Quality Management in PM²: building quality, not chasing defects
Twentieth article in the “Learning PM² Through a Project” series
Quality does not mean “the highest possible level” or “zero defects at any cost”. In a project, it means that deliverables, processes and results satisfy agreed requirements, standards and acceptance criteria, appropriately for their purpose.
In PM², Quality Management plans, assures and controls quality during the project. Waiting until the end to discover whether a deliverable is acceptable makes corrections more expensive and puts schedule and trust at risk.
Quality, scope and acceptance: related but distinct
| Concept | Question |
|---|---|
| Scope | What must be delivered? |
| Quality | With which characteristics, standards and methods must it be delivered? |
| Acceptance | Who verifies and declares that the deliverable meets agreed criteria? |
A deliverable can be in scope but of insufficient quality: for example, the booking feature exists but does not meet agreed performance, accessibility, security or operational criteria.
The quality cycle
flowchart LR
A[Define standards and criteria] --> B[Prepare methods and responsibilities]
B --> C[Prevent defects in the work]
C --> D[Verify with evidence]
D --> E[Correct and improve]
E --> C
In practice, PM² distinguishes:
| Activity | Objective |
|---|---|
| Quality planning | Define standards, criteria, methods, metrics, responsibilities and control points. |
| Quality assurance | Check that selected processes are applied and adequate to prevent defects. |
| Quality control | Examine deliverables and results to verify compliance and identify defects or deviations. |
Quality must be verifiable
Terms such as “high quality”, “modern interface” or “reliable app” are aspirations, not criteria. A useful criterion is observable and testable.
| Weak statement | Verifiable criterion |
|---|---|
| “The app must be fast.” | “Place availability updates within five seconds in 95% of expected operations.” |
| “The screen must be intuitive.” | “A new collaborator completes check-in using the available guide, without assistance, within two minutes.” |
| “Data must be secure.” | “Authorisations comply with defined roles and unauthorised access is denied and logged.” |
Not every project needs sophisticated metrics. Each needs criteria clear enough to prevent acceptance from meaning only “it looks good to me”.
Quality Management Plan: what it should clarify
The quality plan is tailored to the project. For a small project it may be a Project Handbook section; for a complex one it may be a dedicated plan. In both cases it clarifies at least:
| Element | Example |
|---|---|
| Standards and criteria | Acceptance criteria, technical standards, UX, security, applicable regulations. |
| Methods and controls | Reviews, tests, demos, audits, checklists, inspections, regression tests. |
| Roles | Who produces, verifies, approves and handles non-conformities. |
| Evidence | Test reports, review outcomes, acceptance records, metrics. |
| Thresholds and deviation management | When a non-conformity blocks delivery, how an issue is opened, who may decide a waiver. |
The plan need not repeat established organisational procedures. It should make clear how those procedures apply to this project and its deliverables.
Quality assurance: prevent before testing
Quality Assurance looks primarily at the process, not only the product. Examples:
- code review takes place before merge;
- requirements have acceptance criteria before development;
- critical tests are automated or planned;
- the definition of “done” is shared;
- test data and environments are fit for purpose;
- recurring non-conformities are analysed for causes, not only fixed one by one.
The typical question is: are we working in a way that makes the required quality likely?
Quality control: verify the deliverable
Quality Control looks at the result: functional tests, integration tests, performance checks, inspections, completeness checks, document reviews or demonstrations with users.
A quality-control outcome can be:
- compliant;
- compliant with minor observations;
- non-compliant, with a defect or issue to correct;
- subject to a waiver decision, with risk and impact made explicit.
A waiver does not erase a defect. It is a conscious decision to temporarily accept a non-conformity, documenting rationale, owner and follow-up actions.
The Annotatio case: quality of the waiting-list function
For Smart Bookings, the requirement is: when a place becomes available, the first user on the list is notified and has 30 minutes to confirm.
A proportionate quality plan may include:
| Aspect | Required evidence |
|---|---|
| Functional rule | Test with multiple users on the list, answer, time-out and promotion of the next user. |
| Concurrency | Test two nearly simultaneous cancellations or confirmations without double allocation. |
| Notifications | Check on iOS and Android, with documented fallback if push is unavailable. |
| Security and roles | Only authorised managers configure rules; users see only their own bookings. |
| Usability | Trial with a collaborator not involved in development. |
| Acceptance | The manager validates agreed criteria and the result is recorded. |
Testing only the happy path is not enough. Quality requires attention to exceptions, concurrency, roles, integrations and real use.
Cost of quality: prevention costs less than late correction
There is a real trade-off: reviews, tests and checks take time. But a correction discovered after go-live can require urgent analysis, support, communication, loss of trust and new regressions.
Quality is not adding unlimited controls. It is selecting the controls that most reduce risk to the most important deliverables and benefits.
Do not confuse these
| Statement | Correction |
|---|---|
| “Quality is checked at the end.” | It must be planned, assured and controlled throughout the project. |
| “Quality assurance and quality control are synonyms.” | Assurance concerns process adequacy; control verifies the result. |
| “If it works in testing, it is quality.” | It also depends on performance, security, usability, compliance and agreed criteria. |
| “A waiver makes the deliverable compliant.” | It consciously accepts a non-conformity; it does not eliminate it. |
Three questions to anchor it
- Which objective criterion demonstrates that this deliverable has the required quality?
- Which control prevents a defect before it reaches final testing?
- If a non-conformity is not corrected, who can accept the risk and under which conditions?
Flashcards
Q: What is the difference between Quality Assurance and Quality Control?
A: Assurance checks and improves the process used to prevent defects; control verifies compliance of deliverables and results.
Q: What makes a quality criterion useful?
A: It must be clear, observable and verifiable with evidence or tests.
Q: What is a quality waiver?
A: A documented decision to accept a non-conformity, with rationale, impact and any follow-up actions.
Exam-style question
The team completes development of the booking function. The PM proposes accepting it because all user stories are closed, deferring concurrency testing and role verification until after go-live.
What is the correct assessment?
Answer: closing user stories alone does not demonstrate compliance with quality and acceptance criteria. Relevant controls — here, concurrency and authorisations — must produce evidence before acceptance, unless a waiver is explicitly decided by competent governance.
Source
Quality Management, the Quality Management Plan and deliverable-quality management are covered in the European Commission’s PM² Methodology Guide 3.1.
Next: Communication Management, for getting the right information to each person at the useful time, with clear responsibility.