Sprint planning that survives contact with reality

If your team ends most sprints with a pile of unfinished work, the instinct is to blame focus. Usually the real culprit is a plan that assumed a perfect two weeks: no incidents, no sick days, no context-switching, no surprises.

Plan for the team you actually have. Subtract meetings, support rotations and the one person who's out on Thursday before you commit to scope. A sprint that's 70% booked leaves room for the real work that always shows up.

Make dependencies visible on the board, not in someone's head. The task that's blocked on design or a third-party API is the one that quietly sinks the sprint. Flag it on day one, not on day nine.

And keep a short, honest retro. Not a ceremony for its own sake, but two questions: what did we underestimate, and what will we do differently next time? Teams that adjust their estimates every sprint get scarily accurate within a quarter.