Sunday, May 25, 2014

8) QUALITY Management

 |  QUALITY Management
  • everyone in the organization is responsible for the quality (project team for destined parts while PM for project quality), PM is ultimately responsible for the project quality
  • prevention over inspection
  • outliers are singular measurements outside the control limits
  • continuous improvement to ensure quality (quality assurance)
  • some costs of quality will be borne by the organization (organization quality policy, e.g. quality audit, ISO accreditation)
  • Quality: the degree to which a set of inherent characteristics fulfills requirements, decrease rework/costs, increase productivity/stakeholder satisfaction
  • under ‘control’: the process is predictable and repeatable
  • PM: perform continuous improvement activities (quality assurance), verify quality before completion of deliverables (control quality)
  • Grade (fit for use or not) vs Quality (conformance to stated requirements)
  • Accuracy (correctness) vs Precision (consistence, how closely conforms to target, standard deviation is a measure of precision, smaller standard deviation higher precision)
  • Quality Management Concepts
    • Crosby Zero Defects: identify processes to remove defects, quality is built in to the processes
    • Juran Fitness for Use: does the product/service meet customer’s need? i) Grade, ii) Quality conformance, iii) Reliability/maintainability, iv) Safety, v) Actual Use
    • W. Edwards Deming: 85% of quality problem is managers’  responsibility, develop “System of Profound Knowledge” [system = components working together to achieve an aim] i) Appreciation for system, ii) Knowledge about variation (special cause vs common cause) , iii) Theory of Knowledge (built up by prediction/observation/adjustment) , iv) Psychology
    • Six Sigma: achieve 3.4/1 mil defect level (99.999%) using DMAIC (Define, Measure, Analyze, Implement, Control) or [Design for Six Sigma] DMADV (Define, Measure, Analyze, Design, Validate) approach, refine the process to get rid of human error and outside influences with precise measurements, variations are random in nature
    • Just In Time: eliminate buildup of inventory
    • Total Quality Management (TQM): ISO 8402 all members to center on quality to drive customer satisfaction , refine the process of producing the product
    • Kaizen改善: implement consistent and incremental improvement, to reduce costs, cycle time, drive customer satisfaction using PDCA (Plan Do Check Act)
    • The Plan-Do-Check-Act cycle is a way of making small improvements and testing their impact before you make a change to the process as a whole. It comes from W. Edwards Deming’s work in process improvement, which popularized the cycle that was originally invented by Walter Shewhart in the 1930s.
    • Capability Maturity Model Integration (CMMI): improve overall software quality (design, development and deployment)
    • ISO9000: ensures the defined processes are performed in accordance to the plan
  • Quality Management processes are so focused on reviewing EVERY deliverable – not just the final product, but all of the components, designs and specifications too.
Project Quality Management PITTOs:

