Optimizing Last-Mile Delivery Operations
Designed and built a batch delivery routing system that reduced per-delivery cost and improved rider utilization.
The engagement.
Logistics Provider (Confidential)
Software Development
8 months
The client operated a decentralized delivery network with high per-drop costs, inconsistent rider utilization, and poor visibility into daily route efficiency.
A batch delivery platform was designed and built, inspired by the firm Prelp product. The system grouped orders by zone and cut-off time, optimized rider routes, and provided real-time tracking for customers.
What was delivered.
Built custom batching and routing engine
Integrated with existing order management and rider apps
Delivered real-time customer notifications
Created operations dashboard for dispatch teams
28% lower
Cost per delivery45% higher
Rider utilization60% fewer
Customer complaintsThe engagement in detail.
The starting point
The client operated a last-mile delivery network serving a major city. Orders came in throughout the day. Riders were dispatched one order at a time. Per-drop costs were high because riders spent most of their time traveling between pickups, not delivering. Rider utilization was inconsistent: some riders were overloaded, others idle. There was no visibility into route efficiency. Dispatch was manual, based on phone calls and WhatsApp messages.
The design
The team designed a batch delivery platform inspired by the firm Prelp product. The core idea: group orders by zone and cut-off time, then assign a single rider to deliver the entire batch along an optimized route. This reduced travel time between drops, increased rider utilization, and lowered per-delivery cost. The system would integrate with the client existing order management system and rider app.
The build
The batching engine was built first. It consumed orders from the order management system, grouped them by geographic zone and delivery cut-off time, and produced optimized routes. The routing algorithm considered rider capacity, traffic patterns, and delivery windows. The engine was tested against historical order data to validate the batching logic before any live deployment. The operations dashboard was built next, giving dispatch teams real-time visibility into batch status, rider location, and delivery confirmation.
The integration
The platform was integrated with the client existing systems. Orders flowed in from the order management system. Routes were pushed to the rider app. Delivery confirmations flowed back. Customer notifications were automated: order received, order batched, rider dispatched, delivery confirmed. The integration was done incrementally, with a parallel run period where both the old and new systems operated simultaneously.
The results
After full deployment, per-delivery cost dropped significantly. Rider utilization improved as riders spent less time traveling and more time delivering. Customer complaints dropped as real-time tracking and automated notifications replaced phone-call-based status updates. The dispatch team shifted from manual coordination to exception handling. The standard did not move with conditions.