Working once does not mean a solution has the quality needed for its intended use. Nor does a high price, attractive appearance, or long feature list prove quality.
A quality solution delivers its intended outcome under relevant conditions, with risks appropriately managed. This means considering what the person needs, which requirements were agreed, what failure would mean, and what evidence justifies confidence. No universal indicator works equally well for an app, a package, and a delivery service. Quality management is not just the final check: it depends on design, training, processes, and the ability to learn from failure [R01][R02].
1. Define quality by context, not luxury
A component can have a perfect finish yet fail when assembled. An app can look modern while exposing private information. A delivery service can be friendly while arriving late repeatedly. The customer experiences an outcome, not isolated attributes.
To define quality, ask: What are the intended uses? Who is exposed to failure? Which requirements are essential? Which operating conditions could occur? What counts as a defect? What evidence supports acceptance? Separate perceived characteristics—clarity, comfort, confidence—from verified properties—measured strength, access controls, recorded delivery time. Both matter, but are evaluated differently.
Quality does not mean that no one will ever encounter a problem. It means using clear requirements and risk controls, detecting weaknesses, fixing causes, and not claiming results that have not been demonstrated.
2. Identify quality dimensions for the solution
| Solution | Potential quality dimensions | Example evidence |
|---|---|---|
| Booking application | Functional suitability, integrity, usability, security, reliability | Test cases, permission checks, observed sessions |
| Physical product | Dimensional fit, strength, durability, safety, batch consistency | Relevant inspections and tests |
| Delivery service | Timeliness, condition of delivery, courtesy, information, incident recovery | Delivery records, complaints, observations |
| Hybrid solution | Integration and human coordination | End-to-end exercise and handoff checks |
ISO/IEC 25010:2023 defines nine characteristics for ICT and software product quality. It can inform requirements and tests in that domain, but does not alone define the quality of every food product or service [R03]. WCAG 2.2 provides testable web accessibility criteria, though automated checks do not replace manual evaluation and research with relevant users [R04].
Do not turn these dimensions into a mechanical “everything must be perfect” checklist. Select what affects core value and credible risks in your context.
3. Distinguish quality planning, assurance, and control
Quality planning defines the quality needed, how to show it, and who owns it. Quality assurance (QA) covers systematic activities that build confidence requirements will be met. Quality control (QC) focuses on operational checks and inspections that detect deviations. ASQ distinguishes these related concepts [R02].
In food delivery, defining procedures, training staff, and reviewing how work is done can support assurance; checking a particular delivery's relevant conditions can be control. In software, code review, change controls, and automated testing can contribute to QA; observed behavior and detected defects provide QC evidence.
Having a person called “QA” or buying a testing tool is not a sufficient quality system. Criteria, actions, evidence, and a response process must exist.
4. Turn quality attributes into assessable criteria
Use the requirements matrix from P5-04. For each quality attribute record:
- Use context: who uses it and under what circumstances.
- Expected property or behavior: what should happen.
- Verification method: inspection, trial, observation, log, or audit.
- Acceptance criterion: a grounded threshold or condition.
- Evidence: saved result, date, version, and owner.
- Consequence of failure: correction, retest, limited use, or block.
“Fast delivery” expresses intent. “Within the agreed window, measured using pilot records under its stated area and hours” approaches a testable criterion, though it still needs a defined window, procedure, and exceptions. Do not invent targets for accuracy, speed, strength, or accessibility simply to complete a table.
When a quality dimension is still expressed as an adjective, use the requirements and acceptance criteria specification to connect it to a scenario, justified threshold, and concrete check.
5. Use standards without inventing obligations
A standard may guide methods and terminology; a binding requirement might arise through law, contract, or market conditions; certification involves a specific evaluation process and claim. They are not interchangeable.
- ISO 9001:2026 specifies organizational quality management system requirements across industries; it does not provide the full technical specification for every product. Certification is voluntary unless a specific context requires it [R01].
- ISO/IEC 25010:2023 provides an ICT product quality model [R03].
- WCAG 2.2 defines web accessibility conformance criteria at A, AA, and AAA levels; passing selected checks does not establish complete conformance [R04].
