The Software Tester Trait Every Hiring Manager Secretly Wants

Title: Order total wrong

Steps:
1. Add items to cart
2. Apply coupon
3. Go to checkout

Actual: total is wrong
Expected: total should be correct

Severity: High

Title: Checkout total is wrong when SAVE10
       and FLAT50 are stacked

Environment: staging, build 2.14.0, Chrome and Firefox

Steps:
1. Add one item priced 1000 to cart
2. Apply SAVE10 (10%, max 100)  -> total 900, correct
3. Apply FLAT50                -> total shown 820

Expected: 850 (900 - 50)
Actual:   820

Isolation:
- Each coupon alone gives the correct total
- Same wrong value in the cart API response,
  so the issue is server-side, not display
- Not related to browser or user account

Attached: API response, screen recording

CHECK WHETHER A WRONG VALUE COMES FROM THE SERVER

1. Open DevTools (F12) and go to the Network tab.
2. Tick 'Preserve log' so requests survive page loads.
3. Filter by Fetch/XHR.
4. Reproduce the bug (apply the second coupon).
5. Click the request that updates the cart.
6. Check:
   - Status code (200, 400, 500?)
   - Payload tab: what the browser sent
   - Response tab: what the server returned
7. If the response total is already wrong, attach
   the response and mark the bug as server-side.
8. If the response is right but the screen is wrong,
   it is a front-end display bug.

-- Did the order save with the right discount?
SELECT o.order_id,
       o.subtotal,
       o.discount_total,
       o.grand_total,
       COUNT(oc.coupon_code) AS coupons_applied
FROM orders o
LEFT JOIN order_coupons oc ON oc.order_id = o.order_id
WHERE o.order_id = 50231
GROUP BY o.order_id, o.subtotal, o.discount_total, o.grand_total;

-- Find other orders with the same problem
SELECT o.order_id, o.subtotal, o.discount_total, o.grand_total
FROM orders o
WHERE o.grand_total <> o.subtotal - o.discount_total;

Q: A bug happens sometimes but not always. What do you do?

"First I would collect every case where it happened:
time, user, device, data and steps.

Then I would look for what those cases share.
Intermittent bugs usually depend on a hidden condition:
a particular data value, timing between two actions,
a cache, or a specific server in the environment.

I would change one variable at a time to test each guess,
and note it in the bug as I go, so even if I cannot fully
isolate it, the developer starts with what I ruled out."

ISOLATION LOG (example)

Bug: profile photo upload fails sometimes

Try | Changed variable           | Result
----|----------------------------|--------
 1  | Different browser          | Still fails sometimes
 2  | Different user account     | Still fails sometimes
 3  | Same JPG file, many times  | Always works
 4  | Different JPG files        | Fails on 2 of 5
 5  | Compare failing files      | Both above 4 MB
 6  | 3.9 MB vs 4.1 MB file      | 4.1 MB fails

Conclusion: upload limit is 4 MB on server,
but the UI says 5 MB. Not random at all.

WHEN YOU FIND A BUG, ASK BEFORE LOGGING

[ ] Does it happen every time? If not, when?
[ ] Different browser or device?
[ ] Different user, role or account?
[ ] Different data (values, size, special characters)?
[ ] Is the API response already wrong? (Network tab)
[ ] Is the stored data wrong? (SELECT query)
[ ] What is the smallest set of steps that shows it?
[ ] What did I rule out?

Add the answers to an 'Isolation' section in the report.

WEEKLY BUG HUNT (60 minutes)

1. Pick a feature on a practice or demo app.
2. Explore for 30 minutes with a charter, e.g.
   'Explore discount codes with unusual combinations'.
3. For each odd behaviour, spend 10 minutes narrowing
   it with the checklist.
4. Write one isolated bug report with steps,
   expected, actual and an isolation section.
5. Keep the best reports in a folder.

After a few weeks you have real interview stories.

CURIOSITY STORY FOR INTERVIEWS

First symptom I noticed: ______________________
Why I did not stop there: ______________________
What I changed, one at a time:
  1. ______________ -> ______________
  2. ______________ -> ______________
  3. ______________ -> ______________
Layer where it failed (UI / API / data): ________
What I ruled out: ______________________________
How my report helped the fix: __________________

Keep it under 90 seconds when spoken.