In a store, cameras serve three separate purposes: loss prevention, operational control and personal safety. If you expect all three from the same camera, you end up with one of two outcomes: a system that sees everywhere but can’t prove any incident, or a placement that pushes the privacy limit. Below we explain the purpose and required level of detail of cameras for a single store’s checkouts, entrance, sales floor, goods receiving and stockroom points, then how to build a repeatable template for a retail chain.
What cameras solve in retail
Retail losses fall into four groups: external theft, internal loss (staff-related), process and recording errors, and supplier-related discrepancies. The camera’s task differs in each group:
- External theft: identifiable images of people at entry and exit, behavior observation in high-risk aisles.
- Internal loss: matching checkout transactions (refunds, voids, discounts, no-sale cash drawer openings) with video.
- Process errors: the source of discrepancies between boxes counted at goods receiving and the delivery note.
- Personal safety: arguments at the checkout, slip-and-fall incidents, staff safety at closing time.
This list sets the placement priority of the corporate camera system in the store. The areas a loss prevention program focuses on (such as the refund process and high-value aisles) directly change the camera’s level of detail and retention period.
Area-based placement table
The table below is a starting template for a single store. Target levels are written according to the pixel density references in the IEC 62676-4:2014 table.
| Field | Purpose | Target level | Placement note |
|---|---|---|---|
| Entrance–exit | Identifying people | Identification (~250 px/m) | From inside towards the door, close to face height |
| Checkout (each) | Transaction and cash | Identification (belt and drawer) | From the ceiling, behind the cashier |
| In front of checkouts | Customers and queue | Recognition | Facing the checkout line |
| Sales floor | Behavior observation | Observation | Aisle heads, panoramic |
| High-value aisle | Hand movement with products | Recognition | Close to the shelf, narrow angle |
| Goods receiving door | Box count, vehicle | Recognition | From the inside of the door |
| Stockroom / back area | Access and stock movement | Observation, recognition at the door | Stockroom door and rack aisle |
| Cash office / safe room | Cash counting | Diagnostics | From above the table, single entrance |
A common mistake with the entrance camera is placing it facing the door, and therefore the daylight outside. Light coming through the door glass creates backlight; faces turn into silhouettes. In the right placement, the camera meets the person entering a few steps inside the door, under indoor lighting. For people leaving, a second camera can be placed near the door, looking inward. If backlight is unavoidable, a model with true WDR is chosen.
On the sales floor, the goal isn’t to see every shelf in detail but to be able to follow a person’s movement through the store without interruption. Cameras at aisle heads and panoramic cameras in the middle of the ceiling do this job with few devices. A panoramic camera loses detail at the edges. It isn’t used for identification.
Checkouts and POS–video integration
The checkout camera is placed behind and above the cashier, so it sees the cashier’s hands, the area where products are scanned, the cash drawer and, if possible, the customer display. Seeing the checkout screen and the card terminal keypad isn’t wanted; a privacy mask is applied to the terminal zone. Test separately whether this mask is preserved in exported footage as well. Details are in the privacy masking on cameras article.
The real value of the checkout camera emerges when it’s combined with POS data. In POS–video integration, transaction lines produced by the checkout software (product, amount, refund, void, discount, payment type) are passed to the VMS and linked to the image of the relevant checkout camera. Two methods are common in practice:
- Text overlay: transaction lines are burned into the recording as text. It’s simple to set up but can’t be searched; you have to read it by watching the footage.
- Event/metadata integration: transaction data reaches the VMS as separate data and can be queried. Searches such as “all refunds at checkout 3 in the last 30 days” are possible; each result opens the relevant video segment.
Data transfer is done via serial port, TCP socket, file or API, depending on the method the POS software supports. Whatever the method, accuracy depends on two things: POS terminals and cameras using the same time source, and each terminal being mapped to the right camera. If the mapping table isn’t updated when a checkout number changes or a terminal is moved, every search result opens the wrong footage. And noticing this usually takes until an incident review.
If exception reports (refunds, voids, no-sale drawer openings) aren’t reviewed regularly, the integration remains only a post-incident search tool. Define, together with the integration, which report the loss prevention team will review and how often.
