Use case · Liability and incident documentation
The claim arrives weeks later, and the footage is gone
This is the use case where a system is either worth what you paid for it or is worth nothing at all, and the difference was decided at commissioning, not at the moment of the incident. Three things fail: retention, time accuracy and export.
- CSLB C-7 #976737
- BSIS ACO #7501
- CMMC certified
- SAM.gov registered
- Est. 2008
The problem
Nobody finds out on the day
A customer slipped in the entrance in March. A contractor says a forklift clipped his van in the yard. A tenant reports that their delivery was signed for by somebody who was not them. An employee has made an allegation about something that happened in a corridor. A regulator wants to see the loading area on a particular afternoon. A driver disputes the condition of a vehicle at collection.
None of those arrives on the day. They arrive as a letter, a portal notification or a phone call three weeks, two months or a year later, and every one of them turns on the same question: what does the record show. If your system has it, the conversation is short and frequently ends in your favour. If it does not, you are arguing from memory against somebody with a written account, and the absence of footage is itself read as something.
The operational cost of getting this wrong is a specific, exhausting kind of cost. It is the hours spent scrubbing recorders for a date nobody can pin down. The realisation that the camera covering that corner was replaced in June and the old one's recordings went with the box. The export that will not open on anybody else's machine. The timestamp that is eleven minutes out from the incident report and now has to be explained. And the settlement that gets paid because disputing it would cost more than paying it.
It is hard to solve because everything about it is invisible until the day it matters. A system with no coverage of the entrance looks exactly like a system with coverage of the entrance, from the outside. A recorder retaining nine days looks the same as one retaining ninety. A clock eleven minutes out looks like a clock. Nobody discovers any of it during normal operation, because normal operation never asks the system a question.
What the evidence shows
Why the record is probably yours to hold
There is no statistic on this page for how often footage decides a claim, because nobody publishes one. What is sourceable is the surrounding picture, and it points consistently in one direction.
What we will not claim. We could not find a primary source for how often video evidence changes the outcome of a claim, a dispute or an investigation. The figures circulating on that subject come from vendors and comparison sites, which this site excludes on principle. So there is no such number here. Everything below measures something adjacent and is labelled as measuring that.
Very little is cleared
In 2025 San Diego Police Department cleared 1,456 of 21,897 reported property crimes — 6.6%, about one in fifteen. Nationally the FBI reports 17.4% of reported property crime cleared. Cleared means by arrest or exceptional means, not a conviction, and clearances in one year may relate to earlier offences. COUNT, ratio computed.
Reported is not the same as occurred
The Bureau of Justice Statistics estimates 30.5% of property crime and 40.7% of burglary were reported to police in 2024. That survey covers households and persons aged 12 and over — it does not measure crime against businesses, so it cannot tell you a commercial reporting rate. ESTIMATE from a sample survey.
Most incidents never reach anybody
The National Retail Federation reports that 63% of retailers report fewer than half of their incidents to law enforcement. For that share of events no external record is created at all — the only account that exists is the one the business holds. ESTIMATE from a voluntary industry survey of 66 companies.
Sources. Clearance: Federal Bureau of Investigation, Crime Data Explorer, offences and clearances, agency ORI CA0371100 for San Diego Police Department — cde.ucr.cjis.gov/LATEST/webapp; the 17.4% national figure is also stated in the FBI's UCR Summary of Reported Crimes in the Nation, 2025. Period: calendar year 2025. Type: COUNT with the ratio computed. Reporting rates: Bureau of Justice Statistics, Criminal Victimization, 2024 (NCJ 310547), published 29 September 2025, Table 4 — bjs.ojp.gov; Type: ESTIMATE from a sample survey, and BJS began an instrument redesign in 2024 so comparisons with earlier years carry that caveat. Retail reporting: National Retail Federation, The Impact of Theft & Violence 2026, released 30 July 2026 — nrf.com; Type: ESTIMATE from a voluntary, self-selected industry survey of 66 retail companies and 143 brands, which is not a census and not government data. All retrieved: 2 September 2026.
And the direction of travel, stated plainly: recorded crime is falling. Commercial burglary recorded by SDPD dropped from 1,553 offences in 2020 to 1,014 in 2025, the FBI puts estimated national burglary and larceny-theft rates at twenty-year lows for 2025, and the NRF's own most recent finding is a 12.4% decrease in shoplifting incidents in 2025. Nothing on this page argues from a rising risk. It argues that when something does happen, the usable record is very likely to be yours.
Why it resists solving
Four failures, and none of them look like failures
Retention that is shorter than the notice period
The single most common one. Retention is arithmetic — camera count, resolution, frame rate and storage — and it changes silently whenever any of those changes. Four cameras added to an existing recorder shorten it for every channel, and nobody is told. The number that should drive it is not a default, and it is not thirty days because thirty is a round number. It is how long your business genuinely takes to learn that something happened, which for a liability claim is usually far longer than for a theft.
