DPD Croatia:18% Shorter Routes in Zagreb with The Routing LabDPD Croatia:18% Shorter Routes in Zagreb withThe Routing Lab
The Challenge
Can AI-driven planning outperform traditional last-mile routing in the real world?
DPD Croatia wanted to verify whether AI-assisted planning could reduce total distance, keep service levels intact, and improve ETA reliability.
Two levers matter most in last-mile optimisation:
- Accurate geocoding – every incorrect coordinate adds wasted kilometres.
- Constraint-aware planning – routes must minimise distance while respecting time windows, capacities and service rules.
DPD's goal was to confirm whether The Routing Lab could outperform traditional planning tools across both dimensions, producing measurable savings and smoother daily operations.
What We Implemented
A two-pronged approach focused on geocoding precision and route efficiency.
Geocoding Quality Benchmark
We processed 6,060 addresses through multiple geocoders and scored them using The Routing Lab's accuracy engine. Across four key distance bands (all <200 m), our approach outperformed alternatives by +2.9 to +5.6 percentage points. Higher geocoding precision reduces wrong turns, detours and planner guesswork, directly leading to shorter, more reliable routes.
Route Optimisation Simulation
Using The Routing Lab's optimisation models, we automatically planned all Zagreb routes for 7–11 November 2022, optimising for minimum distance while preserving operational constraints. We then compared the optimised totals to DPD's historical route data from the same week.
Results That Delivered
Our optimisation engine reduced total distance by 18.3% while maintaining service quality – demonstrating clear gains in efficiency and reliability.
Optimised distance:
Optimised total distance
Historical distance:
Previous week's performance
Reduction:
Weekly distance reduction
Fewer kilometres mean lower fuel consumption, reduced vehicle wear, and more predictable ETAs, while planners spend less time resolving edge cases.
Real Business Impact
Scaling this optimisation across the full courier network unlocks significant savings and sustained performance improvements:
Monthly Savings:
Fuel, driver hours, vehicle wear
Annual savings:
Conservative 16% average reduction
Reduced CO₂ per year:
Reduced fuel consumption
For a courier network running ~5,000 routes/month, a 16% distance reduction (conservative vs. the 18.3% above) strengthens both profitability and sustainability, lowering cost per route whilst reducing emissions.
Client's perspective
The Routing Lab's advanced route simulation led to a mileage reduction of over 18% across selected routes. Their technology consistently outperformed our existing planning, helping us significantly optimise our delivery network. The collaboration was highly professional and effective.”
Why It Worked
A data-driven planning engine designed for real drivers, real cities and real-world logistics.
Precision Geocoding
The Routing Lab's engine converts raw addresses into high-accuracy GPS points, even with incomplete or ambiguous data, driving fewer detours and more reliable routing.
AI-Assisted Optimisation
Our models plan routes that balance distance, capacity and delivery windows, respecting all operational constraints, and consistently reducing kilometres without compromising service.
Continuous Learning
Every route feeds back live GPS and stop-duration data, enabling the system to refine ETAs and planning accuracy over time.
Operational Fit
Seamlessly integrates with existing workflows, enabling rapid re-planning and real-time adjustments without disrupting drivers or dispatchers.
Ready to Optimise Your Network?
The Routing Lab helps courier networks improve geocoding, optimise daily routes and reduce operational costs at scale.
If you'd like to explore route optimisation or run a pilot on your own network, we'd be happy to help.
Prefer email?
Contact us directly at [email protected]