Traffic and drive-thru sites: speed, timing and which side of the road

A good drive-thru site sits on a road where traffic moves steadily but not fast at the hours you sell, on the side drivers are travelling at those hours, with room to turn in and out without blocking anyone. Live road speed and the jam factor tell you whether traffic is crawling, flowing or racing past at breakfast and dinner; a road that is free-flowing at high speed is passed, and one at a standstill makes the exit a nightmare. Where cars are counted by direction, which today means roads in the United Arab Emirates, you can also check which side carries the morning traffic.

This guide is for quick-service restaurant and coffee brands, fuel and convenience retailers, car wash operators and the property teams who find sites for them. It covers the traffic questions particular to a drive-thru, what the data gives inside and outside the UAE, an example, the mistakes, and an AI employee that screens candidate sites. The general case for roads and sites is in does a busy road help a site.

The traffic questions a drive-thru adds

• Speed at selling hours. Breakfast and the evening meal are when a drive-thru earns most. The road's behaviour at noon matters less.
• Room to slow down. A road moving around half its free-flow speed gives drivers time to see the sign and turn. One at full speed does not.
• A way out. If the road is jammed, cars leaving the lane wait to rejoin and the queue backs up into the car park. A high jam factor at the exit hour is a warning.
• The side of the road. Morning coffee is bought on the way to work, so the inbound side wins in the morning; dinner on the way home favours the outbound side. A U-turn is a lost sale.
• Rivals within a short drive. A competitor two minutes earlier on the same route catches the driver first.

What the data gives you

Traffic Flow works anywhere it has roads measured. For each place it returns the live speed on the roads around it, the free-flow speed, a jam factor from 0 to 10, congestion as a percentage, the level in words, roads closed and the busiest road. It is measured when the flow runs, so you schedule it for 07:30 and 18:30, say, on weekdays. It does not count cars and has no direction split.

In the United Arab Emirates, the Road traffic history dataset in Regional data counts the cars on the road nearest each place, hour by hour since January 2021, in each direction, with their average speed. It turns into cars a day, the busiest hour, the busiest day, a trend and cars by direction. How to read it is in counting cars over time.

An example: three plots in Dubai

A made-up case. A coffee brand is offered three roadside plots in Dubai for a drive-thru. It runs Traffic Flow on weekdays at 07:30 and reads the Road traffic history for the last 90 days. Suppose the results are (invented for illustration):

Invented figures for illustration; not real figures for Dubai.

Plot B has the most cars, but they pass at nearly full speed and split evenly by direction: many will not slow in time. Plot C is on the busy side of the morning but jammed to a crawl, with five rivals nearby; leaving the lane will be slow. Plot A has fewer cars than B, moderate traffic at half speed and two-thirds of the morning flow on its side. For a breakfast-led drive-thru it is the plot to negotiate on.

The same screening in Johannesburg or Monterrey would rest on speeds, jam factor and rivals alone, without car counts or direction, and that is still enough to drop the plainly wrong plots.

Mistakes to avoid

• One reading. Live speeds change by the minute. Run the check on several weekdays and average the results.
• The wrong hour. A road that flows at noon may jam at 07:30. Measure at your selling hours.
• Treating more cars as always better. Fast, heavy traffic is hard to turn out of. Speed and side matter as much.
• Ignoring the busiest road. Traffic Flow averages the roads within the distance you choose; check busiest_road to be sure the figure is not driven by a side street. Adjacent roads (about 150 m) is the tightest setting.
• Promising counts outside the UAE. Elsewhere the data has speeds and congestion only.

Screening drive-thru plots with an AI employee

1. Start with Schedule, weekdays at 07:30, for a fortnight while you decide.
2. Bring in the plots with My data.
3. Add Traffic Flow with Adjacent roads.
4. In the UAE, add Regional data with the Road traffic history dataset, Totals for each place, the last 90 days.
5. Add Nearby places, Each incoming place with counts, for fast food and coffee within 1 km.
6. Add xMap LLM told the hours you sell and asked to judge each plot on speed against free flow, jam factor, rivals and, where present, cars by direction.
7. Send it by Email with Show on a map. In the inbox each morning: every plot's speed, jam factor, rivals, car counts in the UAE, and a short verdict.

A second flow at 18:30 covers the evening. After two weeks, the inbox holds ten readings of each plot at each hour, which is a far better basis than one drive past.