FlowsBook a callSign inCreate free accountSign up
Roads and traffic

Working out the cost of traffic delays for fleets, route by route

Find the slow roads on each route, turn lost minutes into driver hours, missed stops and fuel, and see which routes are worth changing first. A worked example inside.

Part of the guide Location intelligence automation

By the xMap team · · 5 min read

The cost of traffic delays for fleets is the extra time vehicles spend on slow roads, multiplied by what an hour of a vehicle and driver costs you, plus the knock-on costs: overtime, missed delivery windows and the stops a route can no longer fit. You can estimate it per route by comparing how fast traffic moves on the roads around each stop with how fast those roads move when they are clear.

This is for fleet and transport managers, delivery companies, field service firms and anyone whose vans or trucks run the same routes every week. It explains where the cost comes from, how to measure the slow parts of a route, a worked example with invented numbers, the mistakes that inflate or hide the cost, and an AI employee that builds the evidence for you.

Where the money goes

  • Paid time. Every minute a driver sits in slow traffic is paid. Across a fleet and a year, small daily losses add up.
  • Overtime. When a route runs late, the last hour is often paid at a higher rate.
  • Missed windows. A delivery that misses its slot may be refused, rebooked or penalised by the customer.
  • Lost capacity. A route that loses 40 minutes a day may drop a stop, which means another vehicle or a later day for that customer.
  • Fuel and wear. Stop-start driving uses more fuel and brakes than a steady road, though this part is harder to put a figure on.

Measuring the slow parts of a route

You need two things for each stop or road section: how fast traffic usually moves there at the hour your vehicle passes, and how fast it moves when the road is clear. The Traffic Flow block gives both, live, for the roads around each place, along with the jam factor from 0 to 10.

Because the readings are live, run them at the hours your routes actually pass each stop, on enough days to be fair, and keep them. A morning route through a business district should be measured in the morning; an afternoon return through a school zone, in the afternoon.

  1. List the stops on each route with the usual time the vehicle reaches them.
  2. Group stops by the hour they are reached, so one reading covers stops passed at the same time.
  3. Read the roads around each stop at that hour on every working day for three or four weeks.
  4. For each stop, work out the median ratio of current speed to free-flow speed.
  5. Estimate the extra minutes on the stretch before each stop: the time it takes now minus the time it would take at free flow.

A worked example with invented numbers

Everything in this example is made up, to show the arithmetic. A parcel firm in Mexico City runs 30 vans. Take one route of 18 stops. Its drivers report that the morning half is slow and the afternoon half is fine.

The firm measures the roads around each stop for four weeks. In the morning, the eight stops in and around the business district show a median speed of about half their free-flow speed. If the van covers 2 km between each of those stops and free flow there is 40 km/h, each stretch should take 3 minutes and actually takes about 6. That is 3 extra minutes a stop, or 24 minutes across the eight.

Say a van and driver together cost the firm 40 an hour in its own currency, and the route runs 22 days a month. 24 minutes a day is 0.4 hours, so 0.4 x 40 x 22 = 352 a month for this route, before overtime or missed windows. If the same pattern holds on 12 of the 30 routes, the firm is looking at roughly twelve times that each month, and a strong case to move the business district stops later in the day, when the readings show those roads at close to free flow.

Mistakes that distort the figure

  • Counting the whole route as delayed. Most routes are slow in a few places only. Measure stop by stop and add up the slow parts.
  • Using the speed limit as the baseline. Free-flow speed on a city street is often well below its limit. Compare with free flow, or you overstate the loss.
  • One week of data. A week with a festival or a big closure is not typical. Use several weeks and the median.
  • Ignoring closures. A road shut for a month is a detour, not congestion. Check Traffic incidents along the way and treat long closures separately.
  • Expecting vehicle counts. Speed data says how traffic moves, not how much of it there is. Hourly counts exist only for UAE roads.

An AI employee that collects the evidence

Gathering weeks of readings by hand is where most cost studies stall. An AI employee can take the readings on a timetable and write the totals, which is the kind of steady job covered under AI employees for field operations.

  1. Schedule: weekdays at the hours your routes pass the key stops.
  2. My data: your stops, each with its route name and usual arrival hour.
  3. Traffic Flow: the roads adjacent to each stop.
  4. Calculated field: the ratio of speed to free flow, and the extra minutes for a stretch length you store per stop.
  5. A Google Sheets app block adds a row per stop per reading.

After a month, a second flow can read the sheet and ask the Data analyst to total the extra minutes by route and rank them. To act on the result, the Route planner can reorder stops from a new start time, and the template Weekly road disruption report keeps closures in view. The weekly answer lands as an email with a ranked table, and the AI employee's Team inbox thread records each week's totals.

Questions

How do I put a value on a lost minute?

Use your own cost of an hour for a vehicle and driver, including pay and running costs, and divide by 60. Add overtime and missed-window costs separately if you know them.

Does the Route planner use live traffic?

Route planner orders stops for the best visiting order on the roads. Live speeds come from Traffic Flow, which you can read at the hours a route runs.

How many weeks of readings do I need?

Three or four weeks of working days is a reasonable start. Take the median for each hour, so one bad day does not set the figure.

Can I count how many cars cause the delay?

Not with live speed data. Hourly car counts exist only for roads in the United Arab Emirates, in the Road traffic history dataset.

Templates that do this

Read next

More on roads and traffic · Every article · Templates

Put it to work

An AI employee in xMap Flows does this on live data, on its own timetable. Start from a template or describe it in plain words.