2. About The Speaker
Born and raised in the tropical
Costa Rica
Senior Cost Estimation Consultant
at Galorath Inc.
Works with customers around the
world to estimate Software,
Hardware and IT Projects
3. Outline
Background
Agile Estimates – Are they necessary?
Types of Estimates
Size Metric for Agile Project
Story Points or Function Points?
How about both?
Variables (Team Size, Sprint Duration, Velocity)
More on Velocity – Industry Standard Velocity
Sprint 0
Back-End Integration and UAT
The Complete Agile Estimate
5. Jump into the Time Machine… 2001
The Agile Manifesto was
conceived as collection of
loose, iterative, customer
centric development
methods…
- Agile -
… and the difficulty of estimating Agile began!
6. Agile Estimation Shouldn’t be More Difficult
It’s Just Another Methodology
ROM Detailed Estimate
Maximum
Uncertainty
Evolving
Complexity
7. Agile Estimates – Are they necessary?
#NoEstimates
Estimates are difficult to produce
Provide little to no value
Estimation is overhead and should be minimized
#Estimates
Organizations need to do budget planning
Estimates are needed to make informed decisions
CXOs need estimates for accountability to
shareholders
Social Networks have created discord around Agile
Estimates…
8. Parametric Modeling sustains the
#Estimates argument
Extensive Research on Agile
Estimation
Field Experiences Working
with Customers
Best Practices Proven to be
Meaningful and Beneficial
Sizing Methods
Processes
Training
Tools (SEER-SEM Agile
Planner)
9. Agile Estimates have Differing Fidelity
Backlog Estimate
ROM
One Estimate for the
Entire Project
Useful for High Level
Analysis and Budget
Planning
Can be built early with
limited information
Sprint Estimate
Detailed
Estimates are generated
at the Sprint Level
Useful for Team Planning
and Tracking
Considers all information
(Velocity, Team Size,
Sprint Duration )
10. Backlog Estimate = Big PictureFeatures
Project
Characteristics
Team
Dynamics
11. Sprint Estimate = Delivery DetailFeatures
Project
Characteristics
Team
Dynamics
12. Size Continues to be the
Main Driver
Pick a Metric and
Be Consistent!
Function Points
Use Cases
T-Shirt Sizes
13. Project Characteristics
What is it running on?
(Cloud, PC, Mobile, etc.)
What will it do? (CRM,
Database, Embedded)
What Agile Method?
(SCRUM is the most widely
used)
How will you build it? (New,
Modification, Rehost)
How much Governance?
(IEEE, ISO 9001, etc.)
@Galorath we call these
Knowledge Bases
14. Team Dynamics
Development
Team Size
• Optimal Size can
be calculated
• Everyone
available on day
1?
Backlog Size
• What is the team
tasked to
complete?
SPRINT Duration
• Timebox duration for an iteration
• 4 and 2 weeks are the most common
Team Velocity
• What is “doable”
in a SPRINT
• If not known, it
can be computed
(discussed in
slide 20)
15. What About Story Points?
Advantages
Easy to Use
No Training Required
Promotes cross-functional behavior (teams can compare similar
things)
Current Trends
People questioning their usefulness
Challenges
Not Standard/Team Specific
Difficult to Explain
Difficult (maybe impossible) to Benchmark
They are started to be abandoned
Story Points Inflation
16. Quest for Productivity Measurement
Induces Inflation
250
225
200
175
150
125
100
75
50
25
0
StoryPoints
Sprints
Project Monitoring Begins
This is what happens when you don’t have a standard
sizing metric, like Function Points
17. Story Points Must be Tempered by a
Standard Metric – Galorath Approach
Function Points (or SLOC) used as the Underpinning
Metric
Team Variance in Story Points must be accommodated
A FP Counter must understand the Agile Team
“Standards” to normalize the count
Final Estimate will reflect the Team Specific Productivity
19. Same Point Counts - Different Results
Same Story Point Counts
Different Teams
Different Estimates
715 FPs310 FPs
Team A Team B
20. Agile is all about Velocity
Sprint X
Staff = 5 People Duration = 4 Weeks
Effort = 800 Hours Cost = $120K
So, what is the Product Owner getting?
21. Given a Velocity… what can we deliver
for any given project type?
Velocity = Commitments for the Sprint
Velocity is a metric of Software Size/Sprint
It is not magic, it can be predicted
Industry Data provides a predictive model for productivity
20
FPs/Sprin
t
87
FPs/Sprin
t
We call this “Industry Standard Velocity”
Productivity
22. Computing the Number of Sprints
The number of Sprints drives the estimated Cost and
Duration
Backlog Size = 540 SPs
Velocity = 90 FPs/Sprint
Voila, We have an Estimate!!!
23.
24. Industry calls this “Sprint 0”
The work before the Project
can start
For example: Planning and
Setup
We believe Sprint 0 should be
estimated
Duration/Effort may be
More/Less Than a Typical
Sprint
SPRINT 0
Env.
Setup
Team
Setup
Backlog
Plan
25. Back-End Work
Typically, some work needs to be done after you are
done “Sprinting”
Final System Level Integration
User Acceptance Testing
Final Check-out and Certification
Warranty/Maintenance
The Estimate would NOT be Complete if this is
Ignored
Should be Estimated Based on the Overall Backlog
Size
26. The Complete Agile Estimate
Full Estimate Including:
Pre-Sprint Work (Backlog
Planning, team Setup, Env.
Setup)
Sprint Iterations
(Design/Code/Test)
Back-End Work (SIT, UAT, etc.)
27. Our Journey Continues…
At Galorath we have performed extensive research
and worked collaboratively with an international
customer base to understand their estimation needs
in an Agile Environment
All knowledge has been incorporated into the new
Agile Planner capability in SEER-SEM
The journey is just starting… We will continue to
make and offer improvements based upon our on-
going research, staying abreast of industry trends
and listening to the voice of the customer