New: Practical guidance for AI-assisted quality engineering
QA Fundamentals · Cornerstone Guide 07

What Makes a Requirement Testable?

Learn how to recognize testable requirements, improve acceptance criteria, expose ambiguity, and prevent requirements from turning into late testing problems.

10 min readFor QA analysts, developers, product teams, business analysts, and delivery leaders.

Testability means observable evidence

A requirement is testable when the team can determine, with reasonable objectivity, whether the delivered solution satisfies it.

That generally requires a clear condition, expected behaviour, and observable result.

Characteristics of a testable requirement

  • Clear
  • Specific
  • Unambiguous
  • Measurable where appropriate
  • Observable
  • Feasible
  • Traceable
  • Complete enough to design tests

Weak vs testable examples

Performance

Weak: “The page should load quickly.”

Testable: “The page should load within 2 seconds for 95% of requests under the agreed normal workload.”

Authorization

Weak: “Only authorized staff can edit records.”

Testable: “Editor users can modify active records; Viewer users can read but cannot create, edit, or delete records.”

Questions QA should ask

  • What business outcome is required?
  • Who is allowed to perform the action?
  • What happens when input is invalid?
  • What are the boundaries?
  • What if a dependency is unavailable?
  • What data states must be supported?
  • How will success be observed?

Acceptance criteria are not the whole story

QA should also consider negative behaviour, interfaces, data, permissions, error handling, performance, accessibility, and operational impact.

Definition of Ready should enable delivery, not become bureaucracy. Use it to expose missing information before commitment.

Have a real QA situation to solve?

AskAQA connects fundamentals to practical delivery, automation, AI, data, governance, and release decisions.

Ask a QA Question

AskAQA AI assistant

Ask a QA

Ask me about software testing fundamentals, QA concepts, test levels, test types, test cases, requirements, defects, coverage, or other foundational quality topics.

Do not include passwords, confidential information, or personal data in your question.