Digitalisierung

Power Apps: What Does an App Cost and How Long Does a Project Take?

"What does a Power App cost?" is the question we hear most often — and the one most honestly answered with "it depends". Not an evasive answer but the truth: a Power App can cost 4,000 euros or 120,000 euros, and both figures can be right for the same "app", depending on what is underneath.

This article makes that concrete. It belongs to the Power Platform guide for the mid-market and complements Power Automate and GDPR and enterprise readiness. Ballpark figures instead of fixed prices — orders of magnitude you can plan with.


1. Why "the app costs X" is a meaningless question

A Power App is not a product with a price tag but a build commission — like a house whose price depends on square metres, bathrooms and the plot. Four factors determine it almost entirely: the data complexity (a flat SharePoint list is a different world from Dataverse with 14 linked entities), the depth of the business rules ("show me a list" is trivial, "calculate the discount according to seven rules" is software development), the integrations (every connection to an ERP or DATEV is its own small project) and the user base (five people can wait now and then, 200 field staff need load tests and monitoring).

Whoever names a price without knowing these four things is guessing — and you pay the difference later as a change request.

Three size classes — realistic ballpark figures Small €4–15k 1 list · 3–6 weeks Capture · checklist Canvas App Medium €15–45k Dataverse · 2–4 months Logic · 1–2 integrations Model-Driven Large €45–120k+ critical · 4–9 months Load · monitoring Platform project The data determines the class, not the screen Same screen: €6k on a list, €38k on a clean data model. Licences come on top. Source: Medienstürmer

2. Three size classes with honest ballpark figures

Three corridors from real projects, all as build effort excluding licences and operations:

Small — 4,000 to 15,000 euros, 3 to 6 weeks. A Canvas app on a SharePoint list or a lean Dataverse table: capture form, checklist, simple approval. The platform's sweet spot — replaces an Excel that has grown too big, often amortises in under a year.

Medium — 15,000 to 45,000 euros, 2 to 4 months. A Model-Driven app on Dataverse with a real data model, a role concept, several business rules and one or two integrations. The most common case in the mid-market: replacing a wild growth of island solutions. Here the choice between Canvas, Model-Driven and Power Pages significantly determines the effort.

Large — 45,000 to over 120,000 euros, 4 to 9 months. Business-critical, many integrations, a large user base, load and availability requirements. Here Power Apps is no longer a "low-code toy" but a development platform with architecture, tests and lifecycle management — and should be treated as such, with experienced guidance rather than solo.

The honest warning: most budget overruns do not arise from too low a calculation but because an app unnoticed slides from "small" to "medium" — through wishes that sound harmless individually and in total blow up the data model.

3. Where the time really goes (spoiler: not into clicking)

The biggest misconception with low-code: "you click that together in an afternoon." The interface is quick — but rarely more than 20 percent of the effort. The lion's share is devoured by requirements and the data model, integrations, and testing and migration (the first thing cut on tight budgets — and expensive because of it; see the chart). Whoever budgets "just the app" budgets one fifth of the project.

Where the effort really flows Requirements and data model 30–35 % App build (UI and logic) 20–25 % Integrations 15–25 % Test · migration · hardening 15–20 % Rollout and training 5–10 % Clicking is cheap — the thinking before it is the bill A third flows in before the first screen even exists. Whoever saves on it pays it back double. Source: Medienstürmer

4. The running costs nobody talks about

The build is one-off. Three items afterwards are chronically underestimated:

Licences. Power Apps is licensed per user or per app — cheap with standard connectors. As soon as Dataverse, premium connectors or external Power Pages users come into play, the licence becomes a permanent operating expense that scales with the success of the app, not with its effort.

Operations and maintenance. Plan realistically 15 to 25 percent of the build cost per year for updates, connector deprecations and monitoring. Without a maintenance budget an app has an expiry date.

Further development. A well-adopted app generates wishes — a quality signal, but a budget item. Better a small fixed pot than every change as a crisis change request. These three items are not a Power Apps disadvantage — they apply to every piece of custom software, they are just more honestly visible here.

5. How to keep the budget realistic

Four levers keep projects reliably within the corridor: the data model first, then the app — if it becomes too complicated, a dedicated Dataverse architecture belongs underneath, not a list. A fixed scope for version 1 with a negative list of what does not belong to it. Deliver in iterations: a usable version after 4 to 6 weeks uncovers more flaws than three months of specification. Clarify licence costs before the build — otherwise you build an app whose operation is more expensive than its benefit.

That gives no guarantee to the cent — that does not honestly exist with custom software — but a corridor that holds. Worth more than a fixed price that breaks.


Conclusion: the price is honest when the question is

A Power App costs as much as its data, rules, integrations and user base demand — between a few thousand and six figures. Every figure without these four details is guessed.

Three things to take away: the build is one fifth of the project — requirements, integrations and testing are the rest. Licences and maintenance are not a surprise but a planning obligation. And a held corridor beats a broken fixed price. Power Apps is not "cheap software" but faster and more maintainable custom software — a far better deal when you do the maths honestly.

Let us estimate your app — with real numbers

Describe to us what the app should do, and we tell you honestly which corridor you land in — including licences and operations. No estimate into the blue, a reasoned ballpark figure.

Sources