We had 3 “swim lanes” “P1” or “expedite”, Engineering and Design. The lanes converged in the verify column.We had WIP limits for each column or “state”Engineering Lanes: Queue, Requirements & Test Plans, Design, Review& Revision, Code, Review & Revision, Verify, DoneDesign Lanes: Queue, User Experience, UX Review & Revision, Design, Review& Revision, Test Plan, Code, Review & Revision, Verify, Done
We decided to use post it “flags” to indicate a blocked item.We created a parking area for “work not on the board” for visibility sake.
Breaking items into tasks allowed for better tracking of progress, and made it easier for people to see item that they could collaborate on.
We had trouble managing dependencies between the two boards/teamsWe decided to pull together as one team, one board, one goal
Cycle times look good, but would like to see more predictability on Pebbles.
Spike in February cycle times due to Know How items
Standard Deviation is a measure of predictability. The bigger the standard deviation, the less predictable we are.
We still “retro”. We try to look at items that exceed estimates to try to learn from our mistakes.
Throughput dramatically higher in March due to Know How. May throughput down because of Google/Open ID and Network Solutions items. Lots of work, but didn’t deliver until June. Increase in the number of bugs also affects throughput.
This example is from September of last year. Pretty smooth.
Throughput was down because we had 4 large projects going on simultaneously, so smaller items blocked.Infrastructure work (e.g. code cleanup, upgrades, etc.) were not getting done.Team defined classes of service – pebbles, rocks, sand, plus infrastructure and design only.