To effectively manage store performance, retailers need reliable traffic data. Visitor numbers, conversion rates, peak numbers, zone occupancy, marketing campaign effectiveness: all these indicators rely on precise measurement of in-store traffic.
But there's a recurring problem: staff are sometimes counted as visitors. Sales staff, department heads, logistics teams, security guards or managers may pass through the same areas several times in the course of a day. If these passages are included in the footfall figures, the data becomes less representative of customer reality.
The good news is that it is possible to exclude employees from the store count without using facial recognition, personal identification or biometric data.
Thanks to video analysis using artificial intelligence, retailers can distinguish between visitors and staff members on the basis of visible distinguishing features, such as clothing, badges, caps or professional equipment.
Visitor data plays a central role in retail management. In particular, it enables us to measure :
When employees are included in the count, the indicators can be distorted.
A sales assistant may cross the store entrance several times. A manager may move between the sales area and the storeroom. A logistics team may move continuously during opening hours.
These movements are not customer traffic, but they can be counted as such if the measurement system is unable to tell the difference.
The result: traffic appears higher than it really is, conversion rates may appear lower, and operational decisions are based on less reliable data.
A count that doesn't distinguish between visitors and staff can have several consequences.
If staff visits are included in the data, the number of visitors may be overestimated. This distortion is particularly significant in stores, where staff frequently circulate between the sales floor, the storeroom, the checkouts and internal areas.
The conversion rate is generally calculated by comparing the number of purchases with the number of visitors.
If traffic is inflated by staff passing through, the conversion rate mechanically appears lower. This can give the impression that a store is performing less well, when the problem is simply the quality of the data.
Two sales outlets may have very different organizations. One may have a small team, the other a larger one, with more trips to and from the sales floor.
Without staff exclusion, comparing their performance becomes less relevant.
Footfall data is often used to adjust schedules, organize teams, measure the impact of a campaign or optimize the layout of a sales outlet.
If the initial data is biased, so are the decisions that follow.
The method used by XXII is based on a simple approach: distinguish employees by means of visible distinctive signs, without ever seeking to identify individuals.
The aim is not to recognize a person. The aim is simply to recognize a characteristic element associated with the staff.
This could be, for example :
When a person wearing this distinctive sign is detected by CORE, he or she may be automatically excluded from attendance statistics.
Visitors are counted as normal. Employees do not pollute customer traffic data.
One of the key points of this method is that it does not rely on biometric data.
The system does not seek to recognize a face. It does not seek to identify a person. It does not create individual profiles. It does not compare individuals with a database.
It only analyzes the presence or absence of a pre-defined distinctive feature.
For example, if salespeople wear a specific jacket, the AI can learn to recognize this jacket as a criterion for exclusion from the count. If logistics teams wear a high-visibility vest, this can become a signal to differentiate them from visitors.
This logic makes attendance data more reliable, while greatly limiting the stakes involved in identifying individuals.
👉 You can also consult our ethics charter.
Each brand can define the criteria best suited to its environment.
Uniforms are one of the simplest signs to exploit. When a retailer uses uniform outfits for its teams, AI can detect these elements to automatically exclude employees from the count.
A visible badge can also serve as a distinctive sign. There's no need to read the contents of the badge or identify the person. CORE can simply detect the presence of a badge as an indicator of staff membership.
In some sales outlets, staff wear a cap, apron, choker or other special accessory. These elements can be integrated into the differentiation rules.
In logistics, industrial or retail environments, teams may wear vests, safety equipment or high-visibility clothing. These items are particularly useful for differentiating employees from visitors.
Automatic exclusion is generally implemented in several stages.
The first step is to identify the visual elements that differentiate staff from visitors.
This may be a specific outfit, badge, color, accessory or equipment.
Once the distinctive signs have been defined, our CORE software automatically recognizes them in the video streams and transforms this visual cue into a criterion for exclusion from the count.
When a person corresponding to the defined criteria is detected, he or she is removed from the customer traffic indicators.
Dashboards then display more reliable data, focused on actual visitors.
If uniforms change, new equipment is used or the store organization evolves, the rules can be adjusted.
This flexibility enables the system to be adapted to the operational realities of each retailer.
By automatically excluding employees, retailers get a more accurate picture of their actual traffic. The data reflect customer traffic rather than internal store activity.
The conversion rate becomes more relevant, as it is based on a visitor volume that is more faithful to reality.
Retail managers can thus better assess the sales performance of their outlets.
Excluding staff enables stores to be compared on a more homogeneous basis, irrespective of team size or internal organization.
With more reliable footfall data, retailers can better adapt schedules to actual customer flows.
This avoids overstaffing during certain periods and understaffing during peak periods.
Once staff movements have been excluded, area, flow and occupancy analyses become more representative of visitor behavior.
Teams can therefore better understand how customers move around, where they stop and which areas generate the most interaction.
AI-enhanced video analysis transforms video streams into useful data for operational management.
In the case of staff exclusion, the aim is not to identify individuals. The aim is to produce more reliable attendance data.
This nuance is essential.
A non-biometric solution does not seek to name, track or recognize a person. It only observes visual elements useful for analysis: the presence of a uniform, badge, equipment or distinctive sign.
This approach enables retailers to exploit the value of their video streams, while maintaining a logic of confidentiality and data minimization.
Confidentiality is a central issue for in-store video analytics projects.
Employees and visitors alike need to be able to operate in an environment where the technology serves a clear operational purpose, without unnecessary identification.
Automatic exclusion by distinctive signs is precisely the answer to this challenge.
It makes it possible to :
Retail is a dynamic environment, with constant flows, mobile teams and widely differing points of sale.
An effective solution must therefore be :
Staff exclusion by distinctive signs meets these criteria.
It enables retailers to obtain cleaner footfall data without complicating the store experience or adding heavy constraints for staff.
Automatic exclusion of staff from store counting is essential to obtain reliable footfall data.
But this exclusion does not require facial recognition, individual identification or biometric processing.
Thanks to an approach based on the detection of distinctive signs, such as uniforms, badges, caps or professional equipment, retailers can differentiate between staff and visitors simply, efficiently and confidentially.
For retail management, this means more accurate figures, more representative conversion rates, better team optimization and a more precise understanding of customer journeys.
AI-enhanced video analytics can transform existing video streams into reliable operational data, without compromising employee or visitor confidence.