Start with your own loss history
The most credible ROI model uses your own data. Industry statistics are useful for context, and our roundup of business theft statistics is a good place for background, but a finance reviewer will trust what happened at your sites more than a national average.
Gather the last two or three years of records for each site.
- Incident reports, police reports and insurance claims.
- Invoices for stolen items, repairs and replacements.
- Deductibles you paid.
- Hours of downtime or delay after each event.
- Staff time spent on reports, calls and cleanup.
If records are thin, that is itself a finding. Start tracking now, since a baseline makes the next year's model much stronger. Where you are missing numbers, use cautious estimates and label them as estimates.
The cost categories to count
Losses are larger than the item that went missing. Count each category separately so the model is transparent.
- Direct loss. The value of stolen, damaged or vandalized property.
- Deductibles and uncovered losses. What your insurance did not pay.
- Downtime and delay. Lost revenue or schedule slip when operations stop or work has to be redone.
- Cleanup and repair. Labor and materials to restore the site.
- Staff and management time. Hours spent on reports, claims and follow-up.
- Premium effects. Any change in premium or coverage terms that your insurer confirms in writing.
Insurance deserves a conversation. Insurers vary widely in how they treