Route optimization: save 18% on fuel

Fuel accounts for 30% to 40% of the operating cost of a delivery fleet in Spain. And most fleets manage it with two tools: Google Maps when there’s time and “the usual route” when there isn’t. The result is a monthly leak that’s rarely quantified.

In this article we explain how AI route optimization typically reduces fuel consumption by 15% to 20%, what factors the algorithm considers beyond kilometers, and what ROI to expect in a real fleet of 50 vans.

Why an algorithm beats a human planner

A good route planner with 20 years of experience has three operational limitations compared with an AI algorithm:

1. Cognitive capacity: they can’t simultaneously process 50 vans, 600 stops and 200 changing constraints.
2. Historical memory: they don’t remember the actual average loading time at customer #1,234 on Tuesdays at 11:00.
3. Real-time adaptation: when a technician is running 40 minutes late, they can’t recalculate the rest of the day’s stops in seconds.

AI does exactly those three things at once, every five minutes, across the entire fleet. The human planner is freed up to handle exceptions, not routines.

What factors an AI routing algorithm considers

Serious optimization goes far beyond “the shortest route from A to B.” The factors it weighs are:

Vehicle data

• Vehicle type (light van, rigid truck, articulated truck).
• Height, weight and GVW restrictions.
• Average consumption by driving type.
• Current battery status and fuel level.

Trip data

• Historical traffic by time slot and day of the week.
• Real-time traffic (incidents, roadworks, accidents).
• Weather forecast (heavy rain, wind, snow).
• Time-based restrictions (low-emission zones, or ZBE in Spain; pedestrian streets at certain hours; closures for events).

Customer or stop data

• Time window committed to the customer.
• Historical average loading/unloading time at that customer.
• Operational priority (premium customer, stop with an SLA).
• Need for specific equipment (tail lift, refrigeration).

Driver data

• Accumulated driving time (compliance with Regulation (EC) 561/2006).
• Pending break time.
• Specific qualifications (C+E license, ADR).

Combining all of this produces a route that isn’t the shortest in kilometers, but is the most profitable in euros and the most realistic in terms of on-time delivery.

Real case: a fleet of 50 urban delivery vans

Data from a real, anonymized operation — a B2B parcel delivery company in a metropolitan area, 50 Renault Master 3.5-ton vans and similar, 600-700 stops a day.

Starting point (before implementing AI)

• Average consumption: 8.2 L/100 km.
• Jobs completed per van per day: 14.
• Average workday length: 9 hours 20 minutes.
• Stops outside the time window: 18% of the total.
• Monthly fuel cost: €41,000.

After 6 months of AI route optimization

• Average consumption: 6.7 L/100 km (-18%).
• Jobs completed per van per day: 15.7 (+12%).
• Average workday length: 8 hours 45 minutes (-6%).
• Stops outside the time window: 7% (-61% relative).
• Monthly fuel cost: €33,620 (savings of €7,380/month).

Calculating the return

• Fuel savings over 12 months: €88,560.
• Overtime savings from shorter workdays: €22,000.
• Extra billable jobs: +12% capacity without adding headcount.
• Hardware + subscription for 50 vehicles over 12 months: an estimated €18,000 or so.
• Net return in the first year: over €90,000, not counting the value of the additional capacity.

How to roll it out without stopping operations

A legitimate concern when changing routing systems is not wanting to halt operations. The realistic plan is:

1. Hardware installation in a first wave of 5-10 pilot vehicles (15-40 minutes per vehicle).
2. Two weeks of baseline telemetry without changing operations: the algorithm learns the actual times.
3. Gradual activation of optimization in the pilot wave, adjusting thresholds with the fleet manager.
4. Rollout to the remaining vehicles once the pilot pattern is validated, in weekly waves.
5. Monthly review of actual savings vs. baseline and configuration adjustments.

Operations never stop. Integration with the customer’s order system (ERP, TMS, Excel with an API) is usually done in parallel during phases 2-3.

Common mistakes that destroy ROI

Even if the algorithm is good, three operational mistakes can wipe out the savings:

• Not updating the actual loading times at each customer and leaving the system’s default values. The algorithm’s accuracy drops by half.
• Letting drivers ignore the proposed route without explaining the decision. Without feedback, the system doesn’t learn.
• Not measuring the prior baseline properly: if you don’t know what you were consuming before, you can’t prove the savings.

A good provider helps you avoid them from the start. If it doesn’t, it’s selling software, not results.

Less visible secondary benefits

Beyond fuel, serious optimization brings side effects that operations teams value:

• Less stress for the team: more predictable workdays, fewer unexpected emergencies.
• Better compliance with customer time windows: fewer complaints, better NPS.
• Lower CO₂ emissions per kilometer driven: an increasingly important argument in public tenders and RFPs.
• Real visibility for management into what happens every day without having to ask.

These effects don’t show up directly on the P&L, but they’re usually the reason management renews the contract at 18 months.

Conclusion

AI route optimization is one of the most measurable and fastest ROI levers in a delivery fleet: typical savings of 15-20% on fuel, +10-15% operational capacity without adding headcount and a significant reduction in stops outside the time window. A fleet of 50 vans can recover its entire first-year investment in under 4 months.

The condition is having the right hardware, software with real AI algorithms (not static rules) and the operational discipline to feed the system clean data. We also explain it from an SMB perspective in fleet management software for SMBs and from a technical perspective in Trackia’s AI capabilities.

Want us to calculate how much fuel your specific fleet would save? Request a personalized estimate →

Within 24 hours we’ll send you a 12-month savings projection based on your industry, size and operating pattern, with no commitment. If you run an SMB with a sales fleet, you may also be interested in our SMB industry page.