1. The document discusses various techniques for estimating projects, including estimating size and converting it to effort. It covers estimating individual tasks and story points, and using statistics to quantify uncertainty.
2. Key aspects of estimation include studying the requirement, individually estimating size, challenging each other's assumptions, and deciding on a consensus number. Estimates should not be averaged and shortcuts should be avoided.
3. Proper planning involves defining tasks, estimating individual tasks, developing a schedule considering availability and dependencies, and ensuring accomplishments are made through each sprint to deliver value.
Professional Resume Template for Software Developers
About estimating and planning
1. Bart Vanderbeke -- the art of estimating and planning
violin maker: Thomas Meuwissen https://www.thomasmeuwissen.com/
Photo: Rudi Van Beek (http://www.rudivanbeek.be)
Craftsmanship:
The Art of Estimating
and Planning
Bart Vanderbeke
3. Bart Vanderbeke -- the art of estimating and planning
• Choosing the optimal tool…
• Implies you know more than one tool (*)
• Having a well-filled toolbox…
• Makes you more valuable for your employer
• Makes you more employable
• Use your employability in another company as trigger, crowbar to keep
on learning
(*) Tool: method, toolchain, design approach, language,…
What’s in your toolbox?
Law of the Instrument
koevoet
4. Bart Vanderbeke -- the art of estimating and planning
Working in the 21st century
6. Bart Vanderbeke -- the art of estimating and planning
Estimation process
Study Estimate
Challenge
Study Estimate
insight
insight
number
number
Personal
insight
Personal
insight
Collective
Insight
Requirements
Assumptions
Fact
or
Fiction
fact
fiction
THE
Number
’12’
’19’
Decide
7. Bart Vanderbeke -- the art of estimating and planning
Estimating fig
Estimating txt
Size and effort
Complexity+Size
T- shirt size
Scrum
Story Points
Statistics
Goldratt
No Shortcuts
Mindmap
Summary
1. Study
2. Estimate:
• Stick a size number to it
• Individually or in team
3. Challenge:
• Why did Jim come up with a different number than George?
• Surface and eradicate assumptions
4. Decide:
• Converge to a consensus on a number
• Never average
• Never come to a compromise
5. If size was estimated, convert to effort
Elements of Size Estimation
8. Bart Vanderbeke -- the art of estimating and planning
Define the relation between size and effort
estimate “Size” convert Effort
calibration Effort = function(size)
Estimating fig
Estimating txt
Size and effort
Complexity+Size
T- shirt size
Story Points
Statistics
Goldratt
No Shortcuts
Mindmap
Summary
Historical Info
detailed case
analysis
9. Bart Vanderbeke -- the art of estimating and planning
More elaborate ‘size’
size
complexity
How big?
How difficult?
Feature A
Feature B
Effort = function(size, complexity)
Estimating fig
Estimating txt
Size and effort
Complexity+Size
T- shirt size
Story Points
Statistics
Goldratt
No Shortcuts
Mindmap
Summary
10. Bart Vanderbeke -- the art of estimating and planning
Estimating fig
Estimating txt
Size and effort
Complexity+Size
T- shirt size
Scrum
Story Points
Statistics
Goldratt
No Shortcuts
Mindmap
Summary
T shirt size estimation
size
complexity
How big?
How difficult?
Effort = function(size, complexity)
L
S
XL
M
Calibration:
1. Take representative finished project
2. Assign T-shirt size to chunks
3. Calibrate with actual effort
T shirt sizing is an example of ‘thinking in analogies’ as described by David Epstein’s in his book “RANGE”
11. Bart Vanderbeke -- the art of estimating and planning
Scrum
estimate Story
Points
convert person
days
Select some use
cases from the
project and
estimate in more
detail
calibration Effort = function(Story Points)
Calibration:
1. Take representative use case(s)
from project at hand
2. Estimate in most detail
add detail until effort-estimation is possible
3. Calibrate with estimated SP
Estimating fig
Estimating txt
Size and effort
Complexity+Size
T- shirt size
Scrum
Story Points
Statistics
Goldratt
No Shortcuts
Mindmap
Summary
12. Bart Vanderbeke -- the art of estimating and planning
Why (quasi) Fibonacci?
