Logistics site road access: testing the roads around candidate parks
Before a lease, read speeds, jams and closures on the roads around each candidate park at truck hours for a few weeks, and compare the sites side by side.
Part of the guide Location intelligence automation
To judge logistics site road access, read the roads around each candidate park at the hours your trucks and shifts will use them, for several weeks, and compare the parks on the same measures: how much of their free-flow speed the nearby roads keep, how often they jam, how many stretches are closed, which road is the worst bottleneck, and how many roadworks and closures appear nearby. A park that looks perfect on a brochure map can sit behind one junction that locks up every afternoon.
This is for real estate and network planning teams at third-party logistics firms, retailers, manufacturers and e-commerce companies, and for the brokers who shortlist warehouses for them. It covers the questions to ask of the roads, a measuring plan, an invented comparison of three parks near Johannesburg, the errors that skew a shortlist and the AI employee that collects the evidence. It sits under the guide to AI employees for site selection.
Questions to ask of the roads
Every logistics site depends on a short stretch of road between its gate and the nearest major route. Most of what goes wrong happens there. Ask:
- How fast does traffic move on the roads within a kilometre of the gate at 05:00, 14:00 and 18:00, compared with when they are free?
- Which named road is the slowest, and is it the one trucks must use?
- How often are stretches near the gate closed, and are there long-running roadworks?
- What can a van reach within 15 minutes, and does that include the customers or stores it serves?
The Traffic Flow block answers the first two each time it runs. For each candidate it returns the current and free-flow speeds, a jam factor from 0 to 10, a congestion percentage, how many road stretches were measured and how many are closed, and the busiest road by name. Set it to every road within 1 km of the park, because the junction that matters is often not the road outside the gate.
A measuring plan that holds up
- List the parks with a point at each main gate, not the middle of the plot.
- Choose the hours that match your operation: the first trucks out, the shift change, the evening returns.
- Read every park at those hours on weekdays for four to six weeks. Live speeds vary day to day; weeks of readings give a fair median.
- Log roadworks and closures within 2 km of each park on the same timetable with the Traffic incidents block, set to check around each place.
- Draw what a van reaches within 15 minutes of each gate with the Travel area block, and count the stores or customers inside it with Nearby places or your own table.
An invented comparison near Johannesburg
None of these figures is real; they show how the evidence lines up. A distributor compares three parks north and east of Johannesburg over five weeks.
| Park | Median speed kept at 14:00 | Readings with jam factor 6+ | Busiest road | Closures and roadworks seen nearby |
|---|---|---|---|---|
| Park 1 | 0.71 | 4 of 50 | The main arterial | 3 |
| Park 2 | 0.48 | 19 of 50 | Single access road | 9 |
| Park 3 | 0.66 | 7 of 50 | A ring road on-ramp | 2 |
Invented readings: speed kept is current speed divided by free-flow speed.
Park 2 had the lowest rent and the best brochure. It also had a single access road that jammed at the shift change in nearly two readings out of five and carried most of the roadworks seen. Park 1 and Park 3 go forward; the team visits both at 14:00 to see the queues with their own eyes.
Errors that skew a shortlist
- One visit as evidence. A Tuesday morning tour shows one Tuesday morning.
- Reading the wrong hour. Office rush hour and truck hours differ. Measure when your vehicles move.
- Different radii for different parks. Keep 1 km for all, or the comparison measures different roads.
- Ignoring confidence. Each reading has a confidence from 0 to 1. Very early readings on empty roads can be thin; set them aside before taking medians.
- Forgetting the workforce. Roads matter, and so do people who can reach the park. Where population figures exist (among them France, Canada, India, Thailand, Egypt and Australia, not the United Kingdom or the United States), count residents inside a drive area with Demographics.
Have an AI employee collect the evidence
In xMap Flows one flow can gather the readings for every candidate while the team negotiates:
- Schedule: weekdays at your truck hours, for example 05:30, 14:00 and 18:00, in the local time zone.
- My data: the candidate parks, one row per gate.
- Traffic Flow: every road within 1 km.
- Traffic incidents: around each place, 2 km.
- Google Sheets: one row per park per run, so the history builds itself.
- At the end, Data analyst reads the sheet and a Make a report block turns the medians into a short PDF for the board.
The board pack then rests on weeks of readings, not a site tour. Templates that help: Where to build the next plant: workforce and logistics, Warehouses within 15 minutes' drive of the port and Where to put micro-fulfilment hubs. More on the logistics and manufacturing pages.
Questions
How long should I measure the roads around a candidate park?
Four to six weeks of weekday readings at your truck hours gives a fair median. A single day can be thrown off by one breakdown or event.
Can I see how many trucks use the road?
Not with live speeds, which measure how fast traffic moves. Hourly vehicle counts since 2021 are available for roads in the United Arab Emirates through Regional data.
What radius should I use around a logistics site?
Every road within 1 km of the main gate usually catches the junction trucks queue at. Keep the same radius for every park you compare.
Does it show roadworks planned for next month?
No. Traffic incidents shows closures and roadworks present when the flow runs. Logging them on a timetable shows how often a park's roads are disrupted.