A people counter (also called a footfall counter, visitor counter, or people counting sensor) is a device that measures how many people enter, exit, or pass through a physical space. Retailers use the data to calculate conversion rates, optimize staffing, and measure marketing impact; airports, malls, museums, and smart buildings use it for occupancy, queueing, and space planning.
How people counters work: the 5 main technologies
1. 3D stereo vision (AI on edge)
Two lenses create a depth map of the entrance. Modern sensors such as V-Count’s Nano AI process everything on the device, count bidirectionally, exclude staff, and reach up to 99% accuracy independent of lighting — with no personal images recorded (GDPR-compliant by design).
2. Time-of-Flight (ToF)
Infrared light pulses measure distance to moving objects. Accurate in most conditions, though performance and analytics depth vary widely by vendor.
3. Thermal counters
Detect body heat. Fully anonymous, work in darkness, but struggle with crowds, ambient heat, and offer limited analytics (no demographics or zone data).
4. Wi-Fi / BLE tracking
Counts smartphone signals rather than people. Cheap to deploy, but accuracy commonly drifts 15–30% due to MAC randomization, and it raises privacy questions — best for trend data, not conversion math.
5. Beam / infrared break counters
A horizontal beam across a door counts interruptions. The cheapest option, but it cannot separate people walking side by side, distinguish direction reliably in groups, or exclude staff. Typical real-world accuracy: 80–90%.
Accuracy: what the percentages really mean
Vendor accuracy claims only matter if they hold in your conditions: high traffic, side-by-side entries, children, shopping carts, low light. Ask vendors how accuracy was validated (manual video count comparison is the standard), at what traffic density, and whether staff exclusion was active. As a benchmark: beam counters run 80–90%, Wi-Fi counting 70–85%, thermal 90–95%, ToF 95–98%, and leading 3D stereo vision AI sensors up to 99%.
What does a people counting system cost?
Three components drive cost: hardware per entrance, software/analytics subscription, and installation. Entry-level beam devices start under €100 but produce data too rough for conversion analytics. Professional 3D AI sensors with a cloud analytics platform are priced per entrance with a SaaS subscription — see V-Count pricing for a current bundled example (sensor + BoostBI analytics in one price). When comparing, calculate cost per accurate visit: a cheap counter that miscounts 15% of traffic corrupts every KPI downstream.
What to look for in 2026 (buyer’s checklist)
- Accuracy ≥ 99% validated under high traffic, with staff exclusion
- Privacy by design: on-device processing, no facial recognition, GDPR/CCPA compliance
- Analytics platform: conversion, occupancy, queue alerts, multi-site benchmarking — not just raw counts
- Open API for POS/BI/CRM integration
- Total cost per entrance including software and support, not hardware alone
- Global support if you operate in multiple countries
FAQ
Can people counters work in the dark? 3D stereo sensors with active IR projection (e.g., Nano AI) count accurately in complete darkness.
Do they count children? Depth-based 3D sensors count all individuals regardless of height; beam counters often miss children.
Are people counters GDPR-compliant? Only if no personally identifiable imagery leaves the device. Edge-processing 3D sensors are; camera/CCTV-based analytics often are not.
What is the difference between a people counter and CCTV analytics? CCTV was designed for security: accuracy is typically lower, and using it for analytics creates GDPR exposure. Dedicated counters are purpose-built and anonymous.
