[ PRODUCT 03 — LOGISTICS ]

Delivery route planner

Route planning that replaced driver guesswork for a fleet covering three districts.

Python FastAPI PostgreSQL React Docker

[ THE PROBLEM ]

Every driver planned their own route from memory each morning.

[Placeholder client] ran 18 vehicles across three districts. Route order was decided by whoever loaded the van, which meant vehicles frequently crossed paths and doubled back.

Nobody could say why fuel costs rose every quarter, and customers were given delivery windows that were essentially guesses.

[ WHAT WE BUILT ]

  • Automatic route ordering for each vehicle every morning
  • Load balancing so no driver finishes three hours before another
  • A simple driver app with the next stop and one-tap navigation
  • Live delivery status visible to the office team
  • Automatic SMS to customers with a realistic arrival window
  • Weekly reports comparing planned versus actual routes

[ HOW WE WORKED ]

We started by analysing six months of their existing delivery logs. That alone showed which routes were consistently overloaded, before any software was written.

The model uses road distances and historic traffic patterns rather than straight-line distance, which matters considerably on rural routes.

The driver app was deliberately kept to two screens. Drivers were not going to adopt anything more complicated while working.

[ RESULTS — PLACEHOLDER FIGURES ]

What changed

25%Lower fuel spend
95%Deliveries inside the promised window
18Vehicles planned each morning
[ NEXT PRODUCT ]

Restaurant ordering system →

Need something

similar?

Start the conversation

We will tell you honestly what it takes