Upload
qai
View
182
Download
1
Tags:
Embed Size (px)
DESCRIPTION
Different Flavors Of PPMs -S.Sugavaneswaran Sonata Software Ltd.presented at1st International Colloquium on CMMI High Maturity Best Practices held on May 21, 2010, organized by QAI
Citation preview
Different Flavors Of PPMs
-S.Sugavaneswaran
Sonata Software Ltd.
Different Flavors of PPMs
Presented at HMBP 2010
S.Sugavaneswaran
Sonata Software Limited
21-May-10
www.sonata-software.com
3
Agenda
• About Process Performance Models
• High maturity enablers
• Challenges faced in implementation
• Flavors of PPM
• How good they are
4
Need for PPM
• An Earned Value Management dashboard
• How effective is such a report in terms of triggering
process improvement actions?
• Will it help to know which controllable process
factors influence the above outcomes?
Adapted
from the SEI
paper “An
Executive
Tutorial of
CMMI
Process
Performance
Models”
“Delighting customers is what it’s all about, and that comes from
consistent, end-to-end process performance.” – Kevin Weiss
• Relate controllable factors to an outcome
o Y=f(x1,x2,x3…)
• Developed from historical data
• Predict results achieved by following a process
• With a known confidence level
• Help perform “What-if” analysis
• Compose processes for a project
5
Process Performance Models
• IT Consulting and Services company
• Customers across US, Europe, Middle East and APAC
• Services offered
• Product Engineering Services
• Application Development/ Management
• Managed Testing
• Infrastructure Management
• Quality standards adaptation
• ISO 9001
• CMM Level 5
• CMMI v1.2 Level 3
• ISO 20000-1
6
Our Context
7
High Maturity Enablers
• Standardizing size measures for projects
• To normalize process performance
• Enabling sub-process level control
• Effort to create, review and rework
• Options for each sub-process
• Data at the sub-process option level
• Capturing defect injection and detection
8
Implementation Challenges
“The truth is that you always know the right thing to do. The
hard part is really doing it.” – H. Norman Schwarzkopf
• Stakeholder buy-in
• Issues with data availability / stability
• Tool enablement constraints
• Continued involvement of practitioners
9
PPM – Healthy Ingredients
1. Statistical or probabilistic in nature
2. Predict interim and/or final project outcomes
3. Use controllable factors tied to sub-processes
4. Model the variation of predictive factors to forecast
outcome variations
5. “What-if” analysis for project planning/re-planning
6. Connect upstream with downstream activities
7. Enable mid-course corrections
10
PPM Flavors
“All models are wrong, some are useful!” – George Box
• Development project – Continuous simulation
• Sub-process wise process performance
• Prediction with confidence levels
• “What-if” analysis
• Production support – Discrete event simulation
• Process flow depiction and simulation
• Analysis of
• SLA adherence
• Resource utilization
11
Flavor 1
• About the project
• New development (Agile)
• Sprints & stories
• Sprint content decided based on experience
• Developers categorized by skill level
• Model applied
• Monte Carlo Simulation
12
Simulation Highlights
• Objective: To optimize number of stories forming
part of a sprint
• Predictive factors
• Working hours per day
• Number of stories
• Sub-process wise productivity
• Skill levels
• Size of each story
• Team size
13
The Model
• Inputs: Estimated story size and sub-process
productivity distributions
• In each simulation run,
• The model chooses values from sub-process
productivity distributions, arrives at effort
• Predicted effort = Sum of all sub-process efforts
• Effort computed is divided by the available man-hours
per day, giving the elapsed days
• Over time, a profile is built showing the distribution
of likely outcomes (number of days)
• Confidence level indicated for the output
Scenarios
Story
1
Story
2
Story
3
Story
4
Story
5
Story
6
Story
7
Story
8
Story
9
Story
10
Size 30 12 80 2 6
Skill High High High Low High
Understanding &
Analysis0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Design 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Design Review 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Coding 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Code Review 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Code Fix 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Unit Test 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Units Test Fix 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
FIT Testing 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
IT Fix 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
14
Table 1: Model before running the simulation
15
Sample Predictions
Tool Output 1: Release Prediction – 6 Stories
Tool Output 2: Release Prediction – 5 Stories
Scenario 1: Stories
1,2,3,4,5 and 6
Scenario 2: Stories
1,2,3,5 and 6
16
Process Control
Tool Output 3: Sensitivity-Release Prediction
Sub-processes to be
closely monitored:
IT and Coding- High
skill
17
Flavor 2
• About the project
• Production Support
• High volume, short turnaround work
• SLA-driven
• Different ticket priorities
• Three different skill sets
• Model applied
• Discrete Event Simulation
18
Simulation Highlights
• Objectives: To forecast and manage SLA
adherence and Resource utilization
• Predictive factors
• Team size
• Response, analysis and development time
• Arrival pattern of tickets (by priority)
• Wait times
19
The Process Model
SLA Adherence
Tool Output 4: SLA Miss before the model
Tool Output 5: SLA Miss after the model
Probability of SLA breach brought down
21
Resource Utilization
Tool Output 6: Before
the model
Tool Output 7: After
the model
Resource utilization improved as well
Model Flavors vs Healthy
Ingredients
Ingredient Flavor 1 Flavor 2
Statistical, probabilistic… Yes Yes
Predict interim/ final… Yes Yes
Sub-process level factors Yes Yes
Model uncertainty… Yes Yes
Support “What-if”
analysisYes Yes
Connect to downstream.. Yes Yes
Enable course correction Yes Yes
22Table 2: Models vs Healthy Ingredients
23
Conclusion
“Action may not always bring happiness, but
there is no happiness without action.”
- Benjamin Disraeli
Click here for:
High Maturity best practices
HMBP 2010 Presentations
organized by QAI
Click here