A clock that has drifted
Every use of footage is a use against something else — an incident report, a door event, a transaction, a delivery note, a phone record. If the recorder disagrees with all of them the dispute stops being about the event and becomes about your system. Where regulation is specific it is specific about exactly this: California's cannabis surveillance regulations require system time to be measured against United States NIST standards, which is a requirement on licensed cannabis premises and not a general rule, but it tells you what the people who write rules think matters.
Coverage of the wrong thing
Cameras get positioned to cover area, because area is what a plan shows and what a quote is priced against. Liability happens at specific places — the threshold, the ramp, the wet floor by the entrance, the loading edge, the lift lobby, the counter — and at a framing that shows the surface, the person and the sequence — not a wide view containing all three at twelve pixels each.
Nobody ever tested getting it out
An export that needs a manufacturer's proprietary player, or that drops the timestamp, or that only one person in the business knows how to produce, is a system that has the footage and cannot give it to you. This is tested at handover on the systems we commission, and it is one of the first things we check on a system we did not.
What reduces it
Six things that change the outcome
This use case is unusual in that most of the work is commissioning and maintenance instead of hardware.
Positions chosen for what they must prove
Thresholds, ramps, ledges, lift lobbies, counters and loading edges, framed to show the surface, the person and the sequence — specified against the question that will be asked rather than against square footage. That is commercial security camera installation.
Retention, clock and export, verified
What the system actually retains against what it claims, a time-synchronised recorder, a tested export path and a check that every channel is still writing. All invisible faults, all found by looking. That is CCTV maintenance and repair.
Storage and bandwidth that support the number
Retention is decided by what is behind the camera as much as by the recorder. Undersized pathway, a saturated link or a shared switch quietly force the compromises that shorten it. That is structured cabling.
A door record to line the video up against
An access event with a credential, a door and a time turns "somebody was in the corridor" into a specific, corroborated account — and it is the second record that makes the first one hard to argue with. That is commercial access control.
Somebody who saw it at the time
For incidents where a contemporaneous human account matters as much as the recording, watched coverage during the hours it happens produces both, and produces them at the time, not in retrospect. That is remote video monitoring.
A signal record for the events nobody witnessed
Detection with a timestamped, monitored history gives you an independent account of when a door opened, a zone was entered or a system was armed — useful precisely when nobody was there to see it. That is commercial alarm monitoring.
Who has it worst
Where documentation is the whole point
Every business has this exposure. Three types have it as a standing operational condition, not an occasional event.
- Retail. Public trading space, wet floors, thresholds and ramps, a constant flow of people with no relationship to the business, and disputes about transactions that need footage and a till record read against each other. That is retail security and loss prevention.
- Managed office and multi-tenant property. Common parts the landlord is responsible for, contractors on site constantly, tenant-versus-tenant disputes, deliveries signed for by somebody, and a managing agent who has to answer for all of it in writing. That is office and property management security.
- Apartment communities and managed residential property. Common areas, parking, package rooms and resident disputes, with the added constraint that coverage is bounded by privacy expectations and that what may be recorded is a question with a real answer. That is multifamily and apartment security.
The other verticals are on all industries and the full service list is on all commercial services. Where the incident being documented is a theft, the related problem is theft and burglary; where it happens in the lot, it is parking lot incidents.
Two boundaries. We will not tell you what an insurer, a regulator or a court will accept — that is not our field, and a confident answer from a security contractor is worth nothing in that room. What we will do is describe exactly what the system captures, retains and produces, so your own advisers can judge it. And we do not hold a C-10 or a C-16, so we do not work on fire alarm systems; that is coordinated with your own licensed fire alarm contractor. We work on commercial premises only and not on private residences, though managed residential property is in scope.
At survey
What we would look at on your site
This survey spends more time in front of the recorder than on the building, because that is where this problem actually lives.
What the system is really retaining
Not the configured figure — the oldest recording actually present on every channel. These differ more often than people expect, and a channel that stopped writing three weeks ago is a common and completely silent finding.
How long you take to hear about things
The number that sets retention. We want real examples: the last claim, the last dispute, the last time somebody asked for footage, and how long after the event each of those arrived.
The clock, against a reference
Checked on the recorder, on the access system and on anything else that timestamps, because the useful property is not that each is right but that they agree with each other and with the outside world.
An export, actually performed
We do one, on your system, and open it somewhere else. If it needs a proprietary player or loses the overlay, you should know that now, not while an adjuster waits.
The specific places incidents happen
Thresholds, ramps, changes of level, wet zones, loading edges, lift lobbies and anywhere a contractor works. Walked with your own incident history in hand instead of from a floor plan.
Who can view, export and delete
The account list, what each can do, whether there is a shared login, and whether the system records who looked at what. Evidence that nine people have already viewed is worth less than evidence that one named person retrieved.
Findings go in writing whether or not you go on to contract with us. Where the honest answer is that your existing system needs a configuration change, not new hardware, that is what the report will say — and on this particular problem that is a common outcome.
