Custom Software · Berlin

What does custom software cost? — an honest answer

Custom software cost varies enormously — from a few thousand euros for a focused tool to six figures for a complex enterprise platform. The range isn’t vague: it reflects genuine differences in scope, integrations, and ongoing maintenance. Here I’ll walk through what actually drives the price, and how to keep a first project small enough to be useful without betting the company on it.

Not sure what you need yet? A 30-minute call costs nothing and usually brings a lot of clarity.

Why quotes differ so much

Custom software cost: why two studios quote €8,000 and €80,000 for the same brief

It’s rarely dishonesty. Agencies interpret the same brief very differently: one reads “user dashboard” as ten screens with real-time data sync, authentication, and a mobile view; another reads it as a single static page. Both are technically correct, and both quotes reflect genuinely different amounts of work.

There’s also a structural reason: large agencies have account managers, project managers, QA testers, and design leads, all of whom are billed to your project whether or not the task needs all of them. A small studio or a solo developer has lower overhead, which often passes through to the price — though that trade-off comes with fewer redundancies if something goes wrong.

The most useful thing you can do before getting quotes is write down, very specifically, what the software has to do on day one — not what it might do eventually. That one step reduces ambiguity and makes every quote you receive more comparable.

What drives the price

The main cost factors — and how each one adds up

These are the levers that most reliably push a project up or down in cost. None of them are bad choices by themselves — they’re just decisions worth making consciously.

  • Scope & number of features — every additional screen, user role, or workflow branch adds design, development, and testing time. The most common cost driver is a scope that grows during the project.
  • Third-party integrations — connecting to payment providers, CRMs, ERPs, accounting systems, or external APIs adds significant complexity. Each integration needs authentication, error handling, and often ongoing maintenance when the external API changes.
  • Data volume & structure — a simple form that saves to a spreadsheet is very different from a system that processes thousands of records, needs to search or filter quickly, or has to handle sensitive data compliantly (GDPR, financial regulations, health data).
  • User authentication & permissions — adding login, roles (admin vs standard vs read-only), and multi-tenancy (separate data per client) is a substantial piece of work that’s easy to underestimate.
  • Maintenance & hosting — the build cost is only part of the picture. Servers, security updates, bug fixes, and adapting to new browser or OS versions are ongoing. Budgeting zero for maintenance after launch is the most reliable way to accumulate technical debt.
  • AI or automation features — adding AI capabilities (classification, generation, smart search, workflow automation) can be done affordably using existing APIs, but design, prompt engineering, and quality assurance still take real time. More in my AI custom software guide.
Small build vs enterprise

Realistic ranges: small studio, mid-market agency, enterprise

These are rough guides, not quotes. Every project is different, and a simple tool can sometimes cost more than a complex one if the requirements are poorly defined.

Small build / focused tool (€1,000 – €15,000): a single workflow automated, a custom form with backend, a simple internal dashboard, or a lightweight API integration. This is where I typically work with small teams and businesses. The scope is narrow by design — one real problem, solved well.

Mid-market product (€15,000 – €80,000): multi-user SaaS tools, customer portals, internal platforms with multiple integrations, or B2B web applications. Usually involves a small team and several months of work. Quotes in this range from a reputable studio are generally fair reflection of genuine complexity.

Enterprise / large agency (€80,000+): large organisations with existing infrastructure, compliance requirements, legacy system integrations, and internal procurement processes. The premium often reflects overhead and risk management as much as the code itself.

One useful mental shift: instead of asking “how much does custom software cost?” ask “what is the smallest version of this that would make a real difference?” That reframe almost always produces a better first project.

Keeping the first project small

How to scope a first custom software project sensibly

Most custom software projects that go over budget do so because of scope creep — features that seemed minor during the brief, but each added hours to the build. Starting small is not a compromise; it’s a way of learning what you actually need before committing to the full version.

A few things that help:

  • List what’s in scope for day one — then cut it in half. The first version should do one thing well. Everything else can be version two.
  • Accept rough edges in the UI. Good logic in an average-looking interface is far more valuable than a beautiful screen that doesn’t work reliably. Polish is cheap to add later; bad architecture is expensive to undo.
  • Delay integrations where possible. If a manual step works for now, keep it manual. Automate once you’ve confirmed the process is stable and worth the cost.
  • Budget for maintenance from the start. Even a simple tool needs occasional updates. A rough rule: budget 15–20% of the build cost per year for ongoing upkeep, more if the system is connected to third-party services that change frequently.

If you’re interested in how I approach custom builds, the custom software Berlin page covers how I typically scope and price a project. You’re also welcome to get in touch for a free conversation before committing to anything.

Common questions

Custom software cost — FAQ

Is custom software always more expensive than an off-the-shelf tool?+
Not always, but often yes in the short term. An off-the-shelf SaaS tool can cost €50–500 per month and start immediately. Custom software has a higher upfront cost and takes time to build. The trade-off is ownership: you pay once and the software does exactly what you need, with no per-seat fees, no dependency on a third-party’s roadmap, and no data leaving your infrastructure. For most small businesses, a SaaS tool is the right call. Custom software makes sense when no tool does quite what you need, or when you’re dealing with volume, compliance, or integration requirements that off-the-shelf products can’t handle well.
How do I get an accurate quote?+
The more specific you can be about what the software needs to do — step by step, for each type of user — the more accurate any quote will be. Vague briefs produce vague quotes with wide contingency margins. A short user story format works well: “As an admin, I want to upload a CSV, and the system should parse it, validate each row, and flag errors for review.” That gives a developer something concrete to estimate. If you’re not sure how to write this, I can help during a free scoping call.
What’s the difference between a freelancer, a small studio, and a large agency?+
A freelancer is typically one person handling most or all of the work. Lower overhead, often lower price, but limited bandwidth and a single point of failure. A small studio (like WilkiLeads) combines a lead developer or consultant with a trusted network of specialists, giving you broader coverage without large-agency overhead. A large agency has teams, project managers, and account handlers — useful for very large, complex projects but rarely necessary for a focused tool or a first custom build. None of these is the right answer in every case. It depends on the size and complexity of your project, your timeline, and how much risk you can absorb.
Can AI features be added affordably?+
Often yes — adding AI capabilities via existing APIs (OpenAI, Anthropic, Google, etc.) is much more accessible than building a model from scratch. The practical cost is in the design work: deciding exactly what the AI should do, how to handle errors and edge cases, and how to present the output in a way users trust. That work still takes time. For a simple use case — summarising text, classifying records, drafting structured responses — AI can be added to a custom tool for a few thousand euros. More ambitious features (real-time personalisation, fine-tuned models, large-scale embeddings) scale up significantly. I cover this more in the AI custom software guide.

Want a realistic picture of what your project might cost?

I’m happy to have a free 30-minute call, hear what you’re trying to build, and give you an honest estimate — or tell you plainly if an off-the-shelf tool would serve you better.

Get in touch →

No commitment, no pitch — just a straight conversation.