Agile Estimation: What Empirical Process Control Has to Do with Onions
1. Agile Estimation
or
What Empirical Process Control Has to Do with Onions
Ilya Rosenfeld | Fall 2012
2. Welcome
"If you want a guarantee,
buy a toaster".
-- Clint Eastwood
"Entropy increases.
Period."
-- Physics
3. What's really the problem?
Toasters
○ inputs: well-defined
○ outputs: same
○ uncertainty in environment: low
○ planning: top-down 1962 2012
○ control process: repeatable, defined
○ "(mass) production"
Software
○ inputs: poorly defined
○ outputs: different
○ uncertainty in environment: high
○ planning: bottom-up 1962 2012
○ control "practices": empirical, adaptive
○ "evolution", "customization"
4. The Planning "Onion"
● estimation is integral to planning
● periodic to the core
● keeping it real estimation accuracy
inverse of the time frontier
5. "How long is it?"
How long does it take to paint
the house?
How long does it take to build a
piece of software?
6. "How big is it?"
Ok..How big is the house?
● rank by "size" (or complexity)
● calibrate "1"
● it's all relative
● uncertainty increases with size
● program for uncertainty
Story Points
● measure of SIZE (ex: 1+3+2+2+5=13)
● independent of time (or velocity)
7. What is "it", anyway?
informed by: comprised of:
business plan
market research PRODUCTS
strategic clients
operations ... EPICS, THEMES
STORIES
Agile project mgt FEATURES
TASKS
planning cycles stuff to estimate
8. Rinse and repeat.. recursively
● regularity over "process"
○ habits, ceremonies, rituals
● agree to disagree
○ negotiate the gaps away
○ converge
● estimating size
○ expert opinion, analogy, disaggregation
● closing the loop
○ validation, re-calibration
● the books
○ access
○ internal transparency
○ external transparency
9. Head Games
● Where in the Planning Onion are we?
● cross-functional-ness
● uncertainty and estimation scales
○ Fibonacci
○ T-Shirt Sizes
○ linear
○ ...
● which way to split and slice (stories)
○ "vertically": CRUD, MVC, data, ...
○ "horizontally": research, bootstrap, dev, test, doc, depl,
oversight/PM, ...
○ ...
● group poker
● solo poker
10. Versus
● estimation vs planning
● measurement vs estimation
● ideal days vs elapsed duration
● ideal time vs focused time
● story points vs ideal days
● estimates vs commitments
● projections vs forecasts
● estimates vs quotes
● velocity vs productivity
● pull vs push systems
"Prediction is very difficult, especially about the future."
--Niels Bohr
11. Stuff to Read
Good basics on roles, splitting and estmation: http://bit.ly/11PHk
"Agile Estimation and Planning" by Mike Cohn: http://amzn.to/2Bue0T
- free chapter on estimation: http://bitly.com/zzetUf
"Management 3.0" by Jurgen Appelo (big picture): http://amzn.to/V3RIq8
Planning Poker Cards: http://bitly.com/cAPmxN
Planning Poker Apps: http://bit.ly/Suhe5g