the productivity pitfalls of process improvement

The Productivity Pitfalls ofProcess Improvement

Scott GoldfarbQ/P Management Group, Inc.10 Bow StreetStoneham, Massachusetts 02180

Email: [email protected]

Tel: (781) 438-2692FAX (781) 438-5549www.qpmg.com

2© Copyright 2006. Q/P Management Group, Inc.

All Rights Reserved.

Agenda

• Background

• Diagnosing Process Productivity Problems

• Avoiding Pitfalls and Finding Solutions



The Process Productivity Problem

The quest for achieving higher levels of process maturity can result in lower levels of productivity

• Lack of a good productivity baseline disguises the symptoms

• Process improvement overhead can be high

• Newly implemented processes are not always effective

• Equilibrium between process and productivity is difficult to find

• The tradeoffs between quality and productivity are not always well understood



• “Quality is Free”

• Process Improvement Leads to Productivity Improvement.

• If you can measure it, you can manage it!

Three Misconceptions



• “Quality is Free” … until you hit the point of diminishing returns.

• Process Improvement leads to Productivity Improvement with the proper processes and focus.

• If you can measure, analyze and conclude the right things, you can manage it!

Three Corrected Misconceptions



Anticipated Gains

+

Productivity Impact

-

1970’s 1980’s 1990’s 2000’s

OT

Tool Focus

Assess/ Measure

Process Improvement I

Technology Focus

Process Improvement II,

Outsourcing,Offshore,

Y2K

Process Overkill, Quality Focus

MeasureProcess

Streamlining

Industry Trends Impacting Productivity, Quality and Cost Were Considered



Agenda

• Background





• Compare to industry benchmarks

• Compare internal productivity and quality trends and needs

• Diagnosing “Surface Metrics” can help identify the symptoms– Quality – Defects/Function Point– Productivity – Function Points/Hour– Schedule – Duration versus standard or trend– Cost – Cost/Function Point– Process – CMM(I) Assessment

• Tradeoffs between “Surface Metrics” maybe the cause of low productivity

• A “Deep Dive Diagnosis” is often required to uncover specific process productivity problems

Diagnosing the Process Productivity Problem



Diagnose Surface Metrics – CMMI Process Profile

Project PlanningProject Tracking & MonitoringConfiguration ManagementQuality AssuranceRequirements ManagementSubcontract Management

Organization Process FocusOrganization Process DefinitionTraining ProgramIntegrated Software ManagementSoftware Product EngineeringInter-group CoordinationPeer Reviews

Quantitative Process ManagementSoftware Quality Management

Defect Prevention

Process Change ManagementTechnology Change Management

Level 2

Level 3

Level 4

Level 5

Not satisfiedPartially satisfiedFully satisfiedNot applicable

Legend

Symptom: Out of balance processes -Quality and Management Processes satisfied at higher levels



Diagnose Surface Metrics - Productivity versus Quality

Productivity / Quality Comparison

Symptom: Better than Best in Class quality and >Level 3 quality related processes combined with low productivity

Poor Average Best in Class

Quality (Defect/FP)

Average

High Goal

Potential Symptom

Poor

Productivity(FP/Hour)



Diagnose Surface Metrics - Schedule Tradeoff

Productivity versus ScheduleHigh

Productivity

LowCompressed Optimum Extended

Make sure severely compressed or extended schedules are not the real problem



Diagnose Surface Metrics - Schedule Analysis

Schedule Duration by Project Size Category

0

100

200

300

400

500

<50 51-100 101-200 201-500 501-1000Project Size in Function Points

Day

s

Expected Schedule

Symptom: Small projects that have large project schedules



<50 51-100 101-200 201-500 501-1000

Project Size in Function Points

Diagnose Surface Metrics - Cost Analysis

Cost per FP by Project Size Category

Expected Trend

Symptom: Small project Cost/FP is higher than larger projects (500-1000 FPs)

High

Low

Cost/FP



• Symptoms:

