Core Definition
People counting automatically counts how many people cross a point or line — footfall outside, entries at the door, visitors per zone. Higher-level metrics like capture rate, conversion, and dwell all use these counts as numerator or denominator.
How Counting Methods Differ
Infrared sensors are inexpensive but miss people who overlap as they pass, and they can't tell direction or attributes. Wi-Fi signal-based counting depends on device penetration and signal settings. Vision-based counting recognizes people directly on screen, so it can also count attributes like direction, time of day, age group, and gender.
Accuracy is verified by how closely counts match a manual count. Before adopting a system, it's worth confirming whether the stated figures specify the verification standard and when the measurement was taken.
Why Counting Is Foundational
Capture rate is entries divided by footfall, and conversion is payments divided by visits. If the count that serves as the denominator is unstable, every metric built on top of it becomes unstable too. That's why the accuracy and consistency of counting determines the quality of a store's entire data set.
Conversely, once the count is stable, it can be combined with other data such as sales, weather, and promotions to extend into analysis that explains "why."
How SAAI Uses It
saai count uses a single camera to count people crossing a line drawn on screen. It separates footfall lines from entry lines to measure passersby and entrants at the same time, then aggregates the counts by time of day, gender, and age group. In verified field tests, accuracy runs 95 to 96 percent against manual counts. No video is stored. Only the resulting statistics remain.
See it in action