Store Security Cameras Retail Store Security Camera Systems and Commercial Store CCTV Software
Surveillant runs AI video analytics on the store cameras you already have. It watches the register, the exit, the stockroom door, and the high-value aisle, and it tells you what happened instead of leaving a manager to scrub a recorder after close.
Works with any ONVIF or RTSP camera. Keep the cameras and the recorder you already paid for.
- Typical camera count
- 8 to 32 per store
- Cameras required
- Any IP camera you own
- Detects
- People, loitering, after-hours entry
- Multi-location
- One login, all stores
- Audio recording
- Off by default
Store security cameras cover four loss points: the point of sale, the entrance and exit, the stockroom and back door, and the highest-value merchandise on the floor. A typical US retail store runs 8 to 32 cameras. What separates a modern retail store security camera system from an old recorder is not the cameras, it is the software behind them. Recording alone catches nothing in real time, because nobody watches the monitors during a shift. AI video analytics reads every feed continuously, flags after-hours entry and a back door propped open, and lets a manager search months of footage in plain English instead of dragging a timeline. The last National Retail Security Survey the National Retail Federation published put shrink at 1.6 percent of sales, or $112.1 billion, on fiscal 2022 data, and on thin retail margins that is real money walking out the door.
Every Store Already Has Cameras. Almost Nobody Watches Them.
Walk into any retail store in America and you will find cameras in the ceiling. Walk into the back office and you will find a recorder nobody has opened since the last incident. That is the actual state of store surveillance: the recording works fine, and the watching does not happen at all. A single store generates tens of thousands of hours of footage a year, and a human being reviews a rounding error of it, always after the fact, always because something already went wrong.
The losses arrive from several directions at once and they do not look alike. There is the shoplifter who works the blind aisle and walks. There is the group that sweeps a high-value fixture and is out the door before an associate finishes a price check. There is internal loss at the register, the void, the fake return, the sweetheart discount for a friend, which the NRF's survey consistently places in the same league as external theft. And there is the back door, propped open on a warm afternoon during a delivery, which is how product leaves a store without anyone ever setting foot on the sales floor.
Then there is the liability side that has nothing to do with theft. A customer goes down on a wet floor in aisle four. Eight weeks later a demand letter arrives, and by then the footage that would have shown exactly when the spill happened and how fast it was cleaned has already rolled over and been overwritten.
None of this is a camera problem. A store owner with six locations is never going to hire someone to stare at a wall of monitors, and shouldn't have to. Whatever watches those cameras has to watch them on its own, and it has to hand back the handful of clips worth a manager's attention out of the millions of minutes that are not.
AI Video Analytics on the Store Cameras You Already Own
Surveillant is software, not hardware. It connects to your existing IP cameras over ONVIF or RTSP, pulls the streams into the cloud, and runs detection on them continuously. There is no truck roll, no proprietary camera to buy, and no second recorder in the back office.
The system knows your open and close times, so a person moving through the aisles at 3am raises an alert and the same person at 3pm does not. It knows a human from a plastic bag blowing across the lot, which is the difference between an alert you act on and the motion notifications everyone eventually mutes. It flags a stockroom door held open past a threshold you set. And it lets a manager type "person at the register after closing on Saturday" and get the clips back in seconds, instead of scrubbing a timeline for an hour.
For a chain, every store lands in one console. A loss-prevention lead can review last night's after-hours and back-door alerts across all twelve locations over a cup of coffee, which is the only review cadence that has ever survived contact with an actual retail operation.
- Real-time alerts on after-hours entry, loitering, and a propped back door, sent to a phone instead of a monitor nobody is watching
- Plain-English search across every camera and every day of retained footage
- People counting and dwell time, so the same cameras tell you about traffic and not just theft
- Clip export with a chain of custody, for the police report and the insurance claim
- One dashboard across every location, with role-based access per store manager
Where to Put Security Cameras in a Retail Store
Camera count follows loss points, not square footage. A 2,000 square foot boutique and a 2,000 square foot vape shop need very different coverage, because the merchandise and the risk are different. This is the placement most US retailers land on.
| Area | Cameras | What it is for |
|---|---|---|
| Point of sale | 1 per register | Overhead on the drawer and the hands. This is the camera that resolves voids, fake returns, and cash discrepancies. |
| Entrance and exit | 1 to 2 each | Face-height, facing in. The only shot that reliably identifies a person for a police report. |
| High-value merchandise | 1 per fixture or aisle end | Electronics, fragrance, spirits, designer goods. Covers concealment and organized sweeps. |
| Stockroom and back door | 2 to 3 | Internal loss and the propped-door delivery window. Frequently the single highest-yield camera in the store. |
| Sales floor overview | 2 to 6 | Wide coverage for slip-and-fall timelines and general movement. Not for identification. |
| Fitting room entrance | 1 | The entrance only. Cameras inside a fitting room are illegal everywhere in the US. |
| Parking and storefront | 2 to 4 | Smash-and-grab, vehicle plates, and after-hours approach to the door. |
Camera counts are typical ranges for US retail, not a rule. Coverage should follow your merchandise mix and your actual loss history.
