Video analytics is software that automatically analyses camera footage to detect and classify objects and events. In retail it turns ordinary video into structured data, such as how many people entered, where they walked, how long they waited and how crowded an area became.
The software processes video frames, often using machine learning, to identify people as moving objects and follow them across the image. Analysis can happen on the camera, on a local device or in the cloud. The output is data rather than footage, such as counts per hour or a movement path drawn onto a store plan.
Retailers use video analytics for people counting, heat maps, queue monitoring and zone reporting. Hospitality venues and centres use it for occupancy and crowd control. Because it can reuse existing cameras, some businesses can add analytics without new sensors, although accuracy depends on camera angle, resolution and lighting.
The key consideration is privacy and governance. Video can capture identifiable people, so organisations should be clear about what is recorded, what is kept and who can see it, and review their obligations under the Privacy Act and state surveillance laws. Many systems anonymise data at source and keep only counts. Camera positioning for analytics is often different from security coverage.
Video analytics matters because it extends measurement beyond the door. Counting tells you who arrived, while analytics shows what happened next. Purpose-built overhead sensors are usually more accurate for entrance counting, so many stores combine both approaches for different jobs.