Manual TestingPublished: Updated: · 1 month ago12 min read
Types of Software Testing with Examples (2026 Complete Reference)
A complete taxonomy of software testing types — functional, non-functional, manual, automated, black-box, white-box — with a real example and tool recommendation for each.
Avinash Kamble
Founder & QA Engineer at SoftwareTestPilot
Last updated: July 11, 2026 · 12 min read
Software testing is not one activity — it's a family of disciplines. This reference groups every major testing type by category, gives a one-line example, and points you to the tool of choice in 2026. Pair with our Manual Testing Complete Guide and Test Pyramid.
The four main buckets
- Functional — does the app do what the spec says?
- Non-functional — how well does it do it? (performance, security, usability)
- Structural / white-box — is the internal code correct?
- Change-related — do previous fixes still work? (regression, retesting)
Functional testing types
| Type | Example | Typical tool |
|---|---|---|
| Unit testing | Test calculateTax() in isolation | Jest, JUnit, pytest |
| Integration testing | API + DB layer end-to-end | REST Assured, Supertest |
| System testing | Whole checkout flow on staging | Playwright, Selenium |
| UAT | Business users validate a release | TestRail, Jira |
| Smoke testing | Post-deploy sanity of top-5 flows | Playwright + CI |
| Sanity testing | Hotfix spot-check | Manual |
| Regression testing | Re-run test pack after change | Playwright, Cypress, Selenium |
Non-functional testing types
| Type | Example | Typical tool |
|---|---|---|
| Performance testing | Response time under 500 users | k6, JMeter |
| Load testing | System behavior at expected peak | k6, Locust |
| Stress testing | Behavior beyond capacity | JMeter, Gatling |
| Security testing | OWASP Top 10 scan | ZAP, Burp Suite |
| Accessibility testing | WCAG 2.2 audit | axe-core, Lighthouse |
| Usability testing | Task success + time-on-task | Maze, UserTesting |
| Compatibility testing | Chrome / Safari / Edge parity | BrowserStack, Sauce Labs |
| Localization testing | de-DE currency + RTL layouts | Playwright + fixtures |
Structural (white-box) testing
- Statement coverage — every executable statement hit at least once.
- Branch coverage — every if/else path executed.
- Mutation testing — deliberately break code, verify tests catch it (Stryker, PIT).
- Path testing — every logical path through a function.
Black-box vs white-box vs grey-box
| Approach | Knowledge required | Typical tester |
|---|---|---|
| Black-box | Requirements only | Manual QA |
| White-box | Full source code | Developers, SDETs |
| Grey-box | APIs + partial internals | API testers, SDETs |
Exploratory & ad-hoc testing
Not scripted. Testers simultaneously learn, design tests, and execute — highly effective at finding bugs that scripted tests miss. See our deep dive on exploratory testing.
Continue your learning
How to pick the right test type — a decision matrix (2026)
The taxonomy is easy to memorise. Choosing which type to run for a given change is what separates a junior tester from a senior QA lead. Use this decision matrix on any pull request:
| Change type | Must-run test types | Nice-to-have |
|---|---|---|
| New UI component | Unit, component, visual regression, accessibility | Cross-browser E2E |
| New API endpoint | Unit, contract (Pact), integration, security (auth, IDOR) | Load, chaos |
| Refactor with no behaviour change | Full regression suite, mutation testing | Performance diff |
| Database migration | Migration rehearsal on prod-like data, rollback drill, integration | Data quality tests |
| Third-party integration swap | Contract, integration in a sandbox, feature-flag canary | Chaos test for provider outage |
| Copy / content change | Visual regression, i18n snapshot | Accessibility rerun |
Worked example: adding a Save-for-later button to an e-commerce cart
Which test types fire? A senior QA would spec at least:
- Unit — the reducer that adds an item to savedItems.
- Component — the button renders the right icon in the three cart states (empty, one item, over stock limit).
- Integration — POST /api/cart/save persists to the user session and deduplicates.
- End-to-end — add to cart, save, sign out, sign back in, item is still saved.
- Accessibility — button has an aria-label, is reachable by keyboard, and announces state change.
- Visual regression — cart layout unchanged on the four highest-traffic viewports.
- Analytics validation — a cart_item_saved event fires with the correct SKU.
Miss any one of those and something reaches production — usually the analytics event, which product managers discover three sprints later when the funnel report is empty.
Functional vs non-functional at a glance
- Functional answers does the feature do what it should? — unit, integration, system, UAT.
- Non-functional answers does it do it well enough? — performance, load, security, usability, accessibility, compatibility, reliability.
Every release should touch both categories. Teams that skip non-functional testing are the ones that make headlines for outages after a viral launch.
Frequently asked questions
1.How many types of testing are there?
Common taxonomies list 40–60 distinct types, but 90% of QA work centers on 10: unit, integration, system, UAT, smoke, sanity, regression, performance, security, and accessibility.
2.What's the difference between functional and non-functional testing?
Functional testing verifies what the system does. Non-functional testing verifies how well it does it — speed, security, usability, reliability.
3.Is exploratory testing black-box or white-box?
Usually black-box, but a technically savvy tester can do grey-box exploration using APIs, logs, and DB queries alongside the UI.
4.Which testing type should I automate first?
Unit tests (owned by devs), then API integration tests, then the top 20 E2E user journeys — following the test pyramid.
5.Which test type catches the most bugs?
Integration tests catch the highest ratio of production-blocking bugs per test, according to Google 2024 Testing on the Toilet data. Unit tests catch the most bugs in absolute numbers because there are more of them.
6.Is exploratory testing a type?
Yes — it is an approach that spans functional, usability, and security. It is the highest-ROI type per hour because a skilled tester finds bugs no automated suite would think to look for.
7.Do I need every type on every release?
No. Use the decision matrix above. A one-line copy change does not need load testing; a payments refactor absolutely does. Mapping change type to test types is the QA lead daily job.
8.What is the difference between smoke and functional testing?
Smoke is a shallow subset of functional testing focused on build stability. Functional testing covers the full behaviour of every feature and its edge cases.