a family of scampi appraisal methods · pdf filedriven by appraisal scope, organizational...
TRANSCRIPT
Pittsburgh, PA 15213-3890
CMMISM
Sponsored by the U.S. Department of Defense© 2003 by Carnegie Mellon University
SM CMMI, CMM Integration, and SCAMPI are service marks of Carnegie Mellon University® Capability Maturity Model and CMM are registered with the U.S. Patent and Trademark Office
Pittsburgh, PA 15213-3890
CMMI
This material is approved for public release. Distribution is limited by theSoftware Engineering Institute (SEISM) to attendees.
Sponsored by the U.S. Department of Defense© 2003 by Carnegie Mellon University
SM CMM Integration, SCAMPI, SCAMPI Lead Assessor, SCAMPI Lead Appraiser, and SEI are service marksof Carnegie Mellon University. CMM and CMMI are registered in the U.S. Patent and Trademark Office by Carnegie Mellon University.
A Family of SCAMPI AppraisalMethods
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 2
ObjectivesIntroduce the three classes ofSCAMPI appraisal methods,different appraisal needs andtypes, and the SCAMPI B & CProject.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 3
Three Classes of Appraisals
LargeMediumSmallTeam Size
HighMediumLowResource needs
YesNoNoRatings generated
HighMediumLowAmount of objective evidence
Class AClass BClass CCharacteristic
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 4
Appraisal Needs
Getting Started, Understanding the CMMI
Familiarization or Expectation Setting for Future Appraisal
Broad-based Information on Implementation of Practices
In-depth Analysis of Implemented Practices
Gap Analysis Focused on Findings of Previous Appraisal
Incremental Data to Track Improvement Progress
Benchmarking Leading to a Public Statement
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 5
Appraisal Types
Getting-Started Intervention
*Mini Appraisal
*Gap Analysis
*Improvement Monitoring
Delta Appraisal
Incremental Appraisal
Benchmarking Appraisal
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 6
Getting-Started InterventionPurpose: Provide information toan organization learning aboutCMMI.
Description: Expert-directed,presentation & discussion-intensive, broad scope, shallowcoverage method thatemphasizes participation ofaffected groups in facilitatedsessions - with no focus onsufficiency of coverage.
Rough Effort Estimate: One totwo experts on site for 1 to 5days. Planning and coordinationeffort driven by appraisal scope.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 7
Mini AppraisalPurpose: An abbreviated versionof a benchmarking appraisalused for experience and non-rating related outputs.Description: Team-based, dataintensive, narrow scope, deepcoverage method thatemphasizes the use of ObjectiveEvidence and interviews - withreduced focus on datasufficiency.Rough Effort Estimate: Oneexpert on site for 3 to 5 days.Planning and coordination effortdriven by appraisal scope, andexperience level of local teammembers.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 8
Gap AnalysisPurpose: Identify specificdeficiencies in implementedpractices relative to practiceswhich would supportachievement of the goals - for aspecified scope.
Description: Expert-based,interview-intensive, broad scope,shallow coverage method thatemphasizes identification ofweaknesses - with more limitedfocus on sufficiency of coverage.
Rough Effort Estimate: One totwo experts on site for 1 to 3days. Planning and coordinationeffort driven by appraisal scope.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 9
Improvement MonitoringPurpose: Track implementationof process improvement actionsin an organizational unit that haspreviously undergone abenchmarking appraisal.Description: Expert-based,interview & document-intensive,narrow scope, deep coveragemethod that emphasizesidentifying status of changes inprocesses - with more limitedfocus on sufficiency of coverage.
Rough Effort Estimate: One totwo experts on site for 1 to 3days. Planning and coordinationeffort driven by scope andoutcomes of previousappraisal(s).
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 10
Delta AppraisalPurpose: Confirm correctiveactions resulting from abenchmarking appraisal.
Description: Expert-based,interview & document-intensive,very narrow scope, deepcoverage method thatemphasizes confirming specificchanges in processes - withlimited focus on sufficiency ofcoverage for the model.
Rough Effort Estimate: One totwo experts on site for 1 or 2days. Planning and coordinationeffort driven by the nature andmagnitude of weaknessesidentified in the benchmarkingappraisal.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 11
Incremental AppraisalPurpose: Building a profile ofresults across a broad modeland/or organizational scope.
Description: Expert-directed orteam-based, data intensive,narrow scope, deep coveragemethod that emphasizesuse of objective evidence - withstrong focus on sufficiency ofcoverage (in narrowincrements).
Rough Effort Estimate:Appraisal events distributedover time. One expert on site for3 to 8 days. Planning andcoordination effort driven byappraisal scope, and experiencelevel of team members.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 12
Benchmarking AppraisalPurpose: Document a frequentlysought, high level, benchmarkfor an organizational unit.Description: Expert-directed andteam-based, data intensive,broad scope, deep coveragemethod that emphasizes use ofobjective evidence - with strongfocus on sufficiency of coverage.
