Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Agile estimationPrinciples
Cone of UncertaintyV
aria
bilit
y of
Est
imat
es
Time to CompletionDONE
Where is Irene going?
Where is Irene going?
Where is Irene going?
Precision And Accuracy• Your mind is not a precise
time-estimating tool• Your mind can be an accurate
time-estimating tool• Trade precision for accuracy
to succeed
Point of Diminishing Returns
Time
Acc
urac
y of
Est
imat
es
Stop here
Relative estimation
It’s easier to estimate unknown things by comparing them
relatively to things you know
Discussion is the Key• Having group discussions flushes out• Unknown facts• Different points of view• Common understandings
• All opinions are valuable• What happens if only the expert estimates?
An Abstraction: T-Shirt Sizes• S, M, L, XL• Maybe XS and XXL? Be careful of the trap of perceived
precision…• Note: there are no sizes in between major divisions
10 minWall Estimation Technique
• Create a board with 4 T-shirt sizes as columns (S, M, L, XL)• Write the following objects on individual sticky notes,
and put them in appropriate columns based on size• Asia• A 2007 Toyota Prius• The White House• Arkansas• 1977 Cadillac Eldorado• New Mexico
• The Pentagon• All of the Great Lakes• Australia• A one-car garage• Indian Ocean• A Junior High baseball field
An Abstraction: Story Points• A Story Point is a very commonly used abstraction for
agile estimation• On what is a “Story Point” based?• Complexity• Risk• Unknowns• Effort• Cost• All of the above!
Story Points• Using the Fibonacci sequence is common • (0, 1, 2, 3, 5, 8, 13, 21…)• This represents the Cone of Uncertainty• The numbers represent classes of size, no units in between
Comparing AbstractionsAbstraction Classifications UnitT-Shirts • S, M, L, XL Size
Story Points • Typically numeric, in a FibonacciSequence
• i.e., 0, 1, 2, 3, 5, 8, 13, 21
“Story Point” – AKA, some combination of • Complexity• Risk• Unknowns• Effort• Cost
Abstraction advantages• Abstractions prevent unintentional time commitments• Keeps us in a relative estimation state-of-mind• Provides protection against Parkinson’s Law• Allows us to be empirical using new information as it is
available• Saves us from endless re-estimating
Solving for Time later…• Most questions around estimation are questions about
time to complete• We can answer the real question without having to
estimate directly in time units!• Let’s solve for time last
Story Points != Unit of Time• Distribution of time taken to complete a 5 point story
Time
Story Points != Unit of Time
Time
5 8
Story Point Anti-Patterns• Solving for time up front• It’s unsafe to normalize story points (bad SAFe joke)• Story points are intentionally unit-less and relative to each
other• Taking a number in-between• Putting new numbers in between the sequence adds false
precision.
Planning Poker• Another estimation technique• Based on difficulty for the team, not for individuals• Good for diving deeper into PBIs
• The development team doing the work estimates• PO should be available for questions
Technique: Planning Poker• Discuss the story briefly• Each person on the Dev Team selects a
card • Team reveals their selections all at
once• Outliers explain their selection to the
group• Repeat steps 2-4 until (near) consensus• Take the closest number, not the
average
3
2
3 85
Get Ready to go to the Beach
Billy
Johnny
Suzie Carol
Dave
5
2
5 13
8
Get Ready to go to the Beach
Billy
Johnny
Suzie Carol
Dave
8
21
8 88
Get Ready to go to the Beach
Billy
Johnny
Suzie Carol
Dave
Anti-Patterns• Don’t do these:• Estimation differs based on who might
work on it• Rushing to lay down cards before
everyone is ready• One person estimates out loud,
everyone lazily agrees • Little conversation occurs• Estimation is only done with the team
lead and product owner
Anchoring• The act of influencing opinions before a decision is
made• TIP: You can use this to your advantage!• Salary Negotiations:
“Substantial psychological research suggests that, more often than not, negotiators who make first offers come out ahead” - Adam D. Galinsky (Harvard Business School)
Anchored estimates
Avoided discussions
Lazy estimates
5 minMake Your Deck Now
• Create a deck with the following cards:• 0, 1, 2, 3, 5, 8, 13, 20, 40, 100• Do not include: ?, ½, ∞
10 minEstimate this work
Estimate the following using the planning poker technique – use: 0, 1, 2, 3, 5, 8, 13, 20, 40, 100• Plant a tree• Paint the house• Clean the garage• Get ready to go to work• Replace the engine in your car