Amazon Case Study
The Challenge
Amazon. Returns tracking and scanning. Users: Management, warehouse operators, Admin Staff.
Amazon needed an urgent change to its UK returns systems at a time when internal software teams were fully committed. A trusted partner, Prestek, recommended RD Research. After a rapid technical review Amazon was satisfied that, although we are a small company, we had the expertise to deliver.
The brief was to develop a new returns system in one month that could ingest daily truck manifests, scan tens of thousands of items as they came off articulated lorries, and route them to the correct sort bays. Accuracy was vital. Operators needed product images on their handhelds to visually confirm matches and the system had to integrate with Amazon’s APIs and database.
Hundreds of pallets and thousands of items needed to be processed every hour. Every scan had to reconcile to the manifest, create a clear chain and produce reliable reporting for onward resale.
The Development Process
We opened with a rapid discovery on site at one the larger warehouses to map the physical flow from trailer to sort bay, the data fields in manifests, and the barcode formats used. We were told what handheld scanners we had to write the software for.
From that, we produced a tight specification covering handheld screens, scan validation rules, sort logic, exception handling and the data exchange with Amazon.
Wireframes for the handheld app and supervisor console were reviewed within days, so management could confirm the flow before we wrote code. We agreed performance budgets for scans per minute, offline behaviour and image look-ups, along with a simple rollout plan that avoided disruption to live operations.
Build and test ran in parallel. We implemented manifest ingestion, item and pallet scanning, image display for operator confirmation, and automated bay assignment. Exceptions such as missing items or misclassifications triggered prompts and holds.
API integration to Amazon completed the loop so the warehouse could work from live data.
Pilot test proved throughput and accuracy. We trained crews, monitored the first days closely.
The entire programme was delivered several days early and on budget.
Key features
- Real time manifest ingestion, matching scans to LIQN and SPN references
- Handheld workflows with instant barcode validation and product images for visual checks
- Automated routing to sort bays, with pallet creation and unique pallet IDs
- Exception handling with operator overrides, supervisor review and full audit trail
- Live throughput dashboards and shift reports for managers
- Clean API integration to Amazon for compliance and end to end reconciliation
Outcome
The system went live on time and immediately enabled depots to process and resell high volumes of returns with confidence. Scans reconciled to manifests, operators moved faster with on-device images, and supervisors had clear visibility of pallets, bays and exceptions. Once Amazon’s internal replacement was ready it took over, and RD Research was thanked for delivering a mission critical platform under extreme time pressure.
“RD Research delivered a project for us on a very tight deadline and it worked perfectly first time. They did a great job.”
...