– Process – Processes are very mature with a focus on management and quality processes even at higher levels of maturity

– Productivity – Productivity is low compared to process expectations

– Quality – Better than Best in Class quality may be impacting productivity

– Schedule – Schedules are long for small projects, short for large projects

– Cost – Costs are high, especially for small projects

• Conclusion: Current processes are most likely having a positive impact on quality but a negative impact on productivity

• Next Steps: Conduct a “Deep Dive Diagnosis” to uncover specific process productivity problems

Diagnostic Results - Example



• Project Attributes beyond the CMM(I) need to evaluated

• Methods and techniques need to be analyzed in terms of flexibility and efficiency

– Project Management

– Systems Development Methodologies

– Quality Reviews, inspections and testing

– Documentation

• Project effort should be analyzed in detail

• Project schedules by size category should be compared

• Estimating accuracy should be calculated

• Service level and performance goals should be evaluated

• Measurement and governance activities should be reviewed

Deep Dive Diagnosis Can Uncover Root Causes



Deep Dive Diagnosis - Project Attributes

I/S knowledge and experience

User knowledge and experience

Personnel management and support

Development methods

Project management methods

Quality Assurance and control

Testing methods

Measurement

Computer Resources

Office Environment and Support

Below Industry Norm Above Industry NormAttributes Areas

Below Average Average Best in Practice

Symptom: Management and development disciplines look too good to be true… look deeper!



Deep Dive Diagnosis - Estimating Accuracy

Estimating Accuracy – Estimate to Actual

7% 9%

0%

10%

20%

30%

Schedule Variance Effort Varince

+ - % Variance Estimate to Actual

Symptom: Estimate to Actual Variances of < 10% maybe based on self-fulfilling prophecy and “padding syndrome”



Deep Dive Diagnosis - Percent Effort by Phase

Percent Effort by Life Cycle Phase

Potential Symptom

Standard -Small

Projects

Design0% 50% 75%25% 100%

Construction Test Impl.

% Life Cycle by Phase

Req.

Symptom: Effort by phase is consistent regardless of project size, construction is a small percent of all projects



Symptoms:

• Project Management time is > 15% of the total project effort

• QA time (Inspections and reviews) are > 10% of total project effort

• Number of individual staff members per inspection is greater than 10

• Staff are reporting < 70% or >90% of their available time as productive

• The number of individual names charging time to projects are >15 per 100 Function Points

• Even the smallest projects are consistently charging 500-1500 hours

• Unusually time accounting records (everything ends in zeros)

Deep Dive Diagnosis - Effort Analysis



Symptoms:

• Low levels of user involvement resulting in requirements churn

• Significant pages of documentation and numerous updates

• Service levels that continuously improve without regard for cost

considerations

• High levels of measurement and/or governance overhead

without the benefit

Deep Dive Diagnosis - Misc.



Agenda

• Background





Avoiding Pitfalls and Finding Solutions

• Streamline project management and systems development methodologies

– Reduce “required” tasks, deliverables and management checkpoints

– Establish different paths for different project sizes and types– Reduce the guidelines and criteria for selecting quick path

approaches– Make it easy to waive unneeded activities and deliverables

• Reduce the time associated with quality assurance activities– Establish inspection guidelines to eliminate unproductive staff

and activities– Reduce/eliminate Quality Standards Reviews based on project

type and size

• Group small unproductive projects into optimum size productive projects



Avoiding Pitfalls and Finding Solutions

• Relax service level agreements where appropriate in order to better balance service performance with cost considerations

• Revisit measurement and governance activities and eliminate those without meaning or benefit

• Reduce test cycles where possible by analyzing defect removal statistics versus the cost of quality

• Get back to basics with user involvement in requirements definition

• Use measurement to estimate based on good productivity

• Do not promote process improvement solely for the sake of achieving Level X

• Create an organizational awareness that PRODUCTIVITY is also very important!

the productivity pitfalls of process improvement

Documents