A camera bolted to a pole over a storage yard records whatever happens in front of it. Whether it stops anything from happening is a separate question, and the most complete answer available comes from a meta-analysis that is now seven years old.
Eric Piza, Brandon Welsh, David Farrington, and Amanda Thomas published “CCTV Surveillance for Crime Prevention: A 40-Year Systematic Review with Meta-Analysis” in Criminology & Public Policy in 2019 (https://onlinelibrary.wiley.com/doi/abs/10.1111/1745-9133.12419; author summary at https://ericpiza.net/2020/12/02/cctv-review/). The team screened 161 CCTV evaluations dating back to 1978, kept 80 that met their inclusion standards, and ran the meta-analysis itself on 76 of them. It updated a 2008 Campbell Collaboration review and, as of 2026, remains the most current large-scale meta-analysis of camera effectiveness anyone has published. Nothing of comparable size has replaced it. That gap matters: much of the camera industry’s deterrence claim rests on a study old enough to have teenagers, because nobody has repeated the work at that scale since.
What the meta-analysis found
Piza and his co-authors reported a 13% reduction in crime in areas with CCTV compared to matched control areas, an effect their 2019 study found statistically significant (https://onlinelibrary.wiley.com/doi/abs/10.1111/1745-9133.12419). Thirteen percent is real. It is also modest, and it is not the same claim as “cameras stop crime.” It is closer to “cameras, installed and used a certain way, move the number some,” and the “used a certain way” part is what most camera sales pitches leave out.
The gap between watched and recorded
The 2019 Piza meta-analysis split the studies by whether the camera systems were actively monitored or left to record without anyone watching (https://onlinelibrary.wiley.com/doi/abs/10.1111/1745-9133.12419). Schemes with active monitoring, meaning a person watching feeds or reviewing them as part of an operational response, showed roughly a 15% reduction in crime. Passive systems, cameras recording to a hard drive with nobody watching in real time, showed no statistically significant effect at all.
That distinction is the one this piece is built around. A camera that only records is not the same investment as a camera someone is watching, and Piza’s research says they don’t produce the same result. Schemes that combined CCTV with other measures, better lighting, guards, access control, performed best of all, at close to a 34% reduction in the same 2019 analysis. Cameras appear to help most as one piece of a system someone is actively running, not as the whole plan.
Where cameras work best, and where they don’t
Piza and his co-authors found the strongest effect in car parks, close to a 37% reduction, with a significant effect in residential settings too, around 12%, in their 2019 meta-analysis (https://onlinelibrary.wiley.com/doi/abs/10.1111/1745-9133.12419). Car parks are enclosed, have a small number of entry and exit points, and give a monitoring agent a clean view of a limited set of things that can happen there: a window breaking, someone loitering near a vehicle at two in the morning.
The same review found no significant effect in city and town centers, public housing, or public transport. Open, high-turnover environments dilute the signal. Too much ordinary activity is happening for a camera, or a person watching one, to reliably separate a threat from the normal churn of people moving through the space.
A Kentucky storage facility, warehouse, or equipment yard has more in common with the car park case than the city center case: a defined perimeter, a small number of access points, and long stretches when nobody has any legitimate reason to be there. That is why the car park and residential figures from the Piza study are the ones that matter for this kind of property, not the city center result.
Documentation still has value, even without deterrence
None of this makes recording-only systems pointless. A camera with no live monitor still leaves a record after the fact: who came through the gate at 3 a.m., what a damaged unit looked like the day before a claim was filed. Insurance disputes, tenant disagreements, and vendor disputes all move faster with footage than without it.
The Piza meta-analysis measured crime prevention specifically, not the value of having a record. A passive system that shows no measurable deterrence effect can still be worth the cost for documentation. The mistake is assuming it is also doing prevention work, when the 2019 research says it mostly is not.
The industry is formalizing verification
Camera monitoring is not the only place this shows up. On the alarm side, Seattle’s 911 center took in roughly 13,000 residential and commercial burglar alarm calls in 2023, and fewer than 4% of them were confirmed to involve a crime that led to an arrest or a written report, according to the Seattle Police Department’s Monitored Alarms program page, published in 2024 (https://www.seattle.gov/police/community-policing/community-programs/monitored-alarms). As of October 1, 2024, Seattle police dispatch only to alarms backed by audio, video, panic button, or eyewitness evidence that a crime is in progress. An alarm signal alone, with nothing behind it, no longer gets a response in that city.
The monitoring industry is building the same standard nationally. The Monitoring Association’s ANSI/TMA-AVS-01 standard, revised in December 2024, sets up a scoring system called Alarm Validation Scoring, running from Level 0, “no call for service,” to Level 4, “confirmed threat to life,” so a monitoring center can tell a 911 dispatcher how credible a given signal is before anyone gets sent (The Monitoring Association, ANSI/TMA-AVS-01, 2024, https://tma.us/standards/tma-avs-01-alarm-validation-standard/). Certification is UL-audited, and monitoring centers run by companies including ADT, Brinks Home, Guardian Protection, Rapid Response, SimpliSafe, and Vector Security are certified under it.
The direction is the same one the Piza meta-analysis points to. An unverified signal, whether it is an alarm contact or a camera nobody is watching, carries less weight than one a person has looked at and judged.
A decision rule for which sites justify live monitoring
An owner deciding between recording-only cameras and actively monitored surveillance can run each site through three questions.
Is the site enclosed, with a small number of access points, like a fenced yard, storage facility, or parking area, or is it open and high-turnover, like a public sidewalk or a busy retail frontage? The 2019 Piza meta-analysis found its strongest effects in the enclosed, limited-access category. Open, high-traffic sites are closer to the city center case, where the same study found no significant effect.
Does the site have hours when nobody should be there at all? A storage facility overnight, a lumber yard after close, a construction site on a weekend, all carry a baseline expectation: no legitimate activity during those hours. That baseline is what makes analytics and human judgment useful, because anything that shows up against it is worth a look.
What would a response require? If the answer is that someone needs to see what is happening and decide whether to step in, that is a monitoring problem, not a recording problem. If the answer is that proof after the fact is what matters, a passive system covers it.
A site that is enclosed, has real off-hours periods, and needs a real-time decision when something happens is a candidate for active monitoring. A site that is open, busy, and mainly needs a record for insurance or dispute purposes is usually well served by cameras and an NVR on their own, without the added cost of a monitoring contract.
What this looks like done well
Versys builds both kinds of systems. Some clients need a handful of IP cameras and an NVR recording locally, which is the right setup for a small footprint where documentation is the goal. Others, particularly storage facility owners with large perimeters and long unattended overnight periods, need more: a geofenced property boundary built from analytics, live monitoring that can flag a person or vehicle after hours, and an agent on the other end who looks at what triggered the alert, decides whether it is a problem, speaks to whoever is on the property through a paging horn, and calls police if the situation calls for it. The company has built that model for storage operators across Kentucky and Southern Indiana, along with camera mounts on existing buildings and standalone towers, including mobile towers, for sites without a structure to mount to yet.
Deciding what a site needs starts with what happens on that property, not with how many cameras fit on it. An owner in Kentucky or Southern Indiana who wants a direct answer about whether a given property justifies monitoring, or just recording, can call Versys at (270) 358-2200 and walk through it.
