The simplest distinction
Functional testing asks: Does the system do the right thing?
Non-functional testing asks: How well, safely, reliably, and efficiently does it do it?
Non-functional testing asks: How well, safely, reliably, and efficiently does it do it?
A feature can be functionally correct and still be too slow, insecure, inaccessible, unstable, or difficult to operate.
Functional testing
Functional testing validates behaviour against requirements, business rules, user stories, workflows, and interfaces.
Examples include login, calculations, permissions, searches, transactions, approvals, data updates, notifications, API responses, and error handling.
Non-functional testing
| Attribute | Example question |
|---|---|
| Performance | Can the system meet response targets under expected load? |
| Security | Are access, data, and attack surfaces sufficiently protected? |
| Accessibility | Can people with disabilities use the product effectively? |
| Reliability | Does the system behave consistently over time? |
| Resilience | What happens when dependencies fail? |
| Compatibility | Does it work across supported platforms? |
| Recoverability | Can service and data be restored? |
Example: online checkout
Functional
Add items, apply discount, calculate tax, authorize payment, create order, send confirmation.
Non-functional
Remain responsive under peak load, protect payment information, support keyboard-only use, avoid duplicate orders during retry, degrade safely if email is unavailable.
Make non-functional requirements measurable
Weak
“The dashboard should load quickly.”
Better
“The dashboard should render within 2 seconds for 95% of requests under the agreed standard workload.”