Freelancer or Agency for App Development?

Q: What does the developer owe, a contract for work or a service contract?

Under German law a contract for work owes a result, that is working software, and gives you statutory warranty rights plus an acceptance step to which payment can be tied. A service contract owes only the activity, no result, and does not include these defect rights. For an end customer ordering a finished app, the fixed-price contract for work is the fitting and safer model.

Q: Can a single freelancer handle iOS, Android and backend alone?

An experienced senior specialist often covers the whole chain for clearly scoped projects, from the app on both platforms to a lean backend, especially with cross-platform tech from one codebase. The limit is very large, cross-departmental scope with many parallel workstreams. For the typical process app of an SME, the scope usually sits below that threshold.

Q: When is an agency or an in-house team the better choice?

An agency pays off for many parallel disciplines, large scope, permanent operation with guaranteed cover and fixed SLAs, and for corporate sign-off processes. An in-house team makes sense once app demand is permanent and you develop and maintain continuously, because then the fixed personnel cost carries. For a one-off project, by contrast, in-house is the most expensive route.

Freelancer or agency comes down to one question: who carries responsibility after launch? For clearly scoped projects and direct communication, an experienced solo developer is often faster. When many trades come together, the app has to run for years, or you lack in-house tech know-how, a team is the better choice. What decides it is not the legal form but reliability, code ownership and cover for absence.

By Tobias BoehmUpdated June 4, 2026

Freelancer, agency or in-house: the short answer

There is no single right answer, there is the one that fits your project. For a clearly scoped app project with a direct line to the person doing the work, an experienced solo developer is often the fastest and cheapest choice. As soon as many disciplines run in parallel, operation needs guaranteed cover, or corporate sign-off processes apply, an agency plays to its strengths. Your own in-house team only pays off once app demand is permanent, not for a one-off project.

The crucial point most comparisons leave out: the question "freelancer or agency?" mixes two different deals. In one you buy capacity by the hour and carry the risk yourself, in the other you buy a finished result that the provider is liable for. Anyone buying a result at a fixed price asks the risk question differently from the way most ranking articles answer it.

Freelancer, agency and in-house compared directly

The table sets the three routes against the criteria a technical decision-maker actually weighs. All figures are sourced market values, not fixed prices and not quotes. Hourly rates vary strongly by experience and the technology used, which is why ranges appear rather than an invented average. The superscript numbers refer to the source list at the foot of the page.

Freelancer/solo, agency and in-house team compared (market values, German-speaking market, 2026)
Criterion	Freelancer / solo	Agency	In-house team
Cost / day rate	Development freelancer around €94/h¹, app range €75 to €120/h⁵	€90 to €150/h, senior up to 180⁴; the markup bundles PM, QA and design	Senior around €70,000 to €90,000+ gross/year⁴, fully costed around €136,000/year⁶, fixed
Communication path	directly to the builder, short paths	project manager as fixed contact, more layers of coordination	direct and internal, in your own house
Breadth of skills	one person, often a senior specialist; serial capacity	bundled multidisciplinary team: design, backend, QA, DevOps	depending on team size, built up internally
Absence risk / bus factor	bus factor of 1, honestly; defusable through documentation and a transferable state	backup and bench when someone is out, scales by adding team members	fragile in a small team, knowledge lost when someone leaves
Scaling for large scope	limited; ideal for clearly scoped work	parallel workstreams, expandable within days	slow, dependent on recruiting and time-to-hire
Code ownership / lock-in	to be set out by contract, then on par with an agency	usually full transfer of rights plus NDA	internal, full sovereignty over code and knowledge
Maintenance / long term	after handover via a follow-up contract or a managed handover	structured support, SLA possible	permanent, the strongest tie to the company

What matters is the contract model, not the legal form: a fixed-price contract for work owes a result, a service contract owes only activity.

What really matters: buy a result, not hours

Behind the "a freelancer is risky" objection lie two different buying models that are rarely cleanly separated. In the first you buy capacity: a developer by the hour, often on a service-contract basis. Then you carry the risk, the absence cover and the steering yourself. In the second you buy a result: a finished, working app for which the provider owes success, not attendance. That is the distinction the whole decision hangs on.

Underneath it sits the axis of contract for work versus service contract. Under German law, a contract for work (Werkvertrag) owes a result, that is working software, and gives the client statutory warranty rights: subsequent performance, damages, defect remedy. On top comes acceptance, to which payment can be tied². A service contract (Dienstvertrag) owes only the activity, no result, and does not include these defect rights³.

The "a lone fighter is risky" objection mostly targets the first model. An experienced solo developer who delivers under a fixed-price contract for work defuses it structurally, not by assertion. One nuance belongs here: a fixed price is not automatically a contract for work, and time-and-materials billing is not automatically a service contract. The question of bogus self-employment applies almost only to the first model, the developer bought by the hour, and is barely relevant for an end customer ordering a finished result.

The three options honestly: strength and weakness

Each of the three routes has a real strength and an honest weakness. Anyone who only lists the others' weaknesses is selling, not advising.

An important distinction almost every agency article omits: an experienced senior working solo is not the same as a junior, and certainly not the same as a cheap overseas provider for a few hundred euros. Both are readily lumped together to make "freelancer" look young, cheap and unreliable across the board. What truly sets them apart is demonstrable senior experience, verifiable references and a clean contract for work.

