How to do a manual traffic count: vehicle classes, 15-minute intervals, the tally sheet and the peak hour

How a manual traffic count works

A manual traffic count is a person at a fixed point recording every vehicle that passes, by class and direction, in fixed time intervals. The US Federal Highway Administration's Traffic Monitoring Guide (FHWA, December 2022, revised April 2026) defines manual methods as those in which "a human observer" determines and records the numbers, reported by time interval, vehicle class, occupancy or turning movement.

The same guide says manual methods have limited application because of their limited use and high cost, and calls manual or visual counts "often a last resort" for classification where equipment cannot be used. Manual counts keep three jobs: checking automatic equipment, counting where no counter can be laid (busy junctions, shared paths, construction sites), and studies run by students, councils and community groups with no equipment at all. Montana's contribution to the guide names the weakness plainly: a manual count is "a 'one time' observation", so an error cannot be reviewed later as video can.

The steps, as the specifications below agree on them:

1. Choose the site and the line. A safe place with a clear view, and a fixed line across the road, such as a lamp post or a road marking, at which each vehicle is counted.
2. Prepare the sheets. One per direction, with location, date, direction, weather and observer filled in before you start.
3. Count in intervals. A tally mark per vehicle in its class column, a new row every 15 minutes.
4. Total as you go. Add up each row when its interval ends, and each sheet at the end.
5. Check and enter. Compare directions and intervals for anything implausible, then type the totals into a spreadsheet.

Vehicle classes: three schemes compared

Use the classes of the authority your count is for; if none applies, use a simple scheme whose groups are combinations of an official one. The FHWA says exactly that: systems with fewer categories "should be combinations of the FHWA classes".

Three rules from the definitions settle the arguments that start at the roadside:

• A car towing a caravan or trailer is one vehicle, counted as a car, in the DfT's definitions.
• Raised axles do not count. The DfT classifies a lorry with "sleeping axles" by the axles on the road, and the FHWA recommends counting only axles in the dropped position. The FHWA gives the consequence: a quarry lorry can be class 7 loaded and class 6 empty.
• Vans and lorries split at 3.5 tonnes gross weight in the UK, a weight you judge by the vehicle's type, since you cannot see it.

For many purposes fewer classes will do. The FHWA notes that many planning analyses need only a car and truck split, or three or four length bins: cars, small trucks, large trucks and multi-trailer trucks, though "there is no common definition" among states of where a car becomes a truck.

Intervals: why 15 minutes

Record counts in 15-minute rows, because the busiest hour seldom starts on the hour and hourly totals cannot be split afterwards. The FHWA's data formats take vehicle volumes in hourly, 15-minute or 5-minute bins, and its record for walking and cycling counts allows 5, 10, 15, 20, 30 or 60 minutes, asking for "the shortest feasible intervals", longer when counting by hand. The UK census and Papua New Guinea's specification total by the hour; the DfT's raw count files hold one row per direction per hour, where 7 means 7am to 8am.

Hourly rows are enough for a daily total. For anything about the peak, such as junction design, a crossing or a school drop-off, the peak hour and the peak hour factor need 15-minute data, as the worked example below shows.

How long, and when, to count

Count when the traffic you are studying happens: a full weekday for a daily figure, the peak period for a peak-hour study, and a typical day rather than a holiday. The national schemes differ:

The DfT turns each 12-hour count into an annual average daily flow by multiplying it by an expansion factor for the vehicle type, the date and one of 10 road strata (by type of road, location and flow), worked out from about 300 automatic counters, according to its annual methodology note and its 2024 background quality report (published June 2025). A short count on its own measures that day, not the year: converting it needs factors from continuous counters on similar roads, which a highway authority holds.

The tally sheet

A tally sheet has a header for where, when and who, a column for each class, and a row for each 15-minute interval, with totals across and down. Papua New Guinea's manual counting specification (Department of Works and Highways, last reviewed July 2025) asks for "a tick mark for every vehicle in a given classification", a separate sheet for each direction of travel with the direction written on it, and the form filled in as far as possible before going to the site.

Group the marks in fives, the five-bar gate described in the guide to tally marks, so a row can be totalled at a glance. Keep pedal cycles outside the motor vehicle total: the DfT's "all motor vehicles" is "all vehicles except pedal cycles". If you count people walking or cycling, the walking and cycling count instructions count the people passing, and a person each time they pass.

