People counting and store conversion rate, from your existing cameras
People counting is the automatic counting of visitors entering and leaving a store from camera video. EscherSense does it with AI on the security cameras you already have: it separates entries from exits, removes staff from the count and matches net footfall against POS transactions to give you the store's true conversion rate.
When sales drop, the first question is simple: did fewer people come in, or did fewer of them buy? Your till report only answers half of it. Without knowing how many people walked through the door, you cannot tell a traffic problem from an in-store problem.
Reliable door counts turn conversion rate from a month-end estimate into an hourly KPI. Rosters, window changes and promotions can all be judged on the same scale, and when conversion dips at your busiest hour you can read it alongside queue and checkout data to see why.
How it works
Directional counting
An overhead view of the entrance counts entries and exits separately. Someone lingering in the doorway or stepping back in moments later is not counted twice.
Staff and courier filtering
Employees and delivery traffic are identified by uniform, movement pattern and time spent in store, then removed from the visitor count.
POS matching
Net footfall is combined with the number of transactions in the same time window, showing exactly what share of visitors actually bought.
Live dashboard and alerts
Hourly traffic, conversion and store comparisons feed a live dashboard. When live occupancy crosses the limit you set, an alert goes out.
What it measures
Net footfall
Genuine customer traffic after staff and repeat entries are removed.
Conversion rate
Transactions divided by net footfall; tells you whether a sales change comes from traffic or from in-store performance.
Hourly traffic
Shows which parts of the day bring the most visitors; the core input for staff scheduling.
Capture rate
The share of passers-by outside the store who walk in; puts a number on your storefront's pulling power.
Live occupancy
How many people are in the store right now; used for capacity limits and safety rules.
Store benchmarking
Compares stores of different sizes, and different periods, on the same metrics.
In-store use cases
Finding the real cause of a sales dip
If sales fall while footfall holds steady, the problem is inside the store, not at the door. The focus then shifts from marketing to fitting rooms, checkout and sales assistance.
Measuring your own traffic in a mall
The mall's total visitor figure says little about your unit. Measuring passing traffic and walk-ins separately gives you your store's real capture rate.
Fair comparison across a chain
A larger store will naturally take more revenue. Comparing on conversion rate shows which store is best at turning visitors into buyers.
Capacity and occupancy control
On event days or in peak season, the number of people inside is tracked live. When the limit is reached the manager is notified and entry can be managed.
Privacy and GDPR/KVKK
People counting is fully anonymous: no facial recognition, no identification and no biometric records. Each visitor is tracked only by a temporary number while in view, and reports contain aggregate counts rather than individual records. By default, video is processed on a server inside the store and raw footage never leaves the premises; if your IT policy prefers, processing can run in the cloud instead. Retention periods are defined, and the setup is designed around KVKK and GDPR requirements.
Setup and integration
Counting needs one overhead camera above the entrance; the ideal mounting height is 2.6 to 4 metres. Any ONVIF-compatible IP camera that provides an RTSP stream will work, and the feed is read passively from the camera or your NVR, so your recording setup stays as it is. Footfall data flows to POS, ERP, CRM and BI tools via REST API and webhooks, and JSON and CSV exports plug into reporting tools such as Power BI or Looker.
Frequently asked questions
How accurate is camera-based people counting?
EscherSense measures 96.2% counting accuracy in the field, with a typical post-calibration range of 95 to 97%. Because each person is detected and tracked as a separate object, groups walking in side by side are still counted individually. Thresholds are calibrated to your entrance during the pilot.
Will my existing security cameras work for people counting?
In most stores, yes; any ONVIF-compatible IP camera with an RTSP stream can be used. The best results come from a camera looking down on the entrance from 2.6 to 4 metres. We assess your camera angles free of charge before installation and flag any gaps.
Are staff and couriers included in the visitor count?
No, staff and delivery traffic are filtered out of net footfall. They are distinguished by uniform, movement pattern and time spent in the store. Without this filtering, your conversion rate would come out lower than it really is.
How is store conversion rate calculated?
Conversion rate is the number of transactions divided by net footfall over the same period. EscherSense matches visitor data with your POS to produce it hourly, with 92.8% conversion measurement accuracy. If conversion drops during the evening rush, you see it before the store closes.
Why use camera-based counting instead of an infrared counter?
Camera-based counting is both more accurate and far broader in scope. Infrared beam counters can count two people side by side as one, struggle with direction and typically achieve 80 to 90% accuracy in the field. A camera also gives you a retail heatmap, queue and behaviour data from the same feed.
What does a people counting system cost?
Pricing depends on the number of cameras, the modules you choose and whether processing runs in store or in the cloud. Because your existing cameras are used, most projects need no new camera hardware. See pricing for plans, or contact us for a quote for your store.
Related modules
Queue management system and checkout performance analytics
A queue management system continuously measures how many people are waiting at the checkout and how long they wait, then alerts staff when a set threshold is crossed. EscherSense does this with your existing cameras, producing queue length, average wait time and service time per till in real time, so the decision to open another till is based on data.
Explore module →Retail heatmaps and customer path analysis
A retail heatmap is a colour-coded view of how much time shoppers spend in each zone of the sales floor. EscherSense builds it anonymously from your existing ceiling cameras, and its path analysis reveals the most common routes from entrance to checkout, movement between aisles and the dead zones that almost nobody visits.
Explore module →Retail staff productivity and workforce planning analytics
Retail staff productivity analytics measures how employees' time on the floor splits between serving customers and idle time, so rosters can be built around real footfall. EscherSense distinguishes staff by their uniform without identifying anyone; what is assessed is the shift pattern and store process, never the individual employee.
Explore module →This module is also covered in the solution family of CX Teknoloji, the team that builds EscherSense: cxteknoloji.com/en/cozumler/kisi-sayma/