Video analytics automatically classify a visitor as a solo shopper, a couple, a family, or a group by observing whether they enter and exit the store within the same time window as other people. This group detection relies on entry/exit re-identification (ReID), a technique already used to measure dwell time, without requiring additional cameras or models. It transforms a simple foot traffic count into data that qualifies purchasing potential, which can be used to optimize merchandising, staffing, and in-store promotions.
In brick-and-mortar retail, the unit of measurement has long been the individual: we counted entries, segmented profiles, and measured dwell times. But a customer shopping alone behaves differently than one shopping with others. A family with children does not browse the aisles the same way a couple does, and a couple does not shop the same way a group of friends does.
The composition of the group is a key determinant of purchasing behavior, and until now, it has remained invisible in foot traffic data. A raw count of 500 visitors says nothing about actual purchasing potential: this figure could represent 500 solo shoppers or 150 families of 3 to 4 people—two completely different commercial realities.
Group detection is based on a simple principle: if people enter the store together and leave together within the same time frame, they are considered to belong to the same group.
Technically, a group is characterized by three dimensions:
This approach relies directly on the entry and exit timestamps for each individual, which are already generated by the non-biometric re-identification engine. No biometric data, no facial recognition: only anonymized visual characteristics.
| Typology | Technical Definition | What this reveals about purchasing behavior |
|---|---|---|
| Solo | A group of one person, regardless of its composition | Quick shopping trip, individual decision-making, less influenced by family dynamics |
| Couple | Consisting of two adults | Joint decision-making, often longer visit duration |
| Family | At least one child under 18 and at least one adult | Itinerary slowed down by the children; high sensitivity to dedicated spaces and activities |
| Other | Any group that does not fit into any of the three previous categories (e.g., group of friends, coworkers) | Group behavior, potential for cross-selling |
No. For stores already equipped with ReID Entry/Exit, group detection is activated without any infrastructure changes, additional cameras, or specific configuration. It is a native extension of the existing module: clustering is performed directly on the pre-calculated entry and exit time windows.
Once enabled, the feature enriches existing dashboards with:
In practical terms, a retail operations manager is no longer satisfied with simply knowing how many people entered the store: they also know who they came with, and can link this count to actual purchasing potential.
Knowing the breakdown of solo customers versus couples versus families over a given period allows for concrete action on several fronts:
This customer insight—which until now was only accessible through in-store surveys or panels, with the delays and biases that entailed—is now generated automatically every night using existing video feeds.