Every people counter on the market quotes an accuracy figure, and almost every one of them quotes something between 95 and 99 percent. The numbers look interchangeable. They are not. Two systems advertising the same percentage can perform completely differently on your busiest door, because the figure on the spec sheet was measured under conditions that may have nothing to do with yours.
This explains what an accuracy figure actually represents, which real-world conditions move it, and how to run a test that tells you the truth before you sign.
How accurate are people counters?
People counters are typically advertised at 95 to 99 percent accuracy at a single entrance under normal conditions, but the headline number hides what matters. Accuracy depends on the hard cases: groups walking in together, tailgating, low light, wide entrances, and high ceilings. A vendor quoting 99 percent on an empty test corridor may sit far lower on a busy mall door. The only number worth trusting is one measured on your own site, against a manual ground-truth count, during peak hours.
Why "99 percent" is meaningless without conditions
An accuracy figure is only as good as the test that produced it. Four conditions change the result more than the sensor brand does:
- Entrance width. A narrow single-file door is easy. A six-metre mall entrance where people fan out is hard, and a sensor rated for one may struggle with the other.
- Ceiling height. Mounting height changes what the sensor sees. A figure measured at three metres does not transfer to a nine-metre atrium.
- Traffic density. Accuracy at ten people a minute is not accuracy at two hundred. Counters degrade as scenes get crowded, which is exactly when the count matters most.
- Lighting. A camera-based counter that hits 99 percent at midday can fall off sharply at dusk or after close.
A vendor figure with none of these stated is a marketing number, not a measurement.
The hard cases that move accuracy
When accuracy drops, it almost always drops on the same handful of situations:
- Groups. Two or more people entering together are the single largest source of undercounting.
- Tailgating. One person closely following another reads as one event to a simple sensor.
- Strollers and wheelchairs. A pram can read as an extra person, or a wheelchair user plus an aide can read as one.
- Reflections and glass. Polished floors and glass doors create false motion.
A counter's real accuracy is its accuracy on these, not its accuracy on a single calm person walking through.
How to run your own accuracy test
You can settle the question in an afternoon, and the method is simple enough to put in a contract:
- Pick your busiest entrance and your busiest hour.
- Have a person manually tally entries and exits with a clicker for a fixed window, ideally an hour.
- Pull the counter's number for the identical window.
- Accuracy is the counter's count divided by the manual count, expressed as a percentage, with over- and under-counting both noted.
Repeat across a few windows and a couple of doors. See a fuller accuracy test methodology for the full protocol and how to handle multi-door sites.
How the sensing method affects accuracy
The method sets the ceiling on what accuracy is possible. A single beam cannot separate a group, so no amount of tuning fixes group undercounting. A flat camera cannot count what it cannot see in the dark. Depth sensing separates people by shape and works without ambient light. Combining depth at the door with a second signal across the interior closes more of the hard cases than any single sensor can.
Ariadne measures this with Hybrid Fusion, its patented camera-free method. Time-of-Flight depth sensing counts every visitor at the entrances, capturing geometry rather than images, while patented phone signal sensing follows movement through the interior, detecting the signals a phone emits even in airplane mode. The sensor streams both feeds to Ariadne, where Hybrid Fusion combines them into one trajectory per visit and computes counts, dwell, and paths. The streams carry no identifier: no MAC address, no device ID, no biometric data, and no camera is involved. Identifiers are stored only when a visitor explicitly opts in, which keeps the method GDPR-friendly and outside biometric territory.
Accuracy and privacy are not a trade-off
A common assumption is that the most accurate counter must be the most invasive one, a camera watching the door. That is not true. Depth and signal methods reach high accuracy on the hard cases without capturing a face or an identifier. You do not have to choose between counting people well and respecting them.
FAQ
What is a good people counter accuracy?
Vendors typically advertise 95 to 99 percent at a single entrance under stated conditions, but insist on a figure measured at your own site during peak traffic.
Why does my people counter undercount?
Almost always groups and tailgating. A sensor that cannot separate people walking close together will read several as one.
Does lighting affect people counter accuracy?
For camera-based counters, yes, significantly. Depth methods that supply their own light are far less affected.
How do I test my people counter's accuracy?
Manually tally entries for a fixed window at a busy door and compare with the counter's number for the same window.
Is a more expensive people counter more accurate?
Not necessarily. Accuracy comes from the method and the processing, not the price. Test before you assume.
---
