Objective: Write a defect report a developer can reproduce cold — title with values, environment, minimal steps, expected vs actual, evidence, and severity argued from business impact — then close the loop with a retest.
⏱ ≈2 h full lesson · Includes the real repro video (62 s) · English edition — Español · العربية · বাংলা in translation
🧽 The practice business: FreshNest Home Cleaning — a fictional local cleaning company (owner: Maya Torres) invented for this course. Every example and exercise uses FreshNest, so you learn on a complete, realistic business without ever touching a real one.
Tap a title to open that way of learning. Use as many as you need — then take the lesson quiz below.
A bug report has one job: let a developer who has never seen the bug reproduce it cold, understand its cost, and prove their fix. Every field exists for that job. Title: the defect in one line, with the wrong and right values when there are numbers. Environment: where it happened, exactly. Steps: numbered, minimal, with the data used. Expected: quoted from the requirement. Actual: what the software did instead. Evidence: a screenshot, a video, or a log — attached, not described. Severity: argued from business impact, not from how annoyed you are.
Two disciplines separate professional reports from complaints. First, isolate: our example report notes that Standard pricing and Move-Out behave correctly — that one sentence tells the developer the fault lives in the Deep multiplier, not the size calculation, and halves their search. Second, close the loop: a report is not done at "fixed". The same test case re-runs on the fixed build (the retest), and its neighbours re-run too (regression), because fixes break neighbours.
Severity and priority are different decisions. Severity is the damage if the defect ships. Priority is when the team chooses to fix it. A wrong price on the most-booked service is high severity because every quote loses money or starts an argument — argue it in those terms and the priority conversation becomes easy.
DEF-FN-001, the course's real defect report, filed from TC-FN-PRC-001 against the FreshNest practice build. Watch the repro video in the Visual tab first if you can — this is the same bug, performed.
Deep clean estimate uses the wrong multiplier — shows $168 instead of $210 (2 bedrooms).
FreshNest practice site (buggy build), desktop Chrome, 900×800. Reproduced 3/3 attempts.
1) Open the home page. 2) Service = Deep. 3) Bedrooms = 2. 4) Read the estimate.
Expected $210 (Standard $140 × 1.5, REQ-FN-01). Actual $168 — which is 140 × 1.2, pointing straight at the multiplier value.
High. Every Deep quote is 20% under the brief's price on the most common home size; the business loses money or argues later. No customer workaround.
Fixed build, same steps: $210 ✓. Full automated suite: 6 passed. Neighbours PRC-002 and PRC-003 still pass — the fix broke nothing.
One map of the lesson: every field exists so a stranger can reproduce the bug, price it, and prove the fix.
Classify each fragment from a imaginary draft report: is it missing, present but weak, or complete?
1. Steps: "Use the site for a while and the price goes wrong."
2. Title: "Pricing bug"
3. Expected: "The Deep price should be right." Actual: "It isn't."
4. Evidence: screenshot of the estimate reading $168 with Deep + 2 bedrooms selected, plus the failing automated run log.
5. Severity: "High — every Deep quote is 20% under the brief's price on the most common home size."
A bug report has one job. A developer who has never seen the bug must be able to reproduce it cold, understand what it costs, and prove the fix. So the title carries the numbers: shows one hundred sixty eight dollars instead of two hundred ten. The steps are numbered and minimal. The expected result quotes the requirement. The actual result says what the software did. Evidence is attached, not described. And severity is argued from the business: every Deep quote is twenty percent under price, on the most common home size. The report is not finished when someone says fixed. It is finished when the same test case passes on the fixed build, and its neighbours still pass too.
The document that turns "it's broken" into work a developer can do.
You find the price bug at 4:50 pm; the release is at 6. The developer says "just tell me quickly in chat." You…
Explain to a friend: why is "the price is wrong" not a bug report, but DEF-FN-001 is? Name at least four fields and what each one does.
Now check yourself honestly:
| The weak way | The strong way |
|---|---|
No steps, no values, no evidence. "Deep clean price is wrong. Please fix." | Isolation note: Standard and Move-Out are correct. DEF-FN-001: steps, $168 vs $210, screenshot + video + failing test. |
1. A developer replies "can't reproduce." Which fields do you check first?
2. Why does DEF-FN-001 mention that Standard pricing passes?
3. The Deep price defect is called High severity because…
Watch it done — the manual reproduction (62 s): Deep + 2 bedrooms shows $168 instead of $210, then a letters-only phone number is accepted with a success message. This is the exact bug filed as DEF-FN-001.