Security Cameras by Store Type
Retail is not one problem. A liquor store's risk is the after-hours break-in and the underage sale, a jewelry store's is the smash-and-grab, a grocery store's is self-checkout and the dock. We keep a dedicated page for the store formats where the differences actually matter.
Self-checkout walk-offs, the receiving dock, and shrink on a one to three percent margin.
After-hours entry, the register, and the license conditions your state may attach to footage.
Smash-and-grab, case coverage, and the identification shot an insurer will ask for.
ATF recordkeeping context, burglary hardening, and the counter camera.
Forecourt, drive-offs, plates, and the overnight single-clerk shift.
The full loss-prevention picture: external theft, internal loss, and process error.
Four Ways to Run Store Cameras, Honestly Compared
Software on your existing cameras is not the right answer for everyone. Here is what each option actually costs and where each one wins.
| Option | Typical cost | Best for | Where it falls short |
|---|---|---|---|
| DVR or NVR only | Hardware you already bought | A single store that only needs footage after an incident | Nothing is watched live. Review is manual. A stolen recorder takes the evidence with it. |
| AI software on your cameras (Surveillant) | A few dollars per camera per month | Stores and chains that already own working IP cameras and want alerts, search, and cloud evidence without new hardware | No human is dispatched. You still need someone to act on an alert. Analog cameras need an encoder first. |
| Proprietary cloud camera platform | $600 to $3,500 per camera up front, plus licensing | A new build with no existing cameras and budget for a full rip and replace | You throw away working cameras and lock into one vendor's hardware. |
| Monitored remote guarding service | $50 to $200 per camera per month | High-risk sites that need a human to talk down an intruder and call police at 2am | By far the most expensive option, and it scales badly across a lot of cameras. |
If a live human response at 2am is the requirement, a monitored service is the honest recommendation, and many chains run one on a handful of high-risk sites while running software everywhere else. We compare the two in detail in our guide to video monitoring services versus software.
Store Security Camera Questions
How many security cameras does a retail store need?
Most US retail stores run 8 to 32 cameras. The count follows loss points, not floor area: one overhead per register, one to two at each entrance and exit, one per high-value fixture or aisle end, two to three on the stockroom and back door, two to six for wide sales-floor coverage, and two to four outside. A small boutique can be well covered with eight.
Are security cameras in stores legal?
Yes, in every US state, across the sales floor, registers, stockroom, and parking lot. None of those areas carries a reasonable expectation of privacy. Cameras are never permitted inside restrooms, fitting rooms, or lactation rooms. Audio is a separate question: the federal Wiretap Act permits one-party consent, but roughly eleven states require all-party consent, so most retailers leave audio off.
How much do store security cameras cost?
Hardware and installation for a typical store runs from a few thousand dollars to the low tens of thousands, depending on camera count and cabling. Software is the recurring part: basic cloud recording is about $2 to $8 per camera per month, AI analytics add-ons $3 to $15, and full cloud platforms with bundled proprietary cameras cost far more up front. Adding software to cameras you already own is the cheapest route.
How long should a store keep security camera footage?
Thirty days is the practical floor and 60 to 90 days is safer. Slip-and-fall claims, employment complaints, and chargebacks routinely surface weeks after the incident, long after a 14-day recorder has overwritten the evidence. Check whether your insurer, your franchise agreement, or a liquor or firearms license sets a minimum, because in several states they do.
Can I use my existing store CCTV cameras with AI software?
Usually yes. If your cameras are IP-based and speak ONVIF or RTSP, which nearly all cameras sold in the last decade do, Surveillant can pull their streams and add detection and search without replacing anything. Older analog cameras need an encoder, or a DVR that exposes an RTSP stream. The wiring and the mounting do not change.
Do security cameras stop shoplifting?
They help, but not on their own. Research on CCTV finds the clearest deterrent effect where cameras are combined with an active response rather than left recording unattended. A camera that nobody watches mostly produces evidence after the loss. A camera whose feed is analyzed, and where staff visibly act on what it surfaces, is what changes offender behavior. We walk through the evidence in our guide on whether security cameras deter crime.
Put AI on the store cameras you already own
Connect an ONVIF or RTSP camera and see the alerts and the plain-English search working on your own footage. No new hardware, no truck roll.