SAMPLE REPORT

Website and web application

The report below comes from the pre-release test of a B2C e-commerce site. Click a coloured square to jump to its finding.

Cases tested: 528
Total findings: 38
Release blockers: 4

Test summary

528 cases were tested across 3 platforms over seven business days. 38 findings were reported; 4 of them were release blockers and we advised against shipping until they were fixed.

Cases tested: 528
Total findings: 38
Release blockers: 4
Closed on re-test: 35

Severity distribution

4Critical

9High

16Medium

9Low

CLIENT: Anonymous · B2C e-commerce
PACKAGE: Release
DATES: 12–20 January 2026
ENVIRONMENT: staging.v4.2.0

Coverage matrix

Each square is one test case. Coloured squares carry findings; click one to open it.

PassedCriticalHighMediumLow

Finding cards

Five of the 38 findings in the report. Each card is exactly the format you'll see in yours.

5 / 5 showing

CriticalCNRY-104Checkoutchrome 141 · desktop · throttled 3G

Double-tapping pay creates two separate orders

Steps to reproduce

Add a product to the basket and go to checkout.
Throttle the network to "Slow 3G".
Click "Complete order" twice, 300 ms apart.
Open the order list.

Expected: One order is created; the second request is ignored.
Actual: Two separate orders are created and the card is charged twice.
Impact: Direct financial impact plus manual refund work. Reproduced 5/5 times on a slow connection.
Evidence: video-104.mp4 · har-104.json

CriticalCNRY-111Account securitychrome 141 · firefox 145 · ios 26

Sessions on other devices survive a password change

Steps to reproduce

Sign in to the same account in two different browsers.
Change the password in the first browser.
Refresh the second browser and open account settings.

Expected: The second session is revoked and the user is sent to the login screen.
Actual: The second session stays live and account settings remain editable.
Impact: A stolen session cookie keeps working even after a password change, which makes the account recovery flow useless.
Evidence: video-111.mp4 · session-log.txt

HighCNRY-127Product listingchrome 141 · 4G · mid-range android

Category list takes 6.2 seconds with 1,400 products

Steps to reproduce

Open the "Electronics" category, which holds 1,400 products.
Watch the /api/products request in the network tab.
Measure LCP.

Expected: LCP stays under 2.5 seconds and the list arrives paginated.
Actual: The full collection arrives in one 3.1 MB request; LCP is 6.2 seconds.
Impact: Around 30% of mobile users leave before the list renders. Needs server-side pagination.
Evidence: lighthouse-127.html · waterfall-127.png

MediumCNRY-133Bookingios 26 · safari · iphone 15

Date field resets when the keyboard closes on iOS Safari

Steps to reproduce

Open the booking form and tap the date field.
Pick a date, then tap an empty area to dismiss the keyboard.
Look at the date field.

Expected: The chosen date stays in the field.
Actual: The field clears and the user has to redo the step.
Impact: Raises abandonment in the mobile booking flow. Only reproduces on iOS Safari.
Evidence: video-133.mp4

HighCNRY-141Sign-upvoiceover · macos 26 · safari

Form errors are never announced to screen readers

Steps to reproduce

Fill in the sign-up form with VoiceOver running.
Leave the email field invalid and submit.
Listen for announcements.

Expected: The error is announced and focus moves to the first invalid field.
Actual: The error appears visually only; nothing is announced.
Impact: Fails WCAG 2.2 AA 3.3.1. A screen reader user cannot complete sign-up.
Evidence: voiceover-141.m4a · axe-report.json

The report keeps living after handover

The web report is a link, not a dead PDF. Its status updates as you mark fixes, and after the re-test round every finding's outcome shows up in the same place.

A report for another product type

Mobile appiOS and Android, on real devices and real connections.Desktop applicationWindows and macOS, including install, update and offline behaviour.API and backend serviceREST and GraphQL: contract, permissions, boundaries and concurrency.

Don't test the next release alongside your users.

Start a test Read a sample report