Common questions
Incident documentation questions
Almost always retention, and almost always because it was left at whatever the recorder defaulted to. Retention is not a setting you choose in isolation — it is the arithmetic of camera count, resolution, frame rate and how much storage was actually bought, and adding four cameras to an existing recorder silently shortens it for every channel. The second most common cause is that the recorder was overwriting a failing drive, or had stopped writing one channel weeks earlier and nobody looks at a camera until they need it. The fix is unglamorous: work out how long your business genuinely takes to learn that something happened, design storage against that number, not against a default, and check that what the system claims to be retaining is what it is actually holding.
Long enough to still hold it when you first learn you need it, which is a business question instead of a technical one and is different for every operation. A retailer who reconciles weekly is in a different position from a property manager whose first notice of an incident is a solicitor's letter, or a distributor whose shortage shows up at a quarterly count. Some premises have a statutory floor instead of a judgement: a licensed cannabis premises in California, for example, is required by the state regulations to retain surveillance recordings for a minimum of ninety calendar days. That figure applies to cannabis licensees and to nobody else, and we mention it only as an illustration that where a rule exists it beats an opinion. Your insurer or your lease may also set a period, and if so that is the number we design to.
It matters more than almost anything else on the system, and it is the single most common defect we find on installations done by somebody else. Every use of footage depends on lining it up with something outside the video — an incident report, a transaction log, a door event, a delivery note, a witness account, a phone record. If the recorder clock has drifted, everything you produce disagrees with everything else, and the argument stops being about what happened and becomes an argument about your system. It is also worth noting that where regulation gets specific it gets specific about this: the California cannabis surveillance regulations require system time to be measured against United States NIST standards. Time synchronisation is a commissioning step and a maintenance check, and it is cheap to get right.
That depends on the platform, and it should be tested at handover, not discovered under pressure. Some systems export a plain video file that opens anywhere. Some export a proprietary container that needs the manufacturer's player, which is a problem when the person who needs to watch it is an adjuster, an investigator or a solicitor. Some will only export in a way that loses the timestamp overlay, which removes the thing that made it useful. We test an export as part of commissioning, show whoever will actually have to do it how to do it, and tell you plainly if the platform you already own is difficult in this respect — that is worth knowing before you need it rather than after.
Fewer people than usually can, and it should be a deliberate decision instead of a shared login. The practical arrangement on most commercial sites is a small number of named accounts with the ability to review and export, a larger number who can view live if they need to, and a record of who looked at what and when. That last part is worth more than it sounds: if footage of an incident has been viewed by nine people and exported by an unknown account, its usefulness has been damaged before anybody argues about it. Access control on the system is as much a part of incident documentation as the cameras are, and it is a configuration decision taken at handover.
Our default is no, and we will not change it casually. California treats the intentional recording of a confidential communication without the consent of all parties as prohibited under Penal Code section 632. The statute defines a confidential communication as one carried on in circumstances that may reasonably indicate a party desires it to be confined to the parties, and it expressly excludes circumstances in which the parties may reasonably expect the communication may be overheard or recorded — so the rule and its exception have to be read together, and whether a given lobby, corridor or yard falls inside or outside the definition depends on the facts of that space. That is not a determination we make for you. This is not legal advice: if you want audio anywhere on the property, take it to your counsel first and we will design to whatever they tell you.
We do not know, and we are not going to invent a figure. We looked for a primary source on how often footage changes the outcome of a claim, a dispute or an investigation, and there is not one — the numbers circulating on this subject come from vendors and comparison sites, which is the category this site excludes on principle. What we can point to is sourced and adjacent: San Diego Police Department cleared 6.6% of reported property crime in 2025, so an external investigation resolving your incident is not the likely outcome; and the National Retail Federation reports that 63% of retailers report fewer than half of their incidents to law enforcement, so for a large share of events no external record is created at all. Both point the same way without pretending to measure the thing nobody measures.
Four things, and all four are decided before an incident, not during one. Coverage of the place the event will actually happen, at a framing that shows what needs to be shown — which usually means a specified position, not a general view. Enough detail at that distance and in that light to support whatever the footage is meant to establish. Retention long enough to still hold it when you first hear about it. And an accurate clock, so it agrees with every other record. A system can be fully working, fully covered and completely useless on all four counts, and that combination is what we see most often when a business tells us their cameras did not help.
Licensed, yes: CSLB #976737 and BSIS ACO #7501, both checkable on the state registers. What an insurer or a court will accept, no — and you should be wary of any contractor who answers that one. Admissibility turns on facts about your incident and your jurisdiction that we are not qualified to weigh, and a camera company's opinion on it is worth nothing in a hearing. What we can do is make sure the system is not the reason the footage fails: an accurate clock, a retention period you chose deliberately, and an export that leaves the recorder with its metadata intact. Ask your own counsel or your broker what they need; we will build to it.