Skip to content
UA Discuss your project
  1. Home
  2. Resources
  3. A store visitor counter on the cameras you already have: how it counts and what it shows

A store visitor counter on the cameras you already have: how it counts and what it shows

7 min read · Updated 2026-11-11

Author: DuoVision team

For: store and chain owners, retail operations directors, store managers who want to know footfall and conversion

Summary

  • A camera looking at the entrance counts people who cross a set line, separately for entering and leaving. No separate sensor or new hardware is needed if the existing camera sees the entrance at a suitable angle.
  • The numbers give footfall by hour and day of the week. Alongside till receipts it gives conversion, and alongside the queue counter it shows when there are not enough tills.
  • Accuracy depends on camera angle, entrance width, lighting and how people enter: in groups, with strollers, past the door. So it is first measured, comparing with a manual count on the same frames.
  • The system does not recognize faces and does not follow individual people between cameras: it stores numbers by zone and hour.

For: store and chain owners, retail operations directors, store managers who want to know footfall and conversion

How the camera counts visitors

On the frame of a camera looking at the entrance, we set a virtual line or a narrow zone by the door. The system detects people in the frame and records when someone crosses the line toward the entrance or the exit. Each crossing becomes an entry in the counter: time, camera, direction.

From the two series, entries and exits, you also get the current number of people in the store: entries minus exits. The error accumulates, so in practice it is reset, for example to zero after closing, when nobody is in the store. A separate zone lets you leave out employees if they pass through the same entrance; this is usually solved with a separate staff entrance or by excluding a zone. Where a person came from and who they are, the system does not find out.

What the counter shows

By itself the counter gives a series of numbers: how many people came in during each hour of each day. Several useful reports are built from it.

IndicatorWhere it comes fromWhat it is for
Footfall by hour and day of the weekCamera at the entranceShows peak hours and dead hours, helps plan shift schedules
Comparison of days, weeks, promotionsCamera at the entranceWhether a promotion really brought more people or only the number of receipts changed
Conversion: receipts per entryCamera and till systemShows hours when there are many people and few purchases
Queue length at tillsCamera over the checkout areaShows when open tills are not enough; details in the article queues at tills (in Ukrainian)
People in the store nowEntries minus exitsUseful for enclosed spaces with a limit on visitors
Comparison of stores in a chainAll entrance camerasThe same counting method for every location

An example of conversion with illustrative numbers: 200 people came in during an hour, the till rang 60 receipts, so conversion is about 30%. The number is useful over time: does it fall in hours when a till is idle or on days when the layout was changed.

Conversion by receipts: what it shows and what it does not

A receipt and a visitor do not match one to one. A family of three often buys on one receipt. Someone comes in to return an item or ask a price and buys nothing. Someone pays with several receipts. So conversion by receipts is an indicator for comparing hours, days and stores with each other, and it does not give the exact share of people who bought.

That is usually enough for the decisions it is calculated for: where there are many people and few receipts, the question is about the layout, product availability or till speed. More about scenarios for stores is on the retail page.

What accuracy depends on

Counting accuracy at the entrance is determined by the software together with how the camera is placed. The main factors:

  • Angle. A camera above the door, pointing down, sees people separately from each other and counts best. A camera looking at the entrance from the side or from deep in the store sees silhouettes that overlap.
  • Entrance width. Through a narrow door people pass one at a time, through a wide one in groups and several rows.
  • Groups, children, strollers. A tightly moving group can be counted as fewer people than there are. A child in arms or in a stroller also complicates counting.
  • Light. Backlight from the street, glare on the glass of the door and a dark evening reduce image quality.
  • Partitions and vestibules. If there are double doors, the camera must look at one of them, otherwise one person is counted twice.
  • People who stand in the doorway: a security guard, an assistant, a smoker. They are excluded with a zone or with a rule on time spent on the line.

So we do not promise one accuracy figure for all stores. We set the counting up for your site and show the result on your cameras before launch, and you can compare it with what a person counts on the same frames at different hours: morning, lunch, evening, weekdays and weekends. Which camera suits is covered in the article cameras with artificial intelligence: which suit.

A camera that is already there and a separate sensor at the entrance

Separate visitor counters come in different kinds: infrared, with their own camera over the door, with depth analysis. They are installed over a specific entrance purely for counting. It is worth comparing on practical points.

ParameterA camera that is already thereA separate sensor at the entrance
HardwareNothing new is bought if the camera sees the entrance at a suitable angleBought and mounted for each entrance
Angle and heightSet by how the camera was placed for security; may be worse for countingInstalled specifically for counting, so the angle is chosen to be optimal
What else the same stream can doQueues, zones, people in the store, other rules, a card with a frame for checkingUsually only counting
Checking the numberThere is a frame: you can view the moment and see why the number is what it isDepends on the model, often only a number without a frame
A narrow entrance with a clear passageWorks if the angle is suitableOften enough and reliable

If security cameras look at the entrance from a poor angle, and the entrance is wide with many people, a separate sensor over the door may give a better result. We do not claim that a camera on the stream is always more accurate. For some stores it suits, and the benefit is that you do not need to buy separate hardware and queues and other rules are available at once. For the rest it is more convenient to change the camera angle or add one camera right over the entrance.

Shopper privacy

A counter needs only anonymous counting. The system stores numbers: how many people, at which entrance, at what time. It does not recognize faces and does not follow a person between cameras.

The video recording of shoppers itself remains processing of personal data if a person can be identified in it. Article 12 of Law No. 2297-VI requires informing the person about the collection of data, so people should be told about video surveillance at the entrance, for example with a sign, and article 6 requires that the composition of data not be excessive in relation to the purpose (part 3). More in the article video analytics and personal data.

Where to start

  1. Decide which decision the counter must support: shift schedule, evaluating promotions, conversion or queues.
  2. Look at how the existing camera sees the entrance: angle, width of the passage, lighting, whether there is a vestibule.
  3. Prepare till receipts for a few weeks to compare them with entries by hour.
  4. Compare the count with a manual one on the same frames at different hours.
  5. Only after that rely on the reports.

We connect to the cameras you already have, set the counting up for your site and show the result on your cameras before launch. You can write to us through contacts. About losses that are also visible on store cameras, read the article where store losses come from.

Where to apply this

Sources

  1. Law of Ukraine "On personal data protection" of 01.06.2010 No. 2297-VI (article 6 part 3, article 12 part 2; edition of 14.06.2025; checked 11.11.2026, in Ukrainian)

Let’s discuss your site

Request a consultation or a project estimate. We will tell you what your cameras can monitor, where it is best to process the video and how much it will cost.