Story Points
Size, not time. Why Fibonacci's gaps are a feature, why a 4-and-5 disagreement is noise, and why a 13 is a signal to split or spike rather than an estimate to commit to.
Filter inventory by location
Filter inventory by location
Translating points into hours guarantees a fight at the next sprint planning. Hours vary per engineer, per day, per interruption rate. Points are deliberately abstract because the abstraction is useful.
Story points are relative size: complexity, uncertainty, and effort bundled into one number compared to a reference story. The right ticket anchors the 5 to a known 1-pointer rather than to hours.
Team estimation scale
Allowed values: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10.
Team estimation scale
Allowed values: 1, 2, 3, 5, 8, 13. 13 means 'we do not understand this yet, split or spike it'.
A linear 1-10 scale invites the team to argue between 4 and 5 for ten minutes, both of which were the same noise on a Fibonacci scale. The illusion of precision costs more than it adds.
Fibonacci's gaps are the point. Weber's Law: humans cannot reliably distinguish quantities that differ by only a small ratio. A 4 and a 5 are noise. A 3 and an 8 are a real conversation about scope.
Self-serve onboarding for enterprise tenants
Identify the smallest end-to-end onboarding slice
Pulling a 13 into a sprint is taking a bet. Half the time the team will discover the work was actually 21 and the sprint will end with 'in progress' tickets that nobody can land.
13 is a signal that the team is uncertain, not an estimate they can commit to. The right move is to split or spike. A spike that returns three sized stories is cheaper than a 13-pointer that misses the sprint and re-splits under pressure.
How the team points stories
Each engineer estimates how long it would take them. We average. New engineers calibrate over time.
How the team points stories
Reference stories pinned in the team space: a 1 (config flag toggle), a 3 (typed form with two endpoints), an 8 (cross-cutting feature with a migration). All pointing happens against those.
Time-based estimates drift per engineer and per week. Without reference stories, a new joiner's 3 is a senior engineer's 1, and the team's velocity becomes unreadable.
Reference stories anchor the team to a shared baseline. New joiners and old hands point against the same yardstick, which makes the spread compress over time without anyone having to memorize hours.
Managers compare team throughput across quarters
Managers compare team throughput across quarters
Pulling an 8 silently means the team has not decided what happens when it slips. The sprint review will rediscover the same conversation, except now there is in-flight work to split around.
An 8 probably wants to be two stories, but a single sprint can absorb one if the team flags it as a risk and agrees on the re-split rule up front. That is the difference between an honest commitment and an optimistic one.
Sprint capacity
Velocity last sprint: 32. Target this sprint: 32. Pull stories until the total reaches 32.
Sprint capacity
Velocity over the last three sprints: 28, 32, 24 (median 28). Pull stories until the total reaches roughly 28, then leave room for unplanned work.
Pulling to last sprint's number assumes last sprint was typical. The team rotates, holidays land, fires happen. Pulling to that single number guarantees an over-commit half the time.
Velocity is a planning aid, not a target. The median across recent sprints is more honest than the last sprint. Leaving headroom for unplanned work means the team finishes; treating velocity as a target means it becomes a ratchet.
Pointing a story
Three engineers point: 2, 2, 13. Average to 6, round to 5, move on.
Pointing a story
Three engineers point: 2, 2, 13. The 13 saw something the others did not. Talk it through, then re-estimate.
Averaging a wide spread is the worst move. The number you produce was never an estimate; it was a compromise that buries the disagreement. The work will hit the missed scope mid-sprint.
The point of pointing is the conversation. A spread of 2-2-13 is the signal that two of the three are missing something the third sees. Average it and you ship the same misunderstanding into the sprint.