Goods receiving and stockroom areas
A common source of store stock discrepancies is goods receiving. The camera should see the delivery vehicle’s rear door, boxes being brought in and the counting table at recognition level. The goods receiving door usually opens onto outdoor light; the same backlight problem exists here. The camera is placed on the inside of the door, at a side angle to the door.
The number of cameras in the stockroom and back area is usually kept low, but three points aren’t skipped: the stockroom entrance door (who entered, when), the cage or cabinet for high-value products, and the back exit door. In internal loss investigations, the footage opened most is the back exit door. If a door sensor or alarm panel output is connected to the VMS as an event, every opening of the door is found within seconds. For larger storage and shipping points, the warehouse and logistics camera design article gives a detailed area analysis.
A standard template for retail chains
In a chain, designing every store from scratch makes maintenance and review harder; every store becomes a separate exception. Instead, a standard camera template is defined by store format (such as small, medium, large). The template fixes:
- The camera naming rule: names containing the store code, area and sequence number, such as
STR-034-CHECKOUT-02andSTR-034-ENTRANCE-01. - Camera model, lens range and mounting bracket for each area type.
- Stream profiles: recording stream resolution, frame rate and codec; substream settings for remote viewing.
- Retention period and recording mode (continuous or event-based).
- User roles: store manager, regional manager, loss prevention, IT.
- Acceptance test form: a reference frame and image level check for each camera.
The template doesn’t ignore differences on site (ceiling height, column positions, glass façades); it only ensures decisions are made consistently. Deviations per store are recorded and fed back into the next version of the template. Topics specific to multiple sites, such as central monitoring, user management and firmware processes, are covered in the multi-site camera management article, and the infrastructure side on the multi-site central camera management page.
Privacy limits: fitting rooms and staff areas
The limit of a store camera is drawn by legal and ethical rules, not technical ones. Fitting rooms, toilets, nursing rooms and areas where staff change are the places with the highest expectation of privacy; they aren’t covered by cameras. This isn’t open to debate. If a camera that sees the entrance to the fitting room area is used, the inside of the curtain or door mustn’t be in the field of view, and a fixed privacy mask should be applied if needed.
The staff rest area and canteen are kept out of scope unless genuinely needed for security. Audio recording on cameras should be off by default; audio recording needs a separate assessment. Obligations to inform customers and employees, entrance warning signs and employee notices are addressed with the legal and KVKK officer before installation. You can find the general framework in the camera privacy notice under KVKK article and in publications of the Turkish Personal Data Protection Authority. This section is general information, not legal advice.
Implementation checklist
- Is there a separate camera for each checkout, and is the mapping table with POS terminals up to date?
- Are the card terminal keypad and checkout screen masked or out of the field of view?
- Was the entrance camera tested in daytime under backlight conditions?
- Are the goods receiving door and back exit door seen at recognition level?
- Are fitting rooms, toilets and staff changing areas strictly out of scope?
- Is audio recording off?
- Do cameras, the NVR and POS terminals use the same NTP source?
- Was the store set up according to the naming and role structure in the chain template?
For template preparation by store format and implementation, see the scope on the store and retail video surveillance page.
Frequently asked questions
Where should the checkout camera go in a store?
The checkout camera is placed behind and above the cashier, seeing the area where products are scanned, the cashier’s hands and the cash drawer. The card terminal’s keypad is left out of the field of view or masked.
What is POS–video integration for?
It links POS transaction data to the image of the relevant checkout camera. Transactions such as refunds, voids, discounts or no-sale drawer openings become searchable, and each result opens the relevant video segment. For it to work correctly, the checkout–camera mapping and time synchronization must be kept up to date.
Can a camera be placed in front of fitting rooms?
The inside of fitting rooms never falls within camera coverage. If a camera seeing the outside of the fitting room area is to be used, it must be verified that the inside of the rooms isn’t in the field of view, and a fixed privacy mask applied if needed. The topic is assessed with the legal and KVKK officer before installation.
What should a camera template for retail chains include?
The camera naming rule, camera model and lens range by area type, stream profiles, retention period and mode, user roles and the acceptance test form are the template’s basic components. The template is first validated in a pilot store.