- Home
- People Counting
- Accuracy
Accuracy
No counting method is perfect. Even the expensive option of a human counter typically has a 1-2% margin of error. However, the limitations of infrared counting technology are well understood, and easily compensated for with the proper application of sound practices, testing, and statistical correction factors. Infrared counting is considered to have an error margin between 1% and 3% depending on conditions, such as placement of the detectors, visitor characteristics, and traffic volume. Counting Technologies Corporation uses sound statistical correction to minimize any potential error from both overcounting and undercounting.
Overcounting 
Commonly know as false positive readings they typically resulting from:
- Inclusion of individuals, such as children, who are not within a store's target market
- Tall people, with swinging arms
- Customers who pass an infrared beam multiple times
- Tall packages, baby strollers, and other objects that are not people
Where screening by age is a consideration, children placed in the toddler seat of a shopping cart may also be counted, as are unusually tall children, depending upon the height of the infrared unit. A detector height of 56 inches will count nearly all persons 12 years or older, yet few children 10 years old or younger.
Undercounting
Primarily occuring when groups of people walk side-by-side when passing in front of the beam. This will cause multiple persons to be counted as one person, since the beam was not detected between individuals. Undercounting is increased with wide and high-traffic entry/exit doors, but typically occurs as a predictable percentage of overall traffic flow, and correction factors may be applied to minimize any effect. Undercounting may be nearly eliminated when only a single door is used for passage. The accuracy of infrared units can decrease to an error rate of over 3% when more than 400 visitors per hour cross the beam.