“What does it cost us?” is the first question as soon as the idea for our own software emerges in the company – and the answer “deals on it” is correct, but useless. This article shows what the price of an individual software actually consists of, what sizes you should expect, what costs are incurred after the go-live and how you assess an offer before you sign it. What we do in such projects is on our side individual software development.
The short answer
The price of individual software almost always results from a simple calculation: Expenditure in person days × daily rate. Everything else – fixed price, flat rate, “package price” – is just another packaging of the same invoice, where the provider additionally prices the risk of a miscalculation.
The daily rates of experienced development service providers in Germany are roughly between 700 and 1,400 euros. The decisive lever is therefore not the daily rate, but the effort – and this depends almost completely on what the software must be able to do.
Three orders of magnitude from practice
The following classifications are reference values for effort, not a price list. They help to roughly classify an idea before the first conversation takes place.
1. A tool that replaces an Excel solution
A web application for a clearly defined task: collecting, checking, evaluating, exporting data. Few roles, a database, no or a simple interface. The expense here is typically in the range of a few weeks. This is the most common entry point – and often the one with the fastest benefit, because the Excel file that only one person understands is a real operational risk.
2. A portal or specialist application with interfaces
Multiple user groups with different rights, connection to ERP, CRM or accounting, workflows with releases, notifications, reports. Such projects are usually over several months they are implemented in stages. The wide range within this group almost always comes from the interfaces: A well-documented REST API is connected in one day, an old system without documentation can cost weeks.
3. A business-critical system
Software without which a core process stands still: high availability, many integrations, migration of large data stocks, audit obligations or regulatory requirements. Such projects shall cover: over many months they are planned in stages. No one should commission such a system in a single block.
What really drives the price
Two projects with the same heading can differ in effort by a factor of three. Five points explain most of it.
Interfaces. Each connection to another system is a small project of its own. The decisive factor is whether there is a documented interface, who supervises it and whether there is a test environment. If all this is missing, the effort increases – and above all the uncertainty.
Data transfer. If the new software is to start with old data, it must be sighted, cleaned and transferred. This is regularly underestimated. How to plan such a takeover without data loss, we describe in the contribution to lossless data migration.
Roles and rights. An application in which everyone is allowed everything is quickly built. Every additional role, every release level and every rule, who is allowed to see which data, costs effort – in the implementation and in the test.
Unclear requirements. The most expensive sentence in a software project is “We’ll clarify that later”. What remains open at the beginning is later decided under time pressure and often double built.
Operation and security. Where does the software run – in the cloud, in your own data center, at the service provider? Who plays updates, who monitors them? These questions belong in the planning, not in the week before the go-live.
The cost after the go-live
Software is not ready after acceptance. Operating systems, frameworks and libraries get security updates, interfaces of the connected systems change, and users have better ideas after three months of operation than before the start of the project.
As a rule of thumb, for maintenance and minor advancement annually about 15 to 20 percent of the original development costs plan the operating costs for server or cloud. Those who do not provide this budget have an application after two or three years that no one wants to touch anymore – and then face the next new development.
Fixed price or effort?
A fixed price seems safer, but it is only if the scope is really fixed. For a clearly described, manageable project, it makes sense. With everything else, the provider praises the uncertainty – or he calculates narrowly and negotiates every deviation as an amendment.
In practice, a mixed form proves itself: the first stage – such as a prototype or a first runable version – at a fixed price, then further development after effort in short sections with a fixed budget. So you see real results early on and keep control of every additional issue.
What changes AI in costs – and what does not
AI tools have significantly accelerated software development, especially in routine work: standard code, tests, documentation, conversions to existing code. This will benefit mainly small and medium-sized projects, for which this proportion is high.
What AI does not replace are the expensive parts: to clarify what the software should actually be able to do, to define the architecture, to understand interfaces and to technically test the result. Those who promise you a drastic cost reduction through AI alone usually only postpone the effort – into troubleshooting after the go-live.
What do you know about a reliable offer
- Effort is broken down into: – by functions or work packages, not as a single number.
- Assumptions are there. What was assumed, such as interfaces or data quality? Where assumptions are missing, there is also no basis for further demands – or for their defence.
- Test and acceptance are plannedNot just development.
- Operation and maintenance are describedAt least as an option with price.
- The source code belongs to you. Without this commitment, you are permanently bound to a provider.
Is individual software worth it?
Not always. If there is an established standard software for your task that fits 80 percent, it is usually cheaper – even if you have to adjust your processes a bit. Individual software is worthwhile where your process sets you apart from the competition, where standard solutions only work with many detours or where several systems have to work together neatly.
A good service provider also tells you when standard software is the better choice. This is not a sign of a lack of interest, but the opposite.
Conclusion
What individual software costs can be said earlier and more accurately than many think – provided you talk about the scope instead of the technology. The magnitude is usually determined after a single conversation, a reliable number after a short clarification of requirements or a prototype.
If you want to make use of your plan, use our free Software Project Check: In 45 minutes, we assess the scope, procedure and timeframe – and honestly tell you whether a solution is worthwhile. More about our work in the region can be found on the page Software development in Stuttgart.
Would you like to implement this in your company? We support you pragmatically – from the idea to the operation.