Rough Effort Estimate: Oneexpert on site for 5 to 15 days.Planning and coordination effortdriven by appraisal scope,organizational maturity, andexperience level of teammembers.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 13
Sources of Objective EvidenceInstruments
• Organizational assets reflecting evidence ofimplementation of model practices (e.g., mappingtables)
• QuestionnairesInterviews
• Standard structured interviews; on-call interviews;follow-up interviews
• Exploratory or focused questions targeted at manager,practitioners and/or users
Presentations• Briefings, demonstrations
Documents• Hardcopy, softcopy, hyperlinks
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 14
Appraisal Scope
Defining the“Organizational Unit”
Selecting CMMI Model,Representation, and Scope
CMMI-SE/SW
2345 PA1
PA2PA3PA4PA5PA6
0 1 2 3
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 15
Broad Scope, Shallow CoverageTake advantage of low cost data collection methods• achieve broader coverage of the organization
- include all projects, not just 4 for benchmarking• address a larger portion of the model
- match information sources to information needs
Limitations of low cost data collection methods• very limited opportunity to explain nuances
- yes/no questions in interviews don’t pay off• difficult to assure validity of data
- ambiguous questionnaire items don’t help
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 16
Narrow Scope, Deep CoverageMaximize insight with a small sample of the organization• detailed insight about a selected unit in the organization
- potentially include every staff member• full coverage of limited model material
- minimal risk of undiscovered weaknesses
Limitations of using a small sample• limited ability to make generalizations
- process of one stellar project may be misleading• limited coverage of the lifecycle in use
- implementation of future phases may differ fromintended practices
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 17
Required Outputs: SCAMPI B&CCompleted AppraisalDisclosure Statement (ADS)
Identification of data collectiontechniques used
Identification of data collectionsessions conducted
Appraisal plan, annotated withactual data for• time, effort and cost• model coverage• organizational coverage
Official
J. Hancock
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 18
Expected Outputs: SCAMPI B&CStatements of strengthsand/or weaknesses relativeto model practices, goals orother components.
Characterizations ofdifferences in practiceimplementation• across model content
covered in the appraisal• across organizational
units included in theappraisal
Your Results
0
50
100
• Findings statements• Strengths/Weaknesses• Recommendations• ...
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 19
Optional Outputs: SCAMPI B&CDetailed data summariesrelated to a pre-existing:• process improvement target• performance contract• remedial improvement action• set of risk areas in a contract
Orientation and/or training for:• improvement champions• technical and managerial staff• sponsors new to CMMI
A ‘watershed’ event for a groupof co-workers.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 20
Appraisal Reporting TechniquesFormal presentations using“slides” - words and graphics.
Facilitated, interactive,presentations with specificparticipants.
Group or individual meetingsinvolving appraisers andparticipants.
Written material, ranging from‘bulleted’ summaries inpresentation slide format to afull detailed report.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 21
Appraisal Use/Purpose: Examples
Formulating OurWinning Strategy
Fixing ImmediateProblems
Exploring Goals forImprovement &Training Staff
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 22
Appraisal Outcomes: Examples -1
Making Sure OurHouse is in Order
Charting ProgressAgainst Our Plan
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 23
Appraisal Outcomes: Examples -2
Find NewChallengesto Conquer
Recognize Progressand Celebrate Success
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 24
The SCAMPI B&C ProjectSEI working collaboratively with appraisal stakeholders:• prototyping and field trials with collaborator organizations• communication with stakeholders about the project• periodic dissemination of draft material for review• publication/presentation of project results
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 25
SCAMPI B&C Project ObjectivesDevelop, Field Test, and Deploy an integrated suite ofappraisal methods that are:• Developed iteratively with extensive field tests• Based on requirements from users• Integrated and upwardly compatible• Well specified and documented
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 26
Project Goals1. Jointly Develop proven B&C appraisal methods2. Transition B&C competencies to the participating
organizations3. Advance the (SCAMPI) authorization of current and
future lead appraisers4. Deliver value to the participating organizations’
improvement efforts through the delivery of pilotappraisals
5. Jointly produce: technical reports, papers,presentations
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 27
Appraisal Usage ModesInternal process improvement
Acquisition
• Source selection
• Contract/Process monitoring
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 28
Options for the Pilot Appraisals• Appraisal types and timing sequence• Definition of the OU scope (what parts of the organization
are included in the appraisal)• Definition of the reference model scope• Findings
- At the OU level- Discipline specific- Project specific
• These options drove the derivation of the pilot use casesand the pilot enactments.
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 29
Estimated ResourcesTeam Members• Class B (2-9) 3-8 days on-site; 1-2 days training• Class C (1-6) 1-5 days on-site; 0-1 days training
Appraisal participants• 1-5 hours/participant, including (as applicable)
- Attending the opening meeting- Completing a questionnaire or survey- Participating in an interview session- Participating in a validation session- Attending the findings presentation
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 30
SEI’s RoleSteward of the family of SCAMPI appraisal methodsLead developer of SCAMPI B and C methodsDefiner of pilot data to be collected and data analyses tobe performedLeader/Coach of pilot appraisals
CMMI
© 2003 by Carnegie Mellon University Module N SLAT V1.1 page 31
ConclusionNot all appraisal needsmake sense using theSCAMPI A method.
The SEI is developingSCAMPI B & C methods.