What does it cost per month to keep this running and current, and what happens if we stop paying you? If the answer to the second is that the system degrades, the build price was never the price. Our own managed support for systems we built is a published EUR 500 per month, precisely so that the number is on the table before the build starts rather than after.
How to keep the cost down
Three habits save the most money. One: start with a small paid audit or proof of concept before committing to a full build. A few thousand euros spent scoping prevents tens of thousands spent building the wrong thing. Two: own the code and IP, so you are never locked into one vendor. Three: keep one senior team from strategy to launch, so nothing is lost in a handover. If you want a partner who works that way, see custom software development at Crux Digits.
Software built around your process and data, with AI where it changes the outcome. Something working every week, and the code is yours.
Custom software development in the Netherlands →