0
20
40
60
80
100
120
1 2 3 4 5 6 7 8 9 10
Story Points
0
2
4
6
8
10
12
14
1 2 3 4 5 6 7
Story Points
Because you would never doubt
between consecutive values…
Estimating fig
Estimating txt
Size and effort
Complexity+Size
T- shirt size
Scrum
Story Points
Statistics
Goldratt
No Shortcuts
Mindmap
Summary
13. Bart Vanderbeke -- the art of estimating and planning
Estimation and statistics
• Estimate three values for each task:
• M
• E x = µ =
𝑤+4𝑡+𝑏
6
• σ x =
(𝑤−𝑏)
6
• Formalizes
‘risk buffers’, ‘uncertainty’, ‘confidence’ https://en.wikipedia.org/wiki/PERT_distribution
Value description
best Very improbable it will take less than
typical When asked only one value….
worst Very improbable it will take longer
than
PERT distribution (β-distribution)
typical shape
14. Bart Vanderbeke -- the art of estimating and planning
Goldratt
• Make your overall schedule
for a ‘safe duration’
90% probability of success
• Assign duration with 50% prob.
to tasks
• Place a checkpoint at 50%
• Only grant the extra time when
really needed
• Parkinson’s law:
work expands to fill the allotted
time
http://www.pmknowledgecenter.com/node/267
impossible possible safe
task
default
project
default
Only when needed
task
worst case
Estimating fig
Estimating txt
Size and effort
Complexity+Size
T- shirt size
Scrum
Story Points
Statistics
Goldratt
No Shortcuts
Mindmap
Summary
15. Bart Vanderbeke -- the art of estimating and planning
We do not take shortcuts
• If the project ‘becomes better by dropping some practice’
then we should always drop it
• We always do what we think is necessary and sufficient
• We always aim for minimal ‘waste’ in our way of working
• Asking for a faster delivery, may challenge us to organize ourselves better
• The effect of taking a shortcut:
• A shortcut increases the probability to finish earlier, but also to finish later
optimal project approach
taking shortcuts
𝑃𝑟𝑜𝑏 𝑜𝑓 𝑓𝑖𝑛𝑖𝑠ℎ𝑖𝑛𝑔 𝑎𝑡 𝑡𝑖𝑚𝑒 𝑡
𝑡𝑖𝑚𝑒 𝑡
Estimating fig
Estimating txt
Size and effort
Complexity+Size
T- shirt size
Scrum
Story Points
Statistics
Goldratt
No Shortcuts
Mindmap
Summary
16. Bart Vanderbeke -- the art of estimating and planning
What’s in your toolbox?
Estimating fig
Estimating txt
Size and effort
Complexity+Size
T- shirt size
Scrum
Story Points
Statistics
Goldratt
No Shortcuts
Mindmap
Summary
17. Bart Vanderbeke -- the art of estimating and planning
Summary
!
Estimating using the basic elements
is part of your craftsmanship
It must not be compromised
study,estimate,challenge,decide,3 values
18. Bart Vanderbeke -- the art of estimating and planning
Overview
Method Study estimate values Size/effort challenge decide calibration
Expert Judgment individual individual continuous both no expert memory
Wideband Delphi individual individual continuous both team team consensus history
Scrum Poker team
individual
(cards) discrete
Size
(SP) team team consensus
selected
tasks/UC
𝑐𝑜𝑛𝑠𝑒𝑛𝑠𝑢𝑠 ≠ 𝑐𝑜𝑚𝑝𝑟𝑜𝑚𝑖𝑠𝑒
𝑐𝑜𝑛𝑠𝑒𝑛𝑠𝑢𝑠 = 𝑒𝑣𝑒𝑟𝑦𝑏𝑜𝑑𝑦 𝑎𝑔𝑟𝑒𝑒𝑠 𝑎𝑏𝑜𝑢𝑡 𝑡ℎ𝑒 𝑜𝑝𝑡𝑖𝑚𝑎𝑙 𝑐ℎ𝑜𝑖𝑐𝑒
19. Bart Vanderbeke -- the art of estimating and planning
• Conclusion is always consensus
• Never average estimations
• ‘Challenging’ is used to surface assumptions
• convert them in facts or kill them
• Challenging helps creating a shared context in the team
• Use the three-value estimation
• Indicates confidence and helps converging
• Helps in translation to lead time
• No shortcuts
Attention points
21. Bart Vanderbeke -- the art of estimating and planning
1. Only the ‘pure coding’ or
2. (Test) design + Coding + unit testing + bug fixing + handling review
input
• Typically: * 1.6
• Make sure you agree well on what you are estimating
What’s in a name? ‘Task’
What’s in a task?
