Key takeaways
- Quality engineering (QE) treats quality as a property of the whole system of work: requirements, design, code, pipelines, operations and feedback loops. It's not a phase or a department.
- The QE role shifts from gatekeeper to enabler: building test infrastructure, coaching teams, owning quality strategy and risk, and closing the loop from production back to development.
- Measure outcomes, such as delivery performance, escaped defects, reliability against SLOs and customer impact, not testing activity. Research summarised in Accelerate shows speed and stability can improve together.
Where this picks up
You've seen testing in Agile teams, shift-left and shift-right, continuous testing and observability. Quality engineering is what those practices add up to when an organisation takes them seriously.
From testing to QE
| Traditional testing | Quality engineering |
|---|---|
| A phase after development | Continuous, from idea to production |
| A separate team checks others' work | The whole team owns quality; specialists enable it |
| Success = bugs found | Success = fewer bugs created, faster recovery, happy users |
| Manual regression at the end | Automated checks in the pipeline; humans explore and analyse risk |
| Gate at release | Many small feedback loops; release safety through canaries and rollback |
| Reports test counts | Reports risk, outcomes and trends |
This doesn't make testing skills less important. It makes them more widely applied: risk analysis, test design, exploration and critical thinking inform requirements, architecture, pipelines and operations.
What quality engineers do
- Quality strategy and risk: own the test strategy, keep the risk picture current and advise on release decisions.
- Enable, don't bottleneck: build test frameworks, test data tools, environments and pipeline checks that let every developer test well.
- Coach: pair with developers on test design; run three amigos and example mapping; teach exploratory testing.
- Deep testing where it matters: exploratory sessions, performance, security, accessibility and resilience for the riskiest areas.
- Close the loop: analyse escaped defects and incidents; turn them into tests, monitors and process changes.
- Guard the suite's health: speed, flakiness, adequacy and trust.
Measuring what matters
Delivery performance (DORA)
The DORA research programme, summarised in the book Accelerate (2018) by Nicole Forsgren, Jez Humble and Gene Kim, identified four key measures of software delivery:
| Measure | Question |
|---|---|
| Deployment frequency | How often do we deploy to production? |
| Lead time for changes | How long from commit to running in production? |
| Change failure rate | What share of deployments cause a failure needing remediation? |
| Time to restore service | How quickly do we recover from a failure? |
Two measure throughput; two measure stability. Later DORA reports refined the definitions and added reliability, but the core idea remains.
Quality outcomes
- Escaped defects by severity and by the layer that should have caught them.
- Reliability against SLOs and error budget use.
- Customer-facing signals: support tickets per release, complaint categories, accessibility issues reported.
- Suite health: pipeline duration, flaky rate, time to diagnose failures.
For years, many organisations believed that going faster meant accepting more failures, so they slowed releases and added gates to protect quality.
The DORA research, based on survey data from thousands of technology professionals, found the opposite pattern. The highest-performing teams deployed more often and with shorter lead times, and had lower change failure rates and faster recovery than low performers. Practices associated with high performance included test automation, continuous integration, trunk-based development and monitoring, the core of quality engineering.
Lesson: quality and speed reinforce each other when quality is built into the delivery system. Gates at the end slow you down without making you safer; fast, reliable feedback does both.
Leading the shift
- Start with outcomes: agree what quality means for your users and business, and pick a few outcome measures.
- Map the feedback loops: where are defects created, where are they found, how long does each loop take? Shorten the longest, most expensive ones.
- Make testing everyone's job: developers own unit and integration tests; QEs make that easy and teach test design.
- Invest in the platform: fast pipelines, reliable environments, test data tooling, observability.
- Keep deep testing human: protect time for exploratory, risk-based and non-functional testing.
- Learn blamelessly: incident and escaped-defect reviews that focus on systems, not people.
- Grow people: career paths for quality coaches, automation engineers and quality leads, not a single "tester" track.
Do
Report quality as a short story with evidence: "Change failure rate fell from 18% to 7% this quarter; most escaped defects were in pricing rules, so we're adding decision-table tests and a canary for pricing changes."
Don't
Rename the test team "QE", keep the same end-of-cycle gate and measure success by bugs found. The name changed; the system didn't.
Check yourself
A company's release process has a two-week manual regression phase at the end. Change failure rate is high and releases happen monthly. What would a quality engineering approach change first?
Show answer
Look at the feedback loops: automate the most valuable regression checks into the pipeline at the lowest effective layers, so changes get feedback in minutes; introduce smaller, more frequent releases with canaries and fast rollback; move testers' time to risk analysis, exploratory testing and coaching earlier in the cycle; and track DORA measures and escaped defects to show whether the changes work.
