Quality Management and Testing in IT Projects: QA vs QC, Test Levels and UAT

A project delivered on time and within budget that does not work is a failure. Quality management makes sure the project’s deliverables meet requirements and that the processes used to create them are effective. In IT projects it covers planning what “good” means, building quality in and verifying it, above all through testing. This lesson explains the vocabulary, the levels of testing and how to run user acceptance well.

Quality Planning

Quality starts with defining it. Clarify the standards that apply (internal coding standards, security policies, regulatory rules), the measurable quality criteria (for example, fewer than five open high-severity defects at go-live, 99.5 per cent availability, response times under two seconds) and the activities that will check them. Record this in a quality management plan. Include the cost of quality in your budget: prevention and appraisal cost money, but failures cost much more. Rework, support calls, lost reputation and emergency fixes are all costs of poor quality.

Quality Assurance vs Quality Control

AspectQuality assurance (QA)Quality control (QC)
FocusProcess: are we doing the right things the right way?Product: does the deliverable meet requirements?
TimingThroughout the project, preventiveAt deliverable stages, detective
ExamplesProcess audits, peer review standards, coding guidelines, trainingTesting, inspections, defect counts, acceptance checks
Question answered“Will our process produce good results?”“Did we produce a good result?”

In practice, job titles blur the two, but understanding the distinction helps you balance prevention and detection. Do not rely only on testing to “inspect in” quality.

Verification vs Validation

Verification asks, “Did we build the product right?”, checking against specifications through reviews, inspections and tests. Validation asks, “Did we build the right product?”, checking that it meets the users’ real needs, typically through acceptance testing and demonstrations.

Levels of Testing

LevelWhat is testedUsually performed by
Unit testingIndividual functions or componentsDevelopers
Integration testingInterfaces between components and systemsDevelopers and testers
System testingThe complete system against requirementsTest team
User acceptance testing (UAT)Whether the system supports real business processesBusiness users

Alongside these levels are test types: functional (does it do what it should), regression (did the change break something that used to work), performance and load (does it cope with expected and peak usage), security (vulnerability and penetration tests), usability and compatibility (browsers, devices, operating systems). Test automation pays off for regression checks that are repeated frequently, and is a key practice in Agile and DevOps, alongside test-driven development.

Defect Management

A defect (bug) is a deviation from expected behaviour. Track every defect with an ID, description, steps to reproduce, environment, severity, priority, owner and status. Remember the difference:

  • Severity is the technical impact (critical: system crash or data loss; major: key function fails; minor: cosmetic).
  • Priority is the urgency of fixing, decided with the business. A typo in the company name on the home page is low severity but high priority.

A typical defect life cycle is: new, assigned, in progress, fixed, retest, closed (or reopened or deferred). Hold regular triage meetings to decide what to fix, defer or reject, and monitor trends: defects found per week, open by severity, and age. A falling discovery rate with rising test coverage suggests stabilisation; a rising rate late in testing is a red flag.

Entry and Exit Criteria

Agree objective criteria for starting and finishing each test phase. Entry criteria might be “build deployed to test environment, smoke tests passed, test data ready”. Exit criteria might be “100 per cent of planned tests run, no open critical or high defects, 95 per cent of medium defects resolved, performance targets met”. These prevent the common temptation to declare testing done because the date arrived.

Running User Acceptance Testing Well

  1. Plan early: identify UAT participants when requirements are written, and get their managers to release them.
  2. Base tests on business scenarios, not just features: “Process a customer return from request to refund”.
  3. Prepare realistic data and a stable, production-like environment.
  4. Train testers on how to execute scripts and log defects.
  5. Provide support: a daily check-in to answer questions and triage findings.
  6. Define acceptance criteria in advance and record sign-off formally.
  7. Control change: new requests found in UAT go through change control; see Scope Creep and Change Control.

UAT often fails not on technology but on logistics: users are busy, data is wrong or expectations about what is in scope were never aligned.

Quality Tools

  • Cause-and-effect (fishbone) diagrams to trace defects to root causes.
  • Pareto charts to show that a few causes produce most of the defects.
  • Control charts to see whether a process is stable.
  • Checklists and peer reviews for code and documents.
  • Audits and retrospectives to improve processes; see Sprint Retrospective Explained.

Quality in Agile

Agile teams embed quality into each iteration through a shared Definition of Done, automated tests, continuous integration and frequent demonstrations to users. Quality is everyone’s responsibility rather than the job of a separate phase. See Increment and Definition of Done Explained.

Common Mistakes

  • Compressing testing when the schedule slips, which only moves the problem into production.
  • Leaving non-functional testing, such as performance and security, to the end.
  • Starting UAT with an unstable system, wasting users’ goodwill.
  • Counting tests executed instead of risks covered.
  • Skipping acceptance criteria, so sign-off becomes an argument.
Key takeaways
  • QA is about the process and prevention; QC is about the product and detection.
  • Test at four levels: unit, integration, system and UAT, plus regression, performance and security as needed.
  • Distinguish severity (technical impact) from priority (business urgency).
  • Set entry and exit criteria, and prepare UAT early with real business scenarios.

Related Lessons