Freelancer/solo: the strength is the direct line to the builder, short paths and a usually lower total price without an apparatus. The honest weakness is serial capacity and a bus factor of 1, that is cover when they are unavailable.
Agency: the strength is the bundled multidisciplinary team, backup and bench when someone is out, scaling by adding team members, and structured sign-off processes that count for corporate compliance. The honest weakness is the overhead, felt on small projects as onboarding and PM effort, and a higher minimum project size.
In-house team: the strength is the highest control and knowledge retention in your own house. The honest weakness is the fixed annual cost, the slow build-up through recruiting and knowledge lost when someone leaves; in a small team the bus factor is fragile here too.
The sharpened arguments against agencies, such as "Chinese whispers through many layers of coordination" or "the senior sells, the junior builds", are views that some solo providers advertise with. As a blanket fact they are not true: many agencies deliver excellent work, and their strengths remain real.

The unspoken risk: the bus factor

The strongest argument against a solo developer is the bus factor: what happens if the one person is out? Agency pages use it as their main weapon, and the point is fair. I name it openly rather than talk it away. A solo provider is initially a single point of failure, and that has to be honestly acknowledged.

What matters is how strongly that risk actually bites. Three things defuse it materially. First, a contract for work shifts the completion risk to the provider by contract rather than leaving it with the client. Second, a runnable, documented and handed-over code state makes the project continuable at any time, including by another developer. Third, the bus factor is not a purely solo problem: agencies have turnover too, and whoever was responsible for the result may well be long gone by the follow-up order.

That is why I actively guarantee code ownership, documentation and an always-runnable state as standard. With that, "solo equals unsafe" becomes a factual trade-off instead of a knockout argument.

Code ownership and vendor lock-in

Who ends up owning the source code does not depend on provider size but on the contract. Without an explicit clause, the client usually holds only a right of use and not the right to change the code or access the source². Lock-in therefore arises from missing rights and missing documentation, not from the freelancer-or-agency question.

Agencies mostly bundle full transfer of rights plus an NDA into the contract for work. With a solo developer a dedicated clause is often needed, but that evens out the moment they guarantee code ownership and documentation by contract just as well. That is exactly what I guarantee: you own the result, a runnable state and the documentation, no lock-in. Anyone with no access to code and data pays for every follow-up order on the provider's terms, regardless of whether that provider is a solo or an agency.

What the hourly rate leaves out

A low hourly rate does not automatically mean a cheap project, and a high agency rate is not pure margin. The agency rate is a blended rate, a mixed price that contains more than pure development. That is not a trick but a different cost structure worth understanding before you compare hourly rates.

The agency rate covers, beyond the pure development cost, the roles that cannot be billed directly such as project management, QA and testing, UX and UI design and architecture. On top comes the utilisation gap, because not every hour of a team is sellable, and finally sales, administration and profit. The price difference between freelancer and agency pays for exactly this bundle; there is no reliable primary data on actual agency margin in the German-speaking app market, so "the agency earns X percent" is speculation, not fact.

The honest conclusion for a decision-maker: what you compare is the total deliverable over the project's life, not the hourly price. What an app costs in detail, and how effort, day rate and running costs add up, is covered at length in a dedicated cost guide.

Two numbers for context

€94/h¹

Hourly rate in the freelancer Development segment (web, mobile, software), the app-relevant value. IT freelancers overall sit at €102 to €104/h (Freelancer-Kompass 2025), driven up partly by SAP and consulting.

around €136,000⁶

fully costed annual cost for an in-house senior (around €80,000 gross times a factor of 1.7), running on fixed regardless of utilisation. An industry benchmark and model calculation, not a fixed price.

My take

I write this from the very position this page is about: I am the experienced solo developer with an enterprise track record from roles at komoot, IBM and RTL. For clearly scoped projects with a direct line to the builder, that often makes me the faster and cheaper choice. I deliver at a fixed price on a contract-for-work basis, you own the code including documentation, and a paid app check from 3,999 EUR clarifies the scope up front before you commit to end-to-end development from 19,999 EUR.

The other direction belongs here just as honestly: if you need many parallel disciplines at once, permanent operation with guaranteed cover and fixed SLAs, or you have to pass corporate sign-off processes, then an agency or an in-house team is the better choice, not me. Anyone building apps permanently and in breadth is better off long term with their own team.

You do not have to go all-in straight away. A runnable prototype for 11,999 EUR makes your idea tangible within a few weeks and at the same time delivers a transferable state that defuses the absence risk from the start. That turns the blanket question "freelancer or agency?" into a sober decision about your specific project.

Frequently asked questions about freelancer, agency and in-house

Is a freelancer risky for app development, what happens if they drop out?

A solo developer is initially a single point of failure, and that is honest. The risk can be defused through three things: a contract for work shifts the completion risk to the provider by contract, a runnable and documented code state makes the project continuable by others at any time, and code ownership stays with you. With that, "solo equals unsafe" becomes a trade-off, not an exclusion criterion.

What does the developer owe, a contract for work or a service contract?

Which is cheaper, freelancer or agency?

The hourly rate is misleading. An experienced solo developer is usually cheaper because the overhead for project management, sales and office falls away. The higher agency rate is a mixed price that bundles PM, QA and design. What you compare is therefore the total deliverable over the project's life, not the bare hourly price. How an app price is made up is covered at length in a dedicated cost guide.

Can a single freelancer handle iOS, Android and backend alone?

When is an agency or an in-house team the better choice?

Sources

This page draws on the following sources. Market ranges are industry benchmarks corroborated by several independent providers, not official fixed prices. Hourly rates vary by seniority and stack.

Unsure which route fits your project?

In a free intro call I classify your project and tell you honestly whether an experienced solo developer is the right choice or whether a team fits better. Tobias Boehm, senior app developer from Northern Germany.