Sprint Planning

Sprint Planning

A sprint plan is a bet: this team, this much time, against one outcome that matters. Most plans fail because they're packed to 100% of theoretical capacity, chase five goals at once, and ignore the meetings and PTO that eat real days. This skill builds a plan grounded in actual available hours, anchored to ONE goal, with honest cut lines. The job isn't to fit the most work in — it's to commit to the right work and protect it.

Grounded in: Scrum: The Art of Doing Twice the Work in Half the Time — Jeff Sutherland: capacity-honest commitment around one sprint goal. Go deeper (The Product Channel): The Product Channel

When to use this

• Kicking off a new sprint and need to scope what's realistic.
• Sizing a backlog against who's actually available (PTO, meetings, on-call).
• Deciding what's P0 (must ship) vs. P1 vs. stretch.
• Handling carryover from the last sprint without silently overcommitting.
• Sanity-checking a plan someone else drafted: "does this actually fit?"

Before you start (gather these)

Ask for anything missing before planning — don't guess capacity.

• Backlog items — pasted list with rough sizes, or note: "I can pull open issues from Linear/Jira if connected."
• Sprint length — e.g., 1 or 2 weeks; working days only.
• Team + availability — who's on the sprint, plus PTO, holidays, on-call, part-time splits.
• Carryover — unfinished work from last sprint and its remaining estimate.
• Sprint goal candidate — the one outcome this sprint is for (if they have one).
• Recurring overhead — standing meetings, ceremonies, support rotations per person.

Process

Ask before you plan. Ask 2-4 questions and wait for the answers: "What's carrying over from last sprint?" · "Who's out, and for how long?" · "What's the one outcome that makes this sprint worth running?" · "Anything already committed to a date?" Planning a sprint on assumed capacity is the single most reliable way to miss it. If workspace/projects/ holds last sprint, read it first and state what you took from it.

1. Compute realistic capacity. For each person, subtract overhead first, then take the buffer off what's left — the 20% buffer is 20% of the subtotal AFTER subtracting PTO/holidays, meetings, and on-call, not 20% of raw working days. Formula: Available = (working days − meetings − PTO/holidays − on-call) × 0.8. The buffer absorbs interrupts, reviews, and the unknown. Sum each person's Available to a team total. This number, not the calendar, is your budget.
2. Set ONE sprint goal. A single sentence stating the outcome and why it matters — not a list of tickets. If you can't name one goal, the sprint isn't scoped yet.
3. Rank against the goal. Tag each item P0 (goal fails without it), P1 (high value, goal survives without it), or Stretch (nice if time remains). Carryover competes on the same ranking — it is not auto-committed.
4. Fit to capacity. Fill P0 first, then P1, stopping when committed estimates reach your buffered capacity. Draw a visible cut line. Everything below it is stretch or next sprint.
5. Name an owner per item. No item ships without exactly one accountable owner. Flag any person loaded past their individual capacity.
6. Flag dependencies & risks. External teams, design/research not-yet-done, unknowns needing a spike, single points of failure. Each gets an owner and a "what we'll do about it."
7. Protect the buffer. Do not backfill it with stretch work. It absorbs reality; spending it upfront guarantees a miss.

Output template

# Sprint Plan — [Sprint name / dates]

## Sprint Goal
[One sentence: the outcome this sprint delivers and why it matters.]

## Capacity Summary
| Person | Raw | −Meetings | −PTO | −On-call | Subtotal | −Buffer (20%) | Available |
|--------|-----|-----------|------|----------|----------|---------------|-----------|
| [Name] | [N] | [m] | [p] | [o] | [s] | [s×0.2] | [s×0.8] |
| ...    | ... | ... | ... | ... | ... | ... | ... |
| **Team total** | | | | | | | **[N committable days]** |

## Committed (≤ capacity)
**P0 — must ship**
- [ ] [Item] — owner: [Name] — est: [Xd]
**P1 — high value**
- [ ] [Item] — owner: [Name] — est: [Xd]
_Committed total: [Xd] / [capacity Yd]_

## Stretch (only if time remains)
- [ ] [Item] — owner: [Name] — est: [Xd]

## Carryover Handling
- [Item from last sprint] → [re-committed as P0 / deprioritized / dropped / absorbed into [item] (estimate already counted there — not double-counted) + why]

## Dependencies & Risks
- [Dependency/risk] — owner: [Name] — mitigation: [action]

## Definition of Done
- [ ] Merged & deployed to [env]  · [ ] Tested  · [ ] Docs/tracking updated  · [ ] [team's DoD criteria]

Quality bar

• Capacity subtracts meetings, PTO/holidays, on-call and a buffer (~20%).
• Total committed estimate is ≤ buffered capacity (verify the math).
• Exactly one crisp sprint goal, written as an outcome.
• Every committed item has exactly one owner; no person is overloaded.
• Carryover is explicitly re-ranked, not silently rolled in.
• Dependencies and risks each have an owner and a mitigation.

Avoid (anti-patterns)

• Planning to 100% capacity. Interrupts, support, and review are real work. A sprint planned to full capacity is a sprint that misses.
• Ignoring carryover. Unfinished work from last sprint is committed capacity, not a bonus. Count it first, then plan what's left.
• A sprint goal that's a list. "Ship A, B, and C" is not a goal. The goal is the outcome that makes the sprint worth running, and it survives dropping C.
• Stories with no acceptance criteria. They will be re-litigated at review, in front of stakeholders, at the worst possible moment.
• No PTO or holiday check. The single most common cause of a missed sprint, and the easiest to catch.
• Pulling in the biggest unknown. If an item can't be estimated, the sprint's job is to spike it, not to commit to it.

Tips

• Never plan to 100%. Buffered capacity is the real ceiling — full calendars always overrun.
• One goal, not five. If everything is the goal, nothing is. Sub-goals are just P1 items.
• Protect the goal from scope creep. Mid-sprint adds displace something — make the trade visible, don't absorb it.
• Carryover is a smell, not a default. Chronic carryover means you're overcommitting; cut, don't roll.
• Estimate in ideal days, plan in available days — the gap between them is exactly the overhead you must subtract.
• A spike is a commitment too. If an item is too unknown to size, the P0 is the spike, not the build.
• Round to the nearest 0.5 day. Capacity estimates aren't precise to the decimal — round each Available figure to the nearest half-day and treat the "≤ capacity" fit as approximate. Don't manufacture false precision like 5.2 / 6.4 / 0.6 days; 5.0 ≤ 6.5 with a little room is the honest read.