How to write a bug report developers can act on
The seven parts of a good bug report, with a worked example. A clear report gets a bug fixed in one pass instead of three.

A bug report has one job: let someone who was not there see the problem for themselves. When it fails at that, the developer writes back "cannot reproduce", and a day is lost.
Here is the structure we use for every finding, and why each part is there.
1. A title that states the problem
The title should say what is wrong and where, in one line. Someone scanning a list of fifty bugs should know what this one is without opening it.
- Weak: "Checkout issue"
- Better: "Coupon SAVE10 is accepted but the total does not change"
A good test: could two different bugs share this title? If yes, it is too vague.
2. An ID
Give every bug a unique ID, such as SQA-1001-005. It sounds like paperwork until three people are discussing "the checkout bug" and mean three different bugs. An ID ends that.
3. Steps to reproduce
Numbered, one action per step, starting from a known place. Write them so that someone who has never seen the product could follow them.
- Add the Trail jacket to the basket
- Open the checkout page
- Enter coupon
SAVE10and press Apply - Look at the total
Include the exact data you used. "Enter a coupon" is not the same as "enter SAVE10". Often the bug only happens with one particular value.
4. Expected result and actual result
Two short lines, kept separate:
- Expected: the total drops from $119.00 to $107.10.
- Actual: the page says "Coupon applied: 10% off", but the total stays at $119.00.
This is the part people most often leave out, and the most important. Without the expected result, the developer has to guess what you think is wrong. Sometimes the software is behaving as designed and the report is really a question.
5. Where it happened
Browser and version, device, operating system, screen size, and which environment (live, staging, a test build). A bug that appears only in Safari on an iPhone is a very different job from one that appears everywhere.
If you tested in several places, say which ones showed the bug and which did not. "Fails in Safari, works in Chrome" narrows the cause enormously.
6. Evidence
A screenshot for anything visual. A short screen recording for anything that involves timing or several steps. For web apps, the browser console errors and the failing network request if you can get them.
Mark up the screenshot. An arrow pointing at the wrong number saves the reader from hunting for it.
7. Severity
How bad is it for the user? We use four levels:
- Critical: blocks a main journey, or loses money or data. Nobody can check out.
- High: a main feature is broken and there is no easy way round.
- Medium: something is wrong, but people can still finish what they came to do.
- Low: small or cosmetic. A typo, a misaligned icon.
Severity is about impact on the user, not how hard the fix is. Keep those separate, or everything becomes "high".
What to leave out
- Guesses presented as facts. "The database is broken" when all you know is that the page showed an error. Say what you saw.
- Several bugs in one report. One report, one bug. Otherwise it can never be closed.
- Blame. The report describes software, not people.
Why this pays off
A report with all seven parts usually gets fixed on the first attempt. A report missing the steps or the expected result goes back and forth, and each round costs a day.
Every finding we log follows this structure, and the report you receive, as a PDF or an Excel sheet, lists them with their IDs, severity, steps, who reported each one and when. See how that fits into website testing or mobile app testing.