If you manage a fleet – whether it’s car sharing, rentals, delivery vans, or micromobility – you know one painful truth: demand never matches supply perfectly.
Too many vehicles sitting idle? You’re bleeding money.
Too few available at peak times? You’re losing customers.
That’s where demand planning and forecasting come in. But doing it right is harder than it sounds.
Why Demand Forecasting Is Tricky in Mobility
Unlike manufacturing or retail, mobility demand is:
- Highly time-dependent → weekday rush hour vs. weekend leisure trips.
- Location-sensitive → one neighborhood may be oversupplied while another is starved.
- Weather-driven → sunny Saturday = bike and scooter spike; rain = surge in car rentals.
- Event-driven → concerts, football matches, or even strikes can make demand skyrocket in pockets of the city.
Traditional forecasting methods (spreadsheets + historical averages) just don’t cut it anymore. They fail to capture these fast-moving variables and leave operators constantly in reactive mode.
Real-World Examples
- Car rental companies often struggle at airports: too many vehicles booked for morning arrivals, too few for evening demand. The cost of repositioning vehicles between branches eats into margins.
- Micromobility operators like scooter-sharing platforms see sudden spikes during events. Without accurate demand planning, rebalancing trucks end up shuttling scooters back and forth all night.
- Delivery fleets (think last-mile logistics) face seasonal surges – Christmas, Black Friday, or local holidays – that require months of preparation but still often lead to bottlenecks.
Smarter Approaches Emerging
Operators are experimenting with:
- Predictive analytics → combining weather data, city events, and historical usage.
- AI simulations → testing “what-if” scenarios like: What happens if I add 20 EVs to this district next month?
- Dynamic rebalancing → not just forecasting demand, but acting on it in real time by moving assets ahead of predicted surges.
Where AI Is Changing the Game
One example is SWITCH, which has built tools specifically for this.
- Urbiverse platform runs city-scale simulations, helping operators plan fleet size and positioning weeks or months ahead.
- Urban CoPilot handles day-to-day fleet orchestration, using demand signals to suggest where vehicles should be repositioned in real time.
- SWITCH AI Agent integrates these insights into a conversational interface – fleet managers can literally ask, “Where should I add more cars this weekend?” and get a data-backed answer.
It’s not about “set and forget.” It’s about creating a feedback loop where demand forecasts inform operations, and operations feed back into better forecasts.
Want to discover more?
The Bottom Line
Demand planning in fleet management isn’t just about prediction – it’s about prediction + action. The winners in this space will be the operators who:
- Use external data (weather, events, traffic) alongside internal usage patterns.
- Run simulations to stress-test strategies before deploying vehicles.
- Automate the feedback loop so forecasts are continuously improving.
