What makes a people counter miscount
Staff movement is one of the most common sources of miscounting: without staff exclusion, a people-counting system records employees alongside customers, so the footfall figure stops reflecting genuine visitor traffic 1. Staff exclusion is the ability of the system to detect and separate employees from customers, and it improves footfall accuracy, conversion rates, occupancy monitoring and retail analytics without identifying individual people 1. Where exclusion depends on wearable tags, accuracy depends on consistent tag use, so a forgotten or unworn RFID or Bluetooth tag reintroduces the error 1.
Duplication and non-human objects
Counting the same person more than once is a second source of error. The MRC Out-of-Home Measurement Standards draft states that a user should only be counted once for unique measurement, even though a user can have multiple visits during a reporting period, and that the qualifier "Unique" should be limited to records de-duplicated across the whole reporting period 2. The same draft describes location circulation as an explicitly duplicative count of people or pedestrians in a defined location 2. It also notes that controls are required to ensure pedestrian counts are human-only, which points to non-human objects entering a count as a failure mode 2.
Obstructions and transient traffic
The MRC draft observes that obstructions would not necessarily negate the counting of gross impressions, and that some measurement techniques are designed so that transient individuals are purposely excluded 2. Where transient population is handled by adjustment or by design, the draft says users of the data should be told how it was accounted for and how those methods influence reported metrics 2. Measurement frequency should also be as regular as possible and disclosed 2.