Reordering the stops cut 8.7% of the miles.
A food distributor running 18 box trucks to restaurants and grocery markets sent us four weeks of route data from early 2022. First we changed the order of the stops and left everything else alone.
- Who
- Food distribution company
- Data
- Four weeks, 18 box trucks
- Stops reordered
- 8.7% fewer miles
- Fully reoptimized
- 22.6% fewer miles
What was going on
Eighteen box trucks, each rated for 20,000 pounds, ran multi-stop routes to restaurants and markets. Some trips stayed local. Others ran past 400 miles round trip and needed a DOT rest break built in. Nobody had measured how many of those miles the routes actually needed.
What we did
We ran the same four weeks three ways:
- As routed. The stops modeled as close to the original routes as the data allowed.
- Resequenced. The same trucks, trips and stops, run in a different order.
- Open. Any order could go on any truck, in any sequence, inside the real limits: 20,000 pounds a truck, weekday delivery windows of 8am to 5pm, and DOT hours.
The open model was allowed all 18 trucks. It used 13.
Three routes from the data set
| Route | As routed | Reordered | Change |
|---|---|---|---|
| One, 15 stops | 507 mi | 413 mi | −18.5% |
| Two, 15 stops | 372 mi | 317 mi | −14.8% |
| Three, 16 stops | 563 mi | 411 mi | −27.0% |
Why the real number sits in between
Reordering alone took 8.7% of the miles out across the four weeks. Opening everything up took 22.6%. Real customers keep their own receiving hours and habits, so the saving a distributor actually banks lands somewhere between the two. Either way, it saves money.
To keep it going, send us the stop data each day and our routing desk sequences it before the trucks load.
The math behind why five stops can be ordered 120 ways, and fifteen more than a trillion: Navigating Excellence →
The daily version of this study is our routing desk: RaaS+ →