Estimate → Plan
Accomp. & Success
Value
Can’t make it?
What first?
Intermediate stock
22. Bart Vanderbeke -- the art of estimating and planning
• 𝑡𝑜𝑡𝑎𝑙 𝑒𝑓𝑓𝑜𝑟𝑡 = σ 𝑡𝑎𝑠𝑘𝑖 + σ σ𝑖
2
∗ 1.6
• Lead time follows from work distribution
• Using effective availability and dependencies
• adding code freeze
• 4 weeks
• 𝐿𝑇 >
σ 𝑡𝑎𝑠𝑘𝑖+ σ σ𝑖
2
∗1.6
σ
𝑎𝑣𝑎𝑖𝑙%𝑖
100
+ 4𝑤𝑘 ′ >′ 𝑑𝑢𝑒 𝑡𝑜 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑐𝑖𝑒𝑠
From Estimation to Plan
What’s in a task?
Estimate → Plan
Accomp. & Success
Value
Can’t make it?
What first?
Intermediate stock
23. Bart Vanderbeke -- the art of estimating and planning
• Task < week
• An ‘atomic’ task must fit in a week
• Go into the weekend with a feeling of accomplishment
• Sprint
• 2…3 weeks
• Every two weeks a collective feeling of success
Task / Sprint
https://www.malotaux.eu/doc.php?id=21
What’s in a task?
Estimate → Plan
Accomp. & Success
Value
Can’t make it?
What first?
Intermediate stock
24. Bart Vanderbeke -- the art of estimating and planning
What should I define as sprint-goals?
creating value for the team members is a fair sprint goal
create value for stakeholders
What’s in a task?
Estimate → Plan
Accomp. & Success
Value
Can’t make it?
What first?
Intermediate stock
25. Bart Vanderbeke -- the art of estimating and planning
What if we never make it in the sprints?
If you consistently cannot make it
1. Plan less
2. Find the reason why you cannot make it
What’s in a task?
Estimate → Plan
Accomp. & Success
Value
Can’t make it?
What first?
Intermediate stock
26. Bart Vanderbeke -- the art of estimating and planning
What to do first?
What’s in a task?
Estimate → Plan
Accomp. & Success
Value
Can’t make it?
What first?
Intermediate stock
27. Bart Vanderbeke -- the art of estimating and planning
• Production: Avoid intermediate stock
• Software:
• Avoid results waiting to be used
• Software: avoid doing things in parallel just to keep everybody working
• Minimize the work in progress
• Better one thing finished than everything on its way
• Focus of the team
• (Kanban/Lean)
What should I start?
What’s in a task?
Estimate → Plan
Accomp. & Success
Value
Can’t make it?
What first?
Intermediate stock
28. Bart Vanderbeke -- the art of estimating and planning
Conclusion
• Be very conscious what is in (the estimation of) a task
• Estimate → plan 𝐿𝑇 >
σ 𝑡𝑎𝑠𝑘𝑖+ σ σ𝑖
2
∗1.6
σ 𝑚𝑒𝑚𝑏𝑒𝑟𝑖∗
𝑎𝑣𝑎𝑖𝑙%𝑖
100
+ 4𝑤𝑘 ′ >′
𝑑𝑢𝑒 𝑡𝑜 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑐𝑖𝑒𝑠
• Task/Spring: Create your feeling of success
• Value can be for any stakeholder
• Cannot make it? Step 1: stuff less in sprints
• What first?
• Beware of starting too many things
What’s in a task?
Estimate → Plan
Accomp. & Success
Value
Can’t make it?
What first?
Intermediate stock