Worked example: finding the peak hour

A two-hour morning count on one direction of a road gives these all-motor-vehicle totals per 15 minutes. The numbers are made up for the example; the method is the FHWA's definition of the peak hour factor.

1. Slide a one-hour window down the rows. Each one-hour total is four consecutive intervals.
2. The peak hour is the highest: 07:30 to 08:30, with 490 vehicles.
3. The clock hours were 375 (7 to 8) and 442 (8 to 9). Reporting the busier clock hour understates the peak hour by 48 vehicles, about 10%.
4. The peak hour factor. The FHWA's guide defines it as the peak hour volume divided by four times the 15-minute volume, here the busiest quarter-hour inside it, 141: 490 ÷ (4 × 141) = 490 ÷ 564 = 0.87.

A factor of 1.0 would mean the hour's traffic arrived evenly; 0.87 means the busiest quarter-hour ran at a rate of 564 vehicles an hour, 15% above the hour's average. That rate, not the hourly total, is what a junction or crossing has to cope with.

Counting with a counter instead of paper

With a tally counter or a counting app, read and write down the running total at each quarter-hour instead of resetting; the interval counts are the differences between readings. Resetting at the boundary loses the vehicles that pass while you do it. The example above, read cumulatively, would show 62, 143, 247, 375, 516, 633, 729 and 817; 247 − 143 gives the 104 of the 07:30 interval.

Simple Online Counter is a free, browser-based set of counters that keeps its counts in your browser's own storage. Its multi counter runs one named counter per class, such as "Cars northbound", all on screen in grid view with their own plus and minus buttons, and exports every total as a CSV. The tally counter suits a one-class count, such as cyclists on a path. Neither recognises vehicles: the class of each one is your call, and the rules above are there to make those calls consistent. The multi counter page also explains counting in categories with a check sheet, the paper version of the same idea.

Observers, breaks and safety

Plan at least one observer per direction on a busy road, more for many classes or a junction, and relief at least every two hours. Papua New Guinea's specification leaves the number of tally clerks to its project coordinator for each station and has new clerks do a 30 to 40 minute supervised trial count after training. The walking and cycling count instructions from the US National Bicycle and Pedestrian Documentation Project, run with the Institute of Transportation Engineers' Pedestrian and Bicycle Council, set out the staffing:

• one counter per location, except at an extremely busy downtown junction, plus one backup for every five locations;
• counters on site at least 15 minutes before the start;
• relief for a break at least every two hours, and 30 minutes for lunch;
• high-visibility clothing, identification, a watch and a clipboard, and a manual clicker in very busy places.

Stand on public land, off the carriageway, where you can see the counting line without stepping into traffic. Papua New Guinea's specification makes safety the responsibility of the officer in charge and requires signs at the site and a high-visibility vest on every counter. Junction turning counts need a counter per approach or two, because every arm's left, ahead and right movements are separate columns, which is why the FHWA guide notes that intersection counts "may require additional counters".

Checking the count

Before you trust the totals, look for intervals that break the pattern, compare the directions, and re-add every row. The NBPD instructions tell count managers to question results that vary significantly from one period to the next or are "unusually consistent", and to have them explained or discarded. The FHWA's guide uses manual counts the other way round, as the check on automatic equipment; a Florida example in it compares the two hour by hour.

Three checks for any count:

• Row and column sums. The grand total of the class columns must equal the grand total of the interval rows.
• The interval pattern. A quarter-hour that is half or double its neighbours needs a reason, such as a bus arriving with a queue behind it, or a note that the observer was distracted. The same habit of interval snapshots catches errors in a door count.
• Direction balance. Papua New Guinea's specification checks that the two directions of a seven-day count are within ±5% of each other. Over a shorter count, peaks can be one-sided, but a large gap over a whole day is worth explaining before it reaches a report.

Mistakes that spoil a manual count

What this guide does not cover

This guide covers manual classified counts of vehicles, and of people walking and cycling, at a point on a road or path. It does not cover automatic counters, video analysis, speed surveys, origin and destination surveys or the factoring of short counts into annual figures, which highway authorities do from their continuous counting stations. If you are counting for a highway authority or a planning application, its brief and class scheme override everything here, and a count beside a road needs the site safety arrangements that authority requires.