Fast, #NoWaste Agile Estimating
Ron Lichty, http://www.ronlichty.com
Desire to make software development fun
Co-author of The Study of Product Team Performance
(Pointers to summaries and the free study downloads at:
http://www.ronlichty.com/study.html)
Ron's Context
• SW Development is a team sport
• The point is not to do agile, rather to be effective.
• Simplicity - maximizing the work not done. (â€"Agile Manifesto)
Why estimate?
• Org needs to have some level of predictability.
• Provide ROI for Product Owner -- the buck of the bang for your buck
• Stakeholder confidence
Uses an example on how to prepare a pineapple to eat and clean up when training execs on variance of estimation after using waterfall WBS (work breakdown structures) -- 2 teams in audience with different estimates -- shows dramatic imprecision & inconsistency in estimation
Scrum sizing: principles:
• Entire team: different perspectives and experience, understanding, consensus, etc.
• Less than 4 hours
• Entire backlog of stories (3 to 6 months)
• Always size stories relative to other stories: amount of time (or effort) relative to other stories
• Use abstract measures (use Fibonacci-numbered story points)
• Projections leverage velocity
Different sizing options: Planning Poker vs. Two-Pass Release Sizing (aka Steve Bockman method)
2-Pass Release Sizing:
1) Order them smallest to largest
a) have backlog printed on cards -- 1st person reads story and places in middle of table, and then hands deck to second person
b) second person says, "I think this story is smaller than the first one," and gets consent from team
c) keep on going through rest of team that does work -- creates a "snake" of cards
While going through this exercise: a lot of learning around dependencies and other details in the story -- have a person in Jira recording insights that come up
Add ordinal numbers to cards in case a wind comes up :)
Do this with backlog refinement -- when adding new stories, bring out the card snake and do the same process again (but 10-15 minutes) to relatively size the new cards into the snake
Suggests estimating epics with story points as well: order them in the snake,
but top of backlog must be in stories so can be consumed in upcoming sprints.
Can be used with a remote team as well -- suggests Miro (formerly RealTimeBoard) and Mural as tools to put things on a board and zoom in/out.
Side effects
• Delivers shared project understanding
• Surfaces experience and expertise
• Engages the entire team
• Effective chartering exercise
2-Pass Release Sizing is faster than Poker and it guarantees stories are sized relative to each other
Reinforce adding new stories into the snake -- instead of saying a new story is a three, force them to put it into the snake to determine where it goes and go with that estimate
When combined with velocity, provides relative +-20% predictability.
â€"-thank you to Marty Stamps, GE Digital, reporter
No comments:
Post a Comment