These descriptions come from public summaries; commercial standards are not reproduced here. Confirm the relevant edition, jurisdiction, and contractual clauses before asserting compliance. A small pilot does not automatically create exceptions to safety or privacy obligations.
To distinguish voluntary standards from actual obligations, use the investigation of legal and operating requirements applicable to the activity and jurisdiction. Naming a standard does not establish that it is mandatory or that certification exists.
6. Design a quality plan proportionate to risk
Not every characteristic deserves the same effort. A cosmetic defect might be corrected later; a data leak or unsafe component may require stopping delivery. Consider impact, exposure, detection, reversibility, and strength of evidence. Pillar 4's risk matrix evaluates broader threats. Here we specify how those threats are controlled in the actual solution.
| Dimension | Preventive control | Verification | Response on failure |
|---|---|---|---|
| Booking integrity | Change and conflict rules | Alternative-path test cases | Block affected version and fix |
| Confirmation clarity | User-reviewed interaction design | Observe real tasks | Redesign and retest |
| Component consistency | Specification and stable process | Appropriate sample inspection | Hold batch and investigate |
| Delivery timeliness | Capacity planning | Comparable timing records | Limit scope and correct operations |
This is an instructional table. Actual methods, sample sizes, and criteria depend on risk and requirements. Do not replace qualified testing or specialist advice when stakes are high.
The controls described in the plan are carried out through concrete solution tests, with preconditions, expected outcomes, and evidence records. A quality plan without checks does not demonstrate conformity.
7. Measure variation, not just one favorable result
A successful demonstration says little about consistency. Repeat checks across relevant conditions: different shifts, novice operators, devices, material batches, or peak demand. Record how many cases were evaluated, failures, missing records, and protocol changes.
If eight of ten fictional deliveries meet the window and two do not, do not announce “We guarantee punctual delivery.” Investigate causes and decide what can be promised. Nor should percentages from small samples be presented as representative of every condition. Record quality matters as much as the resulting number.
When nonconformance occurs, distinguish an isolated defect, a process problem, and a poorly defined requirement. Sometimes fixing an interface is insufficient because the coordination system still fails.
8. Decide whether quality supports the next commitment
Before broader use or release, review evidence of essential requirements, open defects, recovery controls, residual risks, and support capacity. A review might conclude fit for a limited test, fit for agreed delivery, requires correction, or not fit. These are editorial decision aids, not substitutes for regulatory approvals.
Record approval ownership, evidence reviewed, remaining conditions, and review triggers. Pillar 4's Go/No-Go examines business readiness. This quality plan examines conditions of the solution itself.
Worked example: three solutions, three different failures
A fictional team develops a booking website, a physical tablet mount, and an in-person reception service.
The app stores reservations, but some users leave without seeing confirmation: experience problem and duplicate-booking risk. The mount appears to fit, but one unit bends: strength or variation concern. The reception team is polite but communicates times inconsistently: procedure and coordination problem.
One “works: yes/no” checkbox cannot cover all three. The team defines suitable criteria, logs failures, corrects processes, and repeats checks. None of the examples establishes standards conformance or scaled capacity without further evidence.
Common mistakes and limitations
Treating quality as visual polish; measuring only satisfaction; defining untestable criteria; assuming automation equals a complete audit; thinking management certification proves every product unit; using “certified” without evidence; excluding accessibility needs; hiding defects to meet a date; omitting failure recovery.
Quality is demonstrable within a defined context, not an absolute promise. Its criteria need revision as use, conditions, or evidence change.
What you should have when finished
A solution quality plan specifying relevant dimensions, requirements, preventive controls, verification methods, acceptance criteria, evidence, owners, and responses to failure. It should support release decisions without mistaking quality for marketing.
References
- [R01] ISO. ISO 9001:2026 — Quality management systems — Requirements (public summary, 2026 edition; certification is not universally mandatory). https://www.iso.org/es/norma/9001
- [R02] American Society for Quality (ASQ). Quality Assurance vs Quality Control. https://asq.org/quality-resources/quality-assurance-vs-control
- [R03] ISO. ISO/IEC 25010:2023 — Product quality model (public abstract). https://www.iso.org/standard/78176.html
- [R04] W3C. Web Content Accessibility Guidelines (WCAG) 2.2. https://www.w3.org/TR/WCAG22/