Tools & Techniques
Plan Quality Management
Project Management Plan
Stakeholder Register
Risk Register
Requirements Documentation
Enterprise Environment Factors
Organization Process Assets
Cost-benefit Analysis
Cost of Quality
Seven Basic Quality Tools
Design of Experiments
Statistical Sampling
Additional Quality Planning Tools
Quality Management Plan
Process Improvement Plan
Quality Metrics
Quality Checklists
Project Document Updates
Perform Quality Assurance
Quality Management Plan
Process Improvement Plan
Quality Metrics
Quality Control Measurements
Project Documents
Quality Management and Control Tools
Quality Audits
Process Analysis
Change Requests
PM Plan Updates
Project Document Updates
Organization Process Assets Updates
Control Quality
Quality Management Plan
Quality Metrics
Quality Checklists
Work Performance Data
Approved Change Requests
Project Documents
Organization Process Assets
Seven Basic Quality Tools
Statistical Sampling
Approved Change Requests Review
Verified Deliverables
Change Requests
Validated Change
Quality Control Measurements
Work Performance Info
PM Plan Updates
Project Document Updates
Organization Process Assets Updates
Plan Quality Management
  • quality policy (either organizational or just for the project), methods and procedures to meet the objectives and satisfy customer’s needs
  • identify the quality requirements and document how to achieve
  • Cost-benefit Analysis: cost of implementing quality requirements against benefits
  • Cost of Quality: lowest quality cost is prevention, highest quality cost (poor quality) is rework and defect repair (as high as 5000 times the cost for carrying out unit testing), lost reputation and sales, failure cost may be internal/external (found by customer)
  • Warranty claims are external cost of quality
  • Cost of Quality is the total cost of quality efforts throughout the product’s lifecycle
  • cost of conformance (prevention cost, appraisal cost) vs cost of non-conformance (failure cost [internal/external])
  • internal/external is reference to the project (not the organization)
  • Poka Yoke (mistake proofing), Zero Quality Control (100% source inspection), Voice of Customer and FEMA (Failure Modes of Effects Analysis) are planning tools for quality management
  • Quality Metrics: function points, MTBF (mean time between failure), MTTR (mean time to repair)
  • Marginal Analysis: ROI of quality measures
  • 7 Basic Quality Tools
    • Cause-and-effect / Ishikawa / Fishbone Diagram: for identifying the cause
    • Flowchart: (e.g. SIPOC diagram) for identifying failing process steps and process improvement opportunities
    • Check Sheets (tally sheets): collecting data/documenting steps for defeat analysis
    • Histograms: does not consider the influence of time on the variation that exits within a distribution
    • Pareto Chart: based on 80/20 principle, a prioritization tool to identify critical issues in descending order of frequency, sort of a histogram
    • Statistical Process Control (SPC) Chart: determine if a process is stable/predictable using statistical sampling (assessed by accuracy[conformance] and precision[standard deviation]), identity the internally computed control limits (UCL/LCL) and specification limits (USL/LSL) by the customer/PM
    • run chart is similar to control chart, but without the control
    • usually +-3sigma i.e. a range of 6 sigma
    • a form of time series
    • if a process is within control limit but beyond specification limit, the process is experiencing common cause variation (random) that cannot be corrected by the system, management help is needed (special cause can be tackled but NOT common cause)
    • Stability Analysis / Zone Test: rule of seven (7 consecutive on either side of the mean = out of control), rule of six (six consecutive with a trend = out of control), rule of ten (10 as a saw-tooth pattern around the mean), rule of one (1 point beyond control limit) [signal in the noise]
    • Scatter Diagram: for trending, a form of regression analysis
  • Benchmarking: compare to past activities/standard/competition
  • Design of Experiments (DOE): change several factors at a time for each experiment, to determine testing approaches and their impact on cost of quality
  • Additional Quality Planning Tools
  • Loss Function: a financial measure of the user’s dissatisfaction with product performance
  • Matrix Diagrams: House of Quality (HOQ) used in Quality Function Deployment (QFD) (method to transform user demands [VOC]  into design quality)
  • Kano Model: differentiate features as satisfy, delight or dissatisfy
  • Marginal Analysis: cost-benefits analysis
  • Force Field Analysis (FFA): reviews any proposed action with proactive and opposing forces
  • Process Improvement Plan: process boundaries, configuration, process metrics/efficiencies, targets for improved performance
  • Quality Checklists: checklist to verify a series of steps have been performed
  • The goal is to refine the process so that human errors and outside influences no longer exist, and any remaining variations are completely random
Perform Quality Assurance
  • in the Executing Process Group
  • ensures the quality standards are being followed, to ensure unfinished works would meet the quality requirements
  • by quality assurance department or sponsor/customer not actively involved in the project
  • primarily concerned with overall process improvement for activities and processes (rather than the deliverable)
  • utilize the data collected in Control Quality Process
  • Quality Management Tools
    • Affinity Diagrams: like a mind-mapping diagram, organize thoughts on how to solve problems
    • Process Decision Program Charts (PDPC): defines a goal and the steps involved, useful for contingency planning
    • Interrelationship Digraphs: maps cause-and-effect relationships for problems with multiple variables/outcomes
    • Tree Diagrams
    • Prioritization Matrices: define issues and alternatives that need to be prioritized for decision, items are given a priority score through brainstorming
    • Activity Network Diagrams
    • Matrix Diagrams: e.g. ‘house of quality’ in QFD
    • Kaizen改善, Kanben看板: quality assurance methods developed by Japanese
  • Quality Audit: to verify quality of processes, to seek improvement, identify best practices, reduce overall cost of quality, confirm implementation of approved changes, need quality documentations
  • Quality Review: to review the quality management plan
  • change requests are mostly procedural changes
Control Quality
  • verify the deliverables against customer’s specifications to ensure customer satisfaction
  • validate the changes against the original approved change requests
  • conditional probability (events somewhat related) vs statistical independence (events not interrelated) vs mutual exclusivity
  • statistical sampling for control quality
  • variable (continuous) data: measurements, can do maths on e.g. average
  • attribute (discrete) data: yes/no, no.123, just an identifier, can’t do maths on
  • QC includes the PM process (lesson learnt, budget, scope)
  • tolerance (spec limits, deliverables acceptable) vs control limits (process acceptable)
  • if within control limit but outside tolerance: rework the process to give better precision
  • all control and execution processes may generate lesson learned

No comments:

Post a Comment

Be the first to comment..