From distance guesswork to a national route plan
A national retail network had 330+ sites and no consistent way to plan travel between them. A three-factor distance, traffic and accommodation model gave field teams one repeatable national playbook.
Every route looked reasonable. Not every route was efficient.
Field consultants and operations managers travelled continuously between sites for inspections and reviews, but each province planned that travel independently, by habit, with no shared way to compare what "efficient" actually meant.
Each running field travel independently, with no shared method for comparing efficiency.
Visited on rotating circuits — far more than any team can sequence from memory alone.
On which overnight choice actually minimised next-day travel, versus what felt convenient.
The overnight choice that actually shortens the trip
In an anonymised excerpt from one 39-site provincial cluster, a circuit ends at Site F today with Site C scheduled first tomorrow — a direct 32 km apart, too far to link the same evening. Rather than picking an overnight stop by convenience, the model compares candidates on combined evening-and-morning distance.
Comparing candidates on combined distance, not the last or first site in isolation, consistently beats an overnight choice picked for convenience — here, by roughly a third.
A single broken site coordinate can silently distort a route recommendation before anyone sees it. Every route is checked against implausible readings before a recommendation is generated.
Our route & accommodation optimisation method
The same three-step process behind this result, adapted to your site network, your provinces, and your data.
Distance intelligence
Build a full point-to-point drive-distance grid from precise site coordinates — real road-network distance, not straight-line GPS.
Time-of-day & traffic layer
Layer time-dependent drive durations onto the distance grid, so routes are judged against realistic departure windows.
Overnight logic
Compare candidate overnight points against today's last stop and tomorrow's first, and recommend the one minimising combined travel.
One method, now running in every province
What began as a single-province pilot cluster now runs, unchanged in method, across every one of South Africa's nine provinces — more than 330 trading sites. Extending coverage needed no new methodology, only new site coordinates.
Turn your site network into a routing plan.
If your team plans routes, visits or overnight stays by habit rather than by data, let's find the factors that actually matter for your network.
Start a conversation: flowvisualpatterns.co.za