Why this matters: half of all planning arguments are actually vocabulary arguments where two people use the same word for different things. This is the page to settle them. Each term links to the note that explains it properly.
The work breakdown
| Term | In one line |
|---|---|
| Epic | Large business objective. Spans releases. Explained in Work Breakdown Hierarchy. |
| Feature | Major product capability. Spans sprints. Where ownership lives. |
| User Story | Small testable user outcome. Must fit in one sprint. Craft in Writing User Stories. |
| Product Backlog Item | The same as a User Story, in Azure DevOps’s Scrum template. See Azure DevOps Setup. |
| Task | Technical implementation work. A day or less. |
| Spike | Timeboxed investigation whose output is a decision, not a feature. |
| Enabler | Real work with no direct user outcome, written with an honest reason. |
flowchart TD
A["Epic<br/>WHY the business objective"]
B["Feature<br/>WHAT the capability"]
C["Story<br/>WHAT the small user outcome"]
D["Task<br/>HOW the implementation"]
A --> B
B --> C
C --> D
The three gates
| Term | The question it asks | When |
|---|---|---|
| Definition of Ready | Can we safely start this? | Before the sprint |
| Acceptance Criteria | What exactly should this Story do? | Written at refinement, checked at the end |
| Definition of Done | Is this professionally complete? | Before saying finished |
Acceptance criteria are per Story and change constantly. Definition of Done applies to everything and changes rarely. Definition of Ready is a team practice, not an official Scrum artifact, while Definition of Done is.
Sizing and planning
| Term | In one line |
|---|---|
| Story Points | Relative size: work + complexity + risk + uncertainty. Never hours. See Estimation and Story Points. |
| Fibonacci scale | 1, 2, 3, 5, 8, 13. Widening gaps prevent false precision as size grows. |
| Planning Poker | Estimate privately, reveal together, then discuss the disagreement. The discussion is the point. |
| Reference Story | A completed Story used as the anchor for what a 3 or a 5 means in your team. |
| Velocity | Historical points completed per sprint. A forecast, never a target. See Velocity and Capacity. |
| Capacity | Actual availability this sprint, after leave and support duty. |
| Carry over | Committed work that did not finish. Track it, and name the cause honestly. |
| Sprint Goal | The one sentence objective that gives the sprint focus and lets you refuse unrelated work. |
| Refinement | The ongoing work of clarifying, splitting, sizing upcoming items. See Sprint Planning and Refinement. |
Quality and operations
| Term | In one line |
|---|---|
| Test case | A precondition, an action, and an expected result specific enough that pass or fail is unarguable. See Test Case Strategy. |
| Traceability | The link from test case to acceptance criterion to Story. Without it, coverage is unknowable. |
| Regression suite | The automated tests that run every build. The asset that keeps you fast later. |
| INVEST | Independent, Negotiable, Valuable, Estimable, Small, Testable. The Story quality check. |
| Three Amigos | Product, development and testing look at a Story together before it is called ready. |
| Example Mapping | 25 minute technique: collect rules, one example per rule, and open questions. |
| SLI | What you measure in production. Example: percentage of payments that succeed. SLI SLO and Error Budget |
| SLO | The target for an SLI. Example: 99.9% over 30 days. |
| Error budget | The failure the SLO allows. When it is spent, reliability work takes priority over features. |
| Toil | Manual repetitive operational work that scales with usage. Measure it, then automate it. |
| Postmortem | Blameless write up after an incident. Actions become backlog items or it was theatre. |
The rules worth memorising
- If a tester cannot write a test case for it, it is not a Story yet.
- Nothing enters a sprint without acceptance criteria.
- New capability from feedback becomes a new backlog item, not a bigger current Story. See Feedback and Scope Control.
- Urgent work enters only when something equivalent leaves.
- Done or not Done. No partial credit, ever.
- Points measure size, not time. Velocity forecasts, it does not grade.
- If the behaviour is not written down anywhere, the feature is not finished. See Feature Inventory and Ownership.
- Defend the Sprint Goal, not the task list.
The four things to measure
Not “are people busy”:
- Carry over rate, and the named cause for each item.
- Points added mid sprint, which is the scope growth number.
- Stories entering a sprint without acceptance criteria, which should be zero.
- Behaviour coverage, meaning live features with written criteria and at least one test case.
The full argument for all of this is in Why We Miss Deadlines, and the order to introduce it in is Adoption Plan.