The moment we realised perfect routes don't survive contact with traffic

The problem with 'plan once, execute always'

Every delivery app tells you the same story: optimise your route before you leave. It's sound advice. But it assumes the world stays still. It doesn't.

Dispatchers and drivers know this better than anyone. You plan a route using live traffic data, and 20 minutes later, a crash happens two junctions ahead. Your second delivery is now a crawl. The ETA you gave the customer is rubbish. You're burning fuel idling. Your whole day shifts.

We spent months watching drivers work. The ones using spreadsheets. The ones using maps. The ones using other routing apps. They all had the same coping mechanism: they'd check their phone every few minutes. 'Am I still on track? Is there a faster way?' It's not really driving. It's constant small anxiety with your device.

The standard answer in the industry was a bit glib: 'Just replan if something changes.' But replanning means stopping, opening the app, letting it calculate, waiting for a new route, then switching. For a driver in motion, especially on a bike or in a van in busy traffic, that's a barrier.

Why we chose continuous over reactive

Most routing apps refresh their data once. Maybe twice. DropPilot pulls live traffic from Google Directions API continuously. Not every second, but often enough that if conditions shift, we know.

The key insight was this: the driver shouldn't have to ask. If we're already pulling fresh traffic data, and the best route changes, why make the driver make a decision? We could just tell them. Better still, we could reroute them automatically if they've drifted from the plan.

That's the feature we call smart rerouting on deviation. Here's what happens in practice. A driver is heading to their next stop. Traffic suddenly builds on the planned route. Our system detects two things at once: the new traffic, and the fact that the driver's current position no longer aligns with the fastest path forward. So we recalculate. If there's a meaningfully better route, we push it to their app. No tap required. They see the new directions appear, and they follow them.

It sounds simple. Building it meant deciding what 'better' actually means. Faster? Shorter? Less likely to hit more traffic in the next 30 minutes? We went with what drivers care about most: lower total time to their next stop, factoring in what traffic looks like both right now and in the immediate future.

The ETA that doesn't embarrass you

Here's something nobody talks about in logistics: the emotional weight of a bad ETA. You promise a customer you'll be there at 3.15pm. You're certain. But then you hit something unexpected, and suddenly you're late. It feels like you've already let them down before you even arrive.

Continuous ETA refresh solves part of that. Because we're pulling traffic data every few minutes, your ETA updates in real time. If conditions improve, it gets earlier. If traffic appears, it gets later. The customer sees an honest picture, not a guess made 30 minutes ago when the roads were clear.

For dispatchers managing multiple drivers, this is even more valuable. You're not getting a cascade of 'I'm going to be late' messages at 2.58pm. You're seeing live ETAs across your whole fleet. You can communicate proactively. You can reassign stops if someone's running significantly behind. You're managing the day, not just reacting to surprises.

We tested this with a food delivery service in London running about 40 stops a day across their team. In week one, they logged every conversation where a customer complained about lateness. Then we rolled out continuous ETA refresh and smart rerouting. The next week, complaints dropped by a third. Not because drivers were faster. But because customers knew what was actually happening.

What we learnt about drivers and decisions

We almost made rerouting opt-in. You'd get a notification: 'New route available. Tap to accept.' It made sense on paper, didn't overwhelm them, gave them control.

Then we watched a driver in our beta ignore three rerouting suggestions in a single day. Each one would have saved them 8 to 12 minutes. But they were concentrating on the road, didn't see the notification, or saw it and thought 'I'll deal with it later.' By the time they decided to engage, the window had closed. The suggestion was no longer the best path.

That's when we realised automatic rerouting isn't about removing driver control. It's about respecting their time. If I'm in the middle of navigating traffic, I don't want to open my phone to say yes to a better route. I want to just go the better route. The driver stays in charge; they can always manually reroute if they disagree with what the system suggests. But the default is that we handle the small optimisations.

For couriers, field service technicians, and anyone managing their own deliveries, this changes how they think about their day. Instead of planning at 8am and hoping nothing changes, they're working with a system that's actively watching conditions and keeping them on the fastest path. It's the difference between executing a plan and executing against a moving target.

Why this matters when you've got 20 stops, not two

Small routes forgive you for planning imperfectly. A five stop round where everything's close together? Missing the optimal order costs you maybe five minutes. When you're running 30 or 40 or 50 stops in a day, order matters exponentially more. One bad decision compounds through the whole sequence.

DropPilot uses nearest-neighbour plus 2-opt optimisation to build smart routes from the start. But we knew that wasn't enough. The real world is too unpredictable. So we layered in continuous ETA refresh and smart rerouting. Together, they create something closer to what larger fleets get from dedicated dispatch software: a system that keeps optimising in real time.

A dispatcher managing a team doesn't need to chase every driver. The system is already correcting course. A single driver running a long day doesn't need to stop and replan every hour. The app is already watching. It's the difference between delivery logistics feeling like firefighting and feeling like a system that's actually working with you.