All terms
Glossary · Requirements

Acceptance Criteria

Acceptance criteria are the specific, testable conditions a user story, feature or other deliverable must meet to be accepted as complete and correct. They describe expected behavior from the user's or stakeholder's point of view, including important edge cases and exact messages, so product, design, engineering and QA share one definition of "working".

How Acceptance Criteria Work

The BABOK Guide defines acceptance criteria as the criteria associated with requirements, products or the delivery cycle that must be met to achieve stakeholder acceptance. Mike Cohn calls them conditions of satisfaction. In Ron Jeffries' card, conversation and confirmation model of a user story, they are the confirmation: how the customer will know the work is done.

Two formats are common:

  • Rule-oriented: a checklist of statements, such as "Passwords must contain at least 12 characters."
  • Scenario-oriented: the Given/When/Then structure from behavior-driven development (BDD), introduced by Dan North. Given some context, when an event occurs, then certain outcomes follow. North put it simply: a story's behavior is its acceptance criteria.

Criteria are written during backlog refinement, before development starts, and agreed between the product owner and the team. They then become the basis for test cases, many of them automated.

Good criteria are pass or fail, describe outcomes rather than implementation, and cover the main path plus the edge cases that matter. If a criterion cannot be verified, it is not a criterion yet.

Why Acceptance Criteria Matter

Acceptance criteria remove ambiguity at the point where it is most expensive. Without them, a story can be "done" by the developer's reading and wrong by the product manager's. They also set the boundary of a story, which limits scope creep within a sprint, and they give testers a ready-made starting point.

For AI coding agents, criteria act as the check the output must pass. An agent given clear Given/When/Then scenarios can write tests first and verify its own work against them. For how criteria fit into well-formed tickets, see how to turn OKRs into sprint-ready tickets with clear acceptance criteria.

Acceptance Criteria Example

Story: "As a shopper, I want to apply a discount code at checkout, so that I pay the promoted price."

  • Given a valid code, when I apply it, then the order total drops by the discount and the code appears on a discount line.
  • Given an expired code, when I apply it, then I see "This code has expired" and the total does not change.
  • Given a code is already applied, when I apply a second code, then the second code replaces the first.

The third criterion records a product decision that would otherwise be left to whoever writes the code.

Acceptance Criteria vs. Definition of Done

Acceptance criteriaDefinition of Done
Applies toOne specific story or featureEvery backlog item
DefinesThe behavior that item must showThe quality standard all work must meet
Set byProduct owner with the team, per itemThe team or organization, once

A story is complete only when it meets both its own acceptance criteria and the team's Definition of Done.

Related terms
User Story
A short, user-centered description of a need: as a role, I want an action, so that an outcome. Paired with acceptance criteria.
Definition of Done (DoD)
A shared, explicit standard of everything that must be true before any piece of work counts as complete.
Edge Case
An edge case is a situation at the limits of expected inputs or operating conditions that checks behavior beyond a product’s typical path.
Definition of Ready (DoR)
The criteria a backlog item must meet before a team will pull it into a sprint.
Test-Driven Development (TDD)
A practice of writing a failing automated test before the code that makes it pass, then refactoring, in short cycles.
Product Requirements Document (PRD)
A product requirements document defines what a product or feature must do, why it matters, and how success will be assessed.
Put the method into practice.
Prodstack is the AI product operating system that turns terms like this into shipped, evidence-backed work — from discovery to growth.
Start your 7 days free trial
// Newsletter
Field notes, in your inbox.

Evidence‑driven thinking on discovery, prioritization, specs, and shipping, plus new articles the moment they drop. No noise.