What should regression testing include?

The core user journeys, everything that touches the changed code, data or config, and bugs fixed in recent releases, plus a cross-browser and device pass for UI changes.

What is the difference between regression and smoke testing?

A smoke test is a quick check that the most important flows work at all. Regression testing goes wider, re-checking areas that worked before and might be affected by the change.

How do I report a regression?

Include the last known good version, the first broken version, the suspected change, steps, expected and actual results, and whether to roll back.

Regression Testing Checklist Template (Free)

Quick answer

Re-test three things on every release: the core journeys (sign in, the main job, payment), anything touching the changed code, and bugs fixed in the last few releases. Then run the browser and device pass below. When something that used to work breaks, file it with the last known good version: that one field cuts debugging time the most.

Choose what to re-test

Core journeys, every time. Keep them as named test cases (see the test case template).
Blast radius. For each merged change, list the screens and APIs that use the same code, data or config.
Recent fixes. Re-run the steps from bugs closed in the last two or three releases. Fixed bugs come back more often than new ones appear.
Automate the boring repeats and spend manual time on what changed.

R8 Cross-browser and device pass

When: For UI changes before release. Owner: QA · Time: 20–40 min

R8 checklist (Markdown tick boxes)Paste into a ticket or PR

# R8 Cross-browser and device pass
When: For UI changes before release.
Owner: QA · Time: 20–40 min

Release: ______  Date: ______  Signed off by: ______

## Browsers (latest versions)
- [ ] Chrome on Windows
- [ ] Edge on Windows
- [ ] Safari on macOS
- [ ] Firefox on any desktop
- [ ] Safari on iPhone
- [ ] Chrome on Android

## Sizes
- [ ] 320–375 px (small phone)
- [ ] 768 px (tablet portrait)
- [ ] 1280–1440 px (laptop)
- [ ] 1920 px+ (large desktop)
- [ ] Browser zoom at 200%

## Each changed screen
- [ ] Layout holds and nothing overlaps
- [ ] Forms submit, with validation messages visible
- [ ] Hover-only actions also work by tap and keyboard
- [ ] Long text, long names and empty states look right

A starting browser and device matrix

Browser	OS	Device
Chrome (latest)	Windows 11	Desktop
Edge (latest)	Windows 11	Desktop
Safari (latest)	macOS	Desktop
Firefox (latest)	Windows or macOS	Desktop
Chrome (latest - 2)	Windows 10	Desktop
Safari	iOS (latest)	iPhone
Safari	iOS (latest - 1)	iPhone, small screen
Chrome	Android (latest)	Phone
Samsung Internet	Android	Phone
Safari	iPadOS	Tablet

Regression bug template (T6)

When you know it worked before. The last known good version is the most valuable field: it cuts debugging time.

regression-bug.md (template T6)

### Last known good version or date (required)


### First broken version or date (required)


### Suspected change (PR, commit, flag)
<!-- Optional. Link if you have an idea. -->


### Steps to reproduce (required)
<!-- Start from a known state. One action per step. -->
1. Logged in as 
2. 
3. 

### Expected result (required)
<!-- What should happen. Link the spec or last working version if you can. -->


### Actual result (required)
<!-- What happens instead. Paste the exact error text. -->


### Should we roll back? (required)
<!-- Tick one -->
- [ ] Yes, roll back now
- [ ] No, fix forward
- [ ] Not sure, decide in triage

### Severity (see the severity guide) (required)
<!-- Triage may change it. -->
<!-- Tick one -->
- [ ] S1 Critical: data lost/exposed or core flow blocked, no workaround
- [ ] S2 Major: core feature wrong or failing for a group
- [ ] S3 Minor: works with an easy workaround
- [ ] S4 Cosmetic: looks or reads wrong only
- [ ] Not sure (triage will set it)

### Evidence
<!-- Drag in a screen recording or screenshot. Add console errors, request IDs or log links. -->

Record the video part

Record the broken behaviour, and if you can, the same steps on the last good version for comparison. Bobbin is a free Chrome extension that records your screen, a window or a tab with your microphone and an optional webcam bubble. No account, no watermark, no time limit, and nothing is uploaded: the file stays on your computer.

Click the Bobbin icon in Chrome's toolbar and choose Record screen, window or tab.
Pick your microphone and turn on the camera if you want the webcam bubble.
Press Start recording, choose what to share in Chrome's picker, and wait for the 3-second countdown.
Pause with Alt+Shift+P, stop with Alt+Shift+S, then press Download to save the file.

To keep other tabs and notifications out of it, share just the Chrome tab (the webcam bubble appears only in screen and window recordings). Bobbin saves MP4 where your browser supports it (Chrome 126 or newer), WebM otherwise, and tells you which it used.

File size: our 60-minute 1080p test came to about 246 MB (roughly 4 MB a minute) on a simple synthetic screen; busy real screens can be larger, so a 1–2 minute clip is usually small enough to attach.

From The Bug Report & QA Kit

The kit has 9 release checklists, the full browser/device matrix as a spreadsheet and 25 test cases. The full kit: 10 bug templates for GitHub, Jira, Linear and Markdown, 9 release checklists (126 checks), the severity guide, 8 bug-video scripts and trackers.

See the QA kit ($15, one-time) PDF + ZIP · 7-day refund · more free samples on the page

Regression testing checklist (free template)

Quick answer

Choose what to re-test

R8 Cross-browser and device pass

A starting browser and device matrix

Regression bug template (T6)

Record the video part

Frequently asked questions

Keep reading

Record your screen and camera. Keep the file.