How to define the right performance indicators for a passenger counting system

A passenger counting system generates data that feeds strategic decisions downstream: sizing the service offer, allocating revenue between operators, calculating subsidies, or managing fare-free policies. If that data isn’t reliable, every decision built on top of it becomes fragile. Defining counting performance indicators as early as the specification phase makes it possible to:

• secure a fair technical comparison between several candidates;
• set verifiable contractual thresholds (SLAs);
• plan the control procedures to apply after deployment;
• avoid cost overruns caused by a solution that’s poorly suited to the network.

Without this groundwork, an operator or local authority risks selecting a system based on price alone, with no guarantee of its actual reliability once in operation.

To build performance indicators for a passenger counting system that are genuinely useful, it’s best to organize them into complementary families rather than focusing on a single figure.

Counting accuracy

This is the criterion most often highlighted, and rightly so: an accuracy rate below 95% can significantly distort ridership analysis. It’s worth specifying:

• the stated accuracy rate (typically between 95% and 99%);
• the measurement conditions (low, medium, or high ridership);
• the verification method used (manual control count, video audit, independent third-party organization).

For example, an independent verification carried out by Bureau Veritas on a network equipped with stereoscopic sensors measured an accuracy rate close to 99%, confirming the value of having this criterion audited by a third party rather than relying solely on the supplier’s own figures.

System reliability and availability

A system can be accurate in a lab and still prove unreliable in real-world operation. The reliability indicators for passenger counting to include are notably:

• sensor availability rate over a given period;
• the actual data reporting rate (missing data, sensors out of service);
• robustness under real operating conditions (vibration, lighting, heavy ridership, multiple doors).

Data granularity and depth

Two systems can report the same overall accuracy while offering very different levels of detail. It’s therefore worth assessing:

• counting by door, by vehicle, or by carriage, particularly on rail and metro networks;
• the distinction between boarding and alighting;
• detection of specific objects (bicycles, wheelchairs);
• the ability to produce a real-time occupancy rate, useful for passenger information.

Interoperability and technical integration

A counting system is almost never standalone: it needs to communicate with ticketing, passenger information, and operations tools. Criteria to check:

• availability of a documented API;
• standardized export formats;
• compatibility with the network’s existing systems.

Regulatory compliance

Finally, a criterion often underestimated at the technical specification stage: GDPR compliance. A system based on anonymous detection, without collecting identifiable personal data, significantly reduces legal risk and makes the project easier for passengers to accept.

Need help translating these indicators into your technical specification?

The Acorel team has been supporting transport authorities and operators since 1989 in defining their passenger counting requirements. Discover Acorel’s counting solutions

Not all criteria carry the same weight depending on the type of network. A network preparing for fare-free transport will place particular emphasis on accuracy and its verification, since this data directly determines funding. A high-ridership rail network will instead prioritize per-carriage granularity and real-time occupancy rates, useful for passenger information. Before launching any tender, it’s therefore advisable to:

• List the network’s priority use cases (revenue allocation, passenger information, service sizing, accessibility);
• Assign a weight to each family of performance indicators based on these use cases;
• Define non-negotiable minimum thresholds (for example: accuracy ≥ 95%);
• Plan the post-deployment control procedures within the contract.

Defining good performance indicators for passenger counting remains a foundational step, but it’s only a starting point. Once counting reliability is secured, the most advanced networks look to exploit that data in real time to manage vehicle occupancy, anticipate ridership peaks, and inform passengers before the vehicle even arrives. That is precisely the purpose of Acorel’s Passenger Flow Management solution, designed as the natural extension of reliable counting: it turns validated counting data into an operational tool for managing the network.

Since 1989, Acorel has deployed its automatic counting solutions across more than 20 transport networks in France, with an accuracy rate above 98%, validated by independent audits a concrete illustration of how these performance indicators can be verified over time, not just during a one-off demonstration.

Move from counting to real-time management. Discover Acorel’s Passenger Flow Management solution