Last Mile Route Optimization: Why the Final Stretch Costs the Most

What "Last Mile" Means

Last mile delivery is the final leg of the supply chain — moving goods from a distribution hub to the customer's door. The term comes from telecommunications, where connecting the last stretch to individual homes was disproportionately difficult compared to the backbone network.

In logistics, the same principle applies. Long-haul freight moves thousands of packages on optimized highway routes. Last mile delivery fans those packages out into hundreds of individual stops across residential streets, apartment buildings, and businesses.

Why It Costs So Much

Between 2018 and 2023, the share of last mile delivery out of total shipping costs grew from 41% to 53%, according to Statista. It's now the single most expensive segment of the delivery chain.

The economics are structural:

• Low drop density. A long-haul truck moves thousands of parcels hundreds of miles. A delivery van makes 100–200 stops across 80–120 km of urban driving. The packages-per-kilometer ratio is dramatically worse.
• Time per stop. Each delivery means parking, walking to the door, waiting for the recipient (or finding a safe spot), and returning. That's 2–5 minutes per stop, plus the overhead of navigating to the next one.
• Failed deliveries. When nobody's home, the package goes back to the depot and the route repeats. Industry estimates put first-attempt failure rates at 6–12%, with each re-delivery costing $12–17.
• Customer expectations. Free shipping, same-day delivery, and narrow time windows (driven by Amazon Prime's influence) compress route flexibility and push up costs.

The Optimization Lever

Route optimization is the most direct way to attack last mile costs because it improves the fundamental metric: stops per kilometer driven.

Density-based clustering

Optimization algorithms group nearby deliveries together, maximizing drops per route-mile. Sophisticated solvers account not just for geography but also compatible time windows — two stops 200 meters apart are useless together if one requires morning delivery and the other afternoon.

Dynamic re-routing

Static plans break on contact with reality. New orders, cancellations, traffic, and delays all disrupt the original route. Real-time re-optimization adjusts mid-day, keeping routes efficient as conditions change.

Reducing failed deliveries

Automated ETA notifications give customers accurate arrival windows, so they know when to be available. This alone can reduce failed deliveries by 20–30%, according to multiple logistics providers — eliminating the cost and emissions of re-delivery attempts.

Load sequencing

Planning which packages go on which vehicle, and in what loading order (last stop loaded first), reduces time at each stop and prevents the driver from digging through the van.

The Scale of Impact

UPS's ORION system saves approximately 100 million miles per year — roughly 10 million gallons of fuel and 100,000 metric tons of CO2 annually. They've stated that one fewer mile per driver per day saves $50 million per year.

For smaller operations, the proportional impact is similar: 20–40% reduction in driving distance, 15–25% more deliveries per driver per day. The math scales down as cleanly as it scales up.

The Sustainability Angle

The World Economic Forum projected in 2020 that without intervention, delivery vehicles in the top 100 cities would increase by 36% by 2030, driving a 32% increase in emissions. Route optimization is one of the most cost-effective ways to cut delivery emissions because it requires no new vehicles — just better planning.

Every unnecessary kilometer eliminated is fuel not burned and CO2 not emitted. For companies facing sustainability reporting requirements or operating in low-emission zones (increasingly common in European cities), optimized routing is both an environmental and compliance strategy.

The Last Mile for Service Businesses

"Last mile" isn't just about parcels. Field service companies — technicians, inspectors, installers — face an identical last-mile challenge. Each customer visit has the same dynamics: travel time between stops, time windows, failed appointments, and the need to maximize productive visits per day.

The optimization approach is the same. The constraints are even more complex (skill requirements, equipment, variable appointment durations), but the payoff follows the same curve.