Linear sprint planning: a practical checklist

Sprint planning goes sideways for the same reasons at every team: no capacity model, estimates that drift mid-cycle, no clear cycle goal, carry-over that never gets acknowledged. Here's a checklist that covers the boring stuff teams usually skip.

Before the meeting (about 60 min of prep)

• Groom the backlog — close duplicates, split oversized issues, add missing estimates
• Confirm the Linear cycle dates match real working days (holidays, company offsites)
• Check who's off, on-call, or interviewing — mark availability per person
• Review last cycle: committed vs delivered, carry-over count, estimate drift
• Decide the one thing this cycle must ship — the cycle goal

During planning (30–45 min)

• State the cycle goal first. Not the backlog.
• Pull top candidate issues into the cycle view
• Assign owners before discussing estimates — unowned estimates are fiction
• Confirm every issue has an estimate. Unestimated = not planned.
• Sum committed estimate per person. Compare against that person's realistic capacity.
• Flag overloaded people and reassign or cut
• Surface known blockers per issue ("what will stop me from finishing this")
• Leave 20–30% slack for unplanned work. If you don't, the unplanned work will take slack from planned work.

After the meeting

• Post a one-line cycle summary somewhere visible: goal, headline issues, known risks
• Confirm each assignee knows their top 1–2 issues
• Look at the total: is this cycle realistically deliverable? If not, cut now, not mid-cycle.

Daily during the cycle

• Use a standup view that shows overloaded people, blocked issues, and unestimated work
• Flag upward estimate changes as they happen — that's the early signal of a slip
• If urgent work comes in, something planned goes out. Update Linear explicitly in both directions. "Just adding it" is how cycles miss.

At cycle end

• Compare committed vs delivered estimate
• Compare original estimate vs final estimate for shipped issues — this is estimate drift
• Decide what carries, what closes, what goes back to backlog
• Run a short retro. Two questions: what slowed us down, what do we change next cycle.

The meta-point

Sprint planning isn't a ceremony, it's a forecast. A forecast with no inputs (capacity, history, constraints) is just a wish. This checklist is the minimum set of inputs that makes the forecast honest.

Teams that follow even half of it deliver more predictably than teams that rely on memory and optimism.