Contents
App acceptance testing is the moment when you as the client confirm that the delivered work meets the agreed requirements – with real legal and financial consequences. To get it right, you need acceptance criteria defined in advance, a structured test plan and clear rules for handling bugs. This article shows how quality assurance and acceptance work together in practice.
Why formal acceptance matters
For custom-built software in Germany, the contract is often a contract for work (Werkvertrag) or contains elements of one. Formal acceptance (Abnahme) then usually has significant consequences: payment or a final instalment becomes due, the warranty period begins and the burden of proof for defects may shift. How this is regulated in your specific contract depends on its wording, and the rules differ between jurisdictions – this article is not legal advice. Read the acceptance clauses in your contract carefully and have them reviewed by a lawyer if in doubt.
In practice this means that signing off "in passing" by e-mail, without thorough testing, can become expensive. At the same time, acceptance should not be delayed endlessly over trivialities, because that blocks the launch and strains the relationship. Background on liability can be found in our article on liability for app defects.
Quality assurance does not start at acceptance
Good quality is created during development, not in the final test. Ask your development partner which measures take place continuously:
- Code reviews: every change is reviewed by a second person – including, and especially, AI-generated code.
- Automated tests: unit and integration tests safeguard the business logic and run automatically with every change.
- Testing on real devices: different iPhone and Android models, screen sizes and operating-system versions.
- Sprint acceptance: completed user stories are checked against their acceptance criteria after each sprint.
- Security and privacy review: handling of credentials, encryption, permissions and consent.
When work is checked sprint by sprint, final acceptance is mainly a confirmation – not a surprise.
Defining acceptance criteria properly
Acceptance criteria describe how you measure whether a requirement has been met. They should be specific enough that two people independently reach the same verdict. A weak criterion reads: "Login works reliably." A strong one reads: "Users can sign in with e-mail and password; after three failed attempts a hint to reset the password appears; the password reset e-mail arrives within a few minutes."
Complement functional criteria with non-functional requirements that are easily forgotten: supported operating-system versions, behaviour without an internet connection, accessibility, loading times with typical data volumes and data protection requirements. Put everything in writing, ideally during the concept phase.
Bug severity levels: what blocks acceptance and what does not
Not every bug justifies refusing acceptance. A proven approach is to agree on severity levels in advance:
| Severity | Description | Example | Effect on acceptance |
|---|---|---|---|
| Critical | Core function unusable, data loss or security vulnerability | Payments are not recorded, app crashes on launch | No acceptance; fix before retesting |
| Major | Important function impaired, workaround cumbersome | Push notifications do not arrive on some devices | Usually no acceptance, or only with a short fix deadline |
| Medium | Function usable but faulty or awkward | A list is sorted incorrectly | Conditional acceptance with a list of open items |
| Minor | Cosmetic issues with no effect on use | Typo, slightly misaligned icon | Acceptance; fix in the next update |
The acceptance test step by step
- Create a test plan: derive concrete test cases from the user stories – with starting situation, steps and expected result.
- Prepare the test environment: a test build via TestFlight or Google Play's testing tracks, realistic test data, access to connected systems.
- Choose testers: domain users who know the workflows, plus at least one person seeing the app for the first time. Our article on beta testing with TestFlight and Google Play explains the process.
- Cover devices: test on several iOS and Android devices, including older models your target group uses.
- Document bugs properly: device, OS version, steps to reproduce, screenshot or screen recording and the agreed severity level.
- Retest: check fixed bugs specifically and briefly verify that nothing else has broken in the process.
- Document acceptance: a written record with date, version, result and a list of open items with agreed deadlines.
Checklist before acceptance
- All agreed user stories are implemented and tested against their acceptance criteria.
- There are no open critical or major bugs.
- The app runs on the agreed operating-system versions and device classes.
- Privacy policy, consent handling and account deletion work.
- For apps handling sensitive data, security has been checked – where appropriate through a penetration test.
- Source code, documentation and credentials have been handed over or are in your own repositories and accounts.
- Monitoring and crash reporting are set up.
- Store listings, screenshots and legal texts are ready – see the full app launch checklist.
Conclusion
A good acceptance process is the result of continuous quality assurance, clear acceptance criteria and a structured test plan. Agree on severity levels in advance, test on real devices with realistic data and document the result in writing. That way you protect your investment and still reach launch quickly. If you are planning a project with clear quality standards and a fixed price, start with our offer wizard.
Frequently Asked Questions
Can I refuse acceptance because of minor bugs?
Minor, cosmetic defects generally do not justify refusing acceptance. The usual approach is acceptance with a list of open items and agreed deadlines; the exact rules follow from your contract.
How long should the acceptance phase take?
It depends on the scope. For an MVP a few days are often enough; for larger apps you should plan several weeks including retests.
Who writes the test cases for acceptance?
Ideally both sides together: the client's domain experts know the workflows, and the agency helps with structure and technical test cases. The basis is the agreed user stories and acceptance criteria.
What happens if bugs appear after acceptance?
The warranty or maintenance arrangements agreed in the contract then apply. Document the bug precisely and report it promptly through the agreed channel.