Why Daily Route Optimization Isn’t Enough for Mill Service Fleets

Every routing platform on the market can optimize tomorrow’s loads. You tell it what needs to go where, it builds the routes, and your drivers execute. For a straightforward delivery fleet, that’s enough.

But if 30 to 40 percent of your workload is intraday on-demand mill service — loads that shift hour by hour, schedules that get refined day-of, and bins that fill up faster than anyone predicted — daily optimization is like building a perfect plan that’s wrong by 10am.

The Intraday Reality

Here’s how it actually works for a fleet servicing mills. You get a schedule a week ahead. Then it gets refined. Then it changes again the day before. And then the morning crew goes out, inspects the bins, and calls dispatch to say they need two more trucks — now.

Your dispatcher planned for 10 trucks on that route today. Now they need 12. But the routing platform already locked tomorrow’s plan last night. There’s no mechanism to re-optimize mid-day. So Tyler pulls up the ELD, finds two drivers with available hours, and sends them manually. The route plan is now meaningless. The rest of the day is reactive.

This isn’t an edge case. For companies running bulk waste removal and raw material procurement from mills, this is 30 to 40 percent of the entire operation. Every single day.

“You plan the night before, and by 10am the plan is already wrong. If your routing platform can’t re-optimize intraday, you’re back to dispatching by phone.”

What Platforms Like Descartes Get Wrong

Descartes, to their credit, has a modern cloud-based routing product. It handles daily optimization well. But it can’t look at a longer time horizon. It can’t process your weekly order book and figure out that the flexible Friday delivery should be moved to Wednesday because the truck is already in the area.

And more critically, when three new mill pickups land at 10am, there’s no button to hit that says “re-optimize the rest of today.” There’s gymnastics. Manual overrides. Workarounds that defeat the purpose of having an optimizer in the first place.

The fundamental design assumption of most routing platforms is that tomorrow’s plan is final. For your operation, tomorrow’s plan is a starting point.

What Intraday Re-Optimization Actually Requires

• The system looks at the full order book — not just today, but the whole week — and optimizes across flexible and rigid time windows simultaneously.
• When new loads land mid-day, the system re-optimizes within minutes, recalculating every active route to find the least disruptive way to absorb the change.
• The re-optimization considers every variable in real time: driver location, hours of service, trailer configuration, wash requirements, and customer time windows.
• Dispatchers can override any assignment the algorithm suggests. The system recommends. Humans decide.

This isn’t science fiction. It’s a different design assumption: instead of building a static plan and executing it, you build a dynamic plan that continuously recalculates as reality changes.

Stat: Fleet operators with intraday re-optimization capability report 15–25% fewer manual dispatch interventions and measurably faster response times to urgent load requests compared to daily-only optimization.

The Cost of Not Having It

Every time your dispatcher manually re-routes a driver because the platform can’t handle a same-day change, you’re losing the optimization benefit on that load. Multiply that across 30 to 40 percent of your daily volume, and you’re operating a routing platform that’s only optimizing 60 percent of your workload. The rest is gut feel.

That’s not a technology investment. That’s an expensive half-measure.

Where to Go From Here

We talk about intraday routing and dynamic dispatch on our podcast and YouTube channel. If your routing platform is great at planning tomorrow but useless by mid-morning, that’s a design problem worth solving.

Book a call with the ScaleLabs team and bring your messiest Tuesday. We’ll show you what intraday re-optimization looks like on your actual load patterns.

‍