Traffic Counting Accuracy vs. Precision: What’s the Difference?

Talitha Loftus
Talitha Loftus
Marketing Manager

The retail landscape has a tremendous number of options for entrance traffic counting, each with its own set of pros and cons. But, when it comes to accuracy, video – and, in particular, stereo video (3D) – remains the gold standard.

In the past, retailers would use traditional traffic counting tools and methods to count store traffic – initially through pen and paper, followed by clickers and then by more automated “clicker” technologies such as beam counts and basic cameras.

Luckily, those days are long gone.

Today, when it comes to determining store traffic, there are a variety of technologies available on the market that can count systematically. However, for retailers to get any return on their investment, traffic counts need to be two things 1) accurate and 2) precise.


Why? Simply because poor quality data gives rise to poor business decisions. And, of course, these bad business decisions can then have adverse effects on how your retail business performs, often leading to financial losses or unhappy customers.

A classic way of demonstrating the role of both accuracy and precision in traffic counting is by using the image of a dartboard. The dartboard example below is a visual representation of how different traffic counting solutions, such as stereo video, monocular video, Wi-Fi/Beacons and other types of people counting technologies potentially achieve accuracy and precision.   

In this case, let’s assume that three darts are thrown at the dartboard. Here the bullseye (center) of a dartboard represents the true value. Therefore, the closer the darts land to the bulls-eye, the more accurate they are. Whereas precision is reflected by a measure of how close a series of measurements are to one another.

Figure 1: Dartboard Example

Let’s take a deeper look …


Accuracy vs. Precision: What's the Difference? from RetailNext on Vimeo.

  • Stereo cameras are cameras with two or more lenses with separate image sensors for each lens (as opposed to a singular lens for monocular cameras). Stereo cameras simulate humans’ own binocular vision, with perhaps the most widely-known application being 3D video. It’s the 3D capabilities of stereo video that allow for the greatest accuracy as it’s not “fooled” by shadows, reflections, and the like.
  • Monocular sensors capture images through a single lens camera. A fisheye lens is an ultra wide-angle lens for wide panoramic images. Regardless of the lens, the sensor processes the image, and the output is data counts. However, accuracy can be problematic. That’s because these cameras are incapable of distinguishing adults from children, shoppers from associates, shadows, shopping carts and more.
  • Infrared beams count when a shopper crosses the doorway and “cuts” the beam. The advantage is low cost and simplicity. The challenge is accuracy. The sensors cannot recognize the direction of motion. They also have trouble differentiating between one or more people. Moreover, the system over-counts and under-counts with no data consistency.
  • Wi-Fi is an extremely valuable data source, and is very complementary to traffic counting data. However, as a traffic counting technology itself, Wi-Fi’s limitations make it less than ideal for stores trying to optimize operations and deliver differentiated shopping experiences. Sensors in a retail store are used to pick up Wi-Fi signals from smartphones. Those smartphones aren’t necessarily connected to a Wi-Fi hotspot: a phone in your pocket or purse may periodically look for a Wi-Fi connection so that it always has a good data connection available. However, not every smartphone in the store will be detected, and not everyone in the store is carrying a smartphone. More problematically, Wi-Fi can only drill down of around 20 meters, allowing for false positives of shoppers close to the store, but not in the store, or, if in a big store, in the store, but unknown as to what department, etc.  As such, Wi-Fi sensing best provides only a sample of data about your shoppers.

With that said, when it comes to investing in a traffic counting solution, sometimes the cheaper option is not always the better option. One needs to make sure the system is reliable enough to turn raw data into accurate and actionable insights.

Join the #retail, #inspiringretail and #SmartStore conversations on Twitter @TalithaLoftus & @RetailNext, as well as at www.facebook.com/retailnext.