Acceptance Criteria Examples: Turn Vague Requirements into Testable Outcomes

Acceptance Criteria Examples: Turn Vague Requirements into Testable Outcomes — original editorial illustration
Original AI-assisted illustration.

“The export should work well” is difficult to build, test or accept consistently. Acceptance criteria make the intended behaviour more concrete by describing observable outcomes and the conditions under which they apply. They help a team find disagreements before those disagreements become rework.

The examples below are fictional business rules for a task-management application. They are teaching examples, not requirements you should copy into a product without discussion.

Separate the goal from the conditions of success

A user story can describe who needs something, what they want and why it matters. Acceptance criteria describe the conditions that must hold for that particular story to be accepted. Atlassian describes both the user perspective and acceptance criteria as useful parts of a story.

Example goal: as a project member, I want to export completed tasks so that I can reconcile weekly work. That goal still leaves questions about filters, permissions, dates, empty results and the contents of the file.

Example 1: replace “all relevant tasks” with a defined rule

Vague criterion: the export contains all relevant tasks.

Proposed criterion: for a project member allowed to export a project, a completed-task export contains every task matching the active completion-date filter at the time the export begins.

Now ask the questions the proposal exposes. Which time zone defines the dates? Are boundary dates inclusive? Does the export reflect a snapshot if tasks change during generation? Agree the answers and add examples. Precision is useful only if it reflects the intended business rule.

Four-step framework for Acceptance Criteria Examples: Turn Vague Requirements into Testable Outcomes
Illustrative framework; adapt to your team and agreed requirements.

Example 2: describe an empty result

Vague criterion: handle no results gracefully.

Proposed criterion: when no tasks match the selected filter, the user sees the agreed “No completed tasks match these filters” message and no export file is generated.

A different product might deliberately generate an empty CSV with headers. Both are plausible. The purpose of the criterion is to decide which behaviour this product should have, rather than letting developers and testers choose independently.

Example 3: use Given, When and Then where helpful

Cucumber’s Gherkin reference uses Given for starting context, When for an event and Then for the expected outcome. You can use that structure to discuss examples even without writing automated tests.

Given a project member can view project DEMO but does not have export permission,

When the member attempts to request an export,

Then the request is denied and no export data is returned.

The criterion describes the outcome. A separate technical discussion can decide how permissions are enforced. An interface that merely hides a button does not by itself demonstrate that an export request is denied.

Include boundaries and exceptions without writing a novel

For a date filter, agree the first included moment, the last included moment and the time zone. For an amount limit, discuss below, at and above the boundary. For optional fields, decide whether missing and blank values mean the same thing.

Keep a short table of rule, example input and expected outcome. Add only examples that clarify a decision or a risk. Repeating the same happy path with different names can create volume without improving understanding.

Review criteria together

Ask a business representative, developer and tester to read each criterion. Can they identify the same starting conditions and the same expected outcome? Are terms like “fast,” “easy” and “secure” supported by agreed measurable requirements or separate specifications?

Do not treat acceptance criteria as the whole definition of quality. Team-wide completion standards, accessibility, security and other requirements may apply across many stories. Record links to those standards so that a concise story does not silently exclude them.

A short review checklist

Does the criterion support a real user goal? Is the outcome observable? Are terms and boundaries defined? Have permissions and important exceptions been discussed? Can a person explain how to check it? Are unresolved questions visible rather than hidden in assumptions?

If the answer is unclear, revise the example together. A small set of agreed examples is more useful than a long list of ambiguous statements.

Sources and method

References: Atlassian’s user story guide and Cucumber’s Gherkin reference. The task-export rules and examples are original illustrations.

Related reading

How to choose test management software for a small QA team · Software testing guides · Software buying guides

Prepared with AI assistance. Worked examples are fictional unless stated otherwise.

Comments

Popular posts from this blog

How to Choose Test Management Software for a Small QA Team