Correction Factor Measurement

To determine the appropriate correction factors for each client, we recommend a phased approach of field observation and subsequent data analysis that complements the staged implementation of people counting systems. The process will provide the actual versus electronic count in enough stores to estimate the Base Correction Factor, then increase the stores sampled to improve and validate the correction factors that would be developed. These field costs would be in addition to the monthly DATAssurance service plan costs.

Sources of Error

We typically consider three types of error that impact the accuracy of infrared people counting methods: Employee, Customer, and Technology-related error. By understanding and measuring these influences, more accurate adjustments and predictions may be calculated.

Employee error is essentially the error inherent in the store's business practices. The two primary factors are (a) the frequency of overcounting by non-store personnel such as delivery drivers, and (b) employees who are counted when entering and exiting the store.

Customer error has several sources, which are the result of visitor behaviors and demographics. Examples may be overcounting due to swinging arms, carrying packages, counting 'non-target' children, and visitors passing in and out of the store repeatedly.

Technology error is primarily the result of the limitations of the placement of the counters or the counting system itself to accurately measure the traffic. Every counting technology has certain conditions that contribute to less accurate reporting. Digital video systems may come out of alignment or not discern people from the background. For infrared systems, the primary issue is people walking side-by-side through an entrance. The wider an entrance is and the greater the rate of visitors passing by, the greater the likelihood of undercounting. The height of the counters, and the type of entrance, and the type of door, will also influence the ability for the counting system to measure accurately.