Key takeaways
- Requirements describe what the software must do and how well. Testers read them critically: a requirement that's ambiguous or untestable is a bug in the requirement.
- Traceability links each requirement to the tests that check it, and each test back to its requirement. A requirements traceability matrix (RTM) shows it in one table.
- Traceability answers three practical questions: is everything covered? What's the status of each requirement? What must be retested when something changes?
What requirements look like
Requirements come in many shapes, depending on the team:
- Business requirements: the goal. "Reduce abandoned carts by 10%."
- Functional requirements: what the system does. "The user can save a card for future payments."
- Non-functional requirements: how well it does it. "The payment page loads in under 2 seconds for 95% of users." See functional vs non-functional testing.
- User stories with acceptance criteria, in Agile teams. See acceptance criteria and BDD.
Whatever the format, your job is the same: understand what "correct" means, so you can check it.
Reading requirements like a tester
A good requirement is clear, complete, consistent and testable. Read each one and ask:
| Question | Weak requirement | Better |
|---|---|---|
| Can it be measured? | "The search should be fast." | "Search results appear within 1 second for 95% of queries." |
| Is there only one meaning? | "Users can upload large files." | "Users can upload files up to 50 MB; larger files show an error." |
| What about errors? | "The user enters their date of birth." | "...Users under 18 see 'You must be 18 or older to register'." |
| Does it contradict another one? | Page 3: "guest checkout allowed"; page 7: "login required to pay". | Ask which is right, and get one of them fixed. |
| Is anything missing? | "Send a confirmation email." | When? From which address? In which language? What if it fails? |
Asking these questions is static testing, and it's the cheapest testing there is: you find the bug before anyone writes the code.
Traceability: connecting requirements and tests
Traceability means you can follow a line from each requirement to the tests that cover it, and from there to the results and bugs.
- Forward traceability: requirement → tests. Is every requirement tested?
- Backward traceability: test → requirement. Does every test have a reason to exist?
A traceability matrix
The simplest form is a table:
| Requirement | Description | Test cases | Status | Bugs |
|---|---|---|---|---|
| REQ-01 | Save a card during checkout | TC-01, TC-02, TC-03 | ✅ Passed | none |
| REQ-02 | Pay with a saved card | TC-04, TC-05 | ❌ Failed | BUG-112 |
| REQ-03 | Delete a saved card | TC-06 | ⏳ Not run | none |
| REQ-04 | Expired cards can't be used | none | ⚠️ Not covered | none |
At a glance, you see that REQ-04 has no tests at all, REQ-02 is blocked by a bug, and REQ-03 hasn't been tested yet. That's information the team needs before a release decision.
What traceability is good for
- Coverage: find requirements with no tests before release, not after.
- Status reporting: "18 of 20 requirements pass; one fails (BUG-112); one is untested." That's far more useful than "we ran 240 tests".
- Impact analysis: when REQ-02 changes, you know exactly which tests (TC-04, TC-05) to update and rerun.
- Audits and compliance: in regulated fields like medical, automotive, banking and aviation, traceability is often required, to prove every requirement was verified.
A bank's mobile app team was preparing a release with a new money transfer feature. Every test case passed. The release was scheduled for Monday.
On Friday, a tester built a quick traceability matrix from the user stories. One acceptance criterion had no test linked to it: "Transfers above 5,000 lei require confirmation with a one-time code." It had been added late, during a meeting, and never made it into the test cases.
The tester tested it that afternoon. Transfers above the limit went through without asking for the code. The release was moved by three days, and the fix was tested properly.
Lesson: "all tests passed" only means something if the tests cover all the requirements. Traceability is how you know.
Do
Link every test to the requirement or acceptance criterion it checks, and report status per requirement, not just per test.
Don't
Assume a requirement is covered because "there are lots of tests in that area". Check the links.
Check yourself
The product owner changes a requirement: the free shipping threshold goes from 200 to 150 lei. How does traceability help you, and what would you retest?
Show answer
The matrix shows which test cases are linked to the free-shipping requirement. Those must be updated (new boundary values: 149.99, 150, 150.01) and rerun. It may also show related requirements that use the same threshold, like cart messages or promotions, which should be checked too. Without traceability, you'd be guessing which tests to touch.
