
- Client
- Peddler
- Industry
- Logistics · Final-mile delivery
- Engagement
- 2022 – 2026 · 1.2 years
- QA team
- 1 full-stack QA engineer
- Platforms
- Desktop Web · Mobile Web · PWA · API · Payment Gateways
- Manual Testing
- UI/UX Testing
- Cross-browser Testing
- Usability Testing
- Performance Testing
ZenHub
Postman
GitHub
Google Chat
Figma
Mollie
JMeter
BrowserStack
About the project
Peddler is a green logistics and last-mile delivery company operating primarily in the Netherlands and wider Europe. Cargo bikes and light electric vehicles run from city micro-hubs, delivering parcels same-day for retailers, brands, 3PLs and international marketplaces.
Behind the fleet is software: shipper booking by web, API and bulk upload, rate calculation, Mollie payments, live track-and-trace, customer notifications, and a PWA used by hub teams and riders. Marketplace volume means those systems see sharp peaks, with parcels arriving in large batches rather than one at a time.
QAonTOP provided one full-stack QA engineer across 1.2 years of active engagement between 2022 and 2026, covering key build and scale-up phases, working in Peddler’s ZenHub board and alongside design in Figma.
What made this hard
In same-day logistics, a software defect turns into a missed delivery within hours. Quality had to cover the whole parcel lifecycle — with one engineer.
- CHALLENGE 01
Delivery windows leave no slack
Same-day cut-offs, time slots and postcode coverage rules decide whether a parcel can be booked. One wrong rule means missed deliveries, not just a ticket in the backlog.
- CHALLENGE 02
Payments in the Dutch market
Checkout runs through Mollie with iDEAL, cards and refunds. Failed, cancelled and expired payments all needed to leave bookings in a clean, recoverable state.
- CHALLENGE 03
Marketplace peaks
Batches of international marketplace orders arrive at once. Booking, label and tracking endpoints had to stay responsive at those volumes.
- CHALLENGE 04
Many screens, many users
Shippers on desktop, recipients tracking on phones, hub staff and riders on a PWA out on the street — often on older devices and patchy connections.
How QAonTOP tested Peddler
-
01
Learn the flows, write them down
Mapped the parcel lifecycle from booking to proof of delivery and turned it into a risk-based test plan in ZenHub, so every story carried acceptance criteria QA could test against.
Tools ZenHub · GitHub
-
02
Design-to-build checks
Reviewed designs before build and compared each implemented screen with Figma, catching layout, copy and state issues while they were still cheap to fix.
Tools Figma
-
03
Payments, end to end
Covered Mollie test-mode scenarios — paid, failed, expired, cancelled and refunded — and verified the booking, invoice and notification outcome for each.
Tools Mollie · Postman
-
04
APIs and cross-browser
Postman collections for booking, rates, labels and tracking. BrowserStack runs across Chrome, Safari, Firefox and Edge on desktop, Android and iOS, including PWA install and offline behaviour.
Tools Postman · BrowserStack
-
05
Load before the peaks
JMeter scenarios simulating marketplace batch imports and tracking-page spikes, run ahead of busy periods, with results shared in Google Chat for fast decisions.
Tools JMeter · Google Chat
The engagement in numbers
Full-stack QA engineer across web, PWA, API and payments
Test cases written across the parcel lifecycle
Defects found and verified before release
Mollie payment outcomes validated end to end
Browser and device combinations on BrowserStack
Faster response on peak-load paths after tuning
Before and after the engagement
| Area | Before QAonTOP | With QAonTOP |
|---|---|---|
| Requirements | Stories written without testable acceptance criteria; QA started after the build was done. | Acceptance criteria agreed in ZenHub before development, with QA reviewing designs in Figma. |
| Payments | Only the happy path tested; failed or cancelled iDEAL payments could leave bookings half-created. | Every Mollie outcome covered, with bookings, invoices and notifications verified for each. |
| Peak load | Slow, unresponsive pages when marketplace batches and sale traffic arrived together. | JMeter-tested peaks, a queueing approach for heavy traffic and roughly 70% faster response on the busiest paths. |
| Cross-browser & PWA | Safari and older Android issues reported by users in the field. | Planned BrowserStack coverage on real devices, including PWA install and offline scenarios. |
| Release confidence | Hotfixes after a release were routine. | A regression checklist per release, severity-based bug flow and QA sign-off on release notes. |
What we covered
Manual Testing
Functional coverage of booking, rates, labels, tracking, returns and hub operations.
UI/UX Testing
Design-to-build comparison against Figma for layout, copy, component states and responsiveness.
Cross-browser Testing
BrowserStack runs across desktop and mobile browsers, plus PWA behaviour on real devices.
Usability Testing
Task walkthroughs for shippers, recipients and hub staff, with findings fed back to design.
Performance Testing
JMeter load and spike tests on booking, tracking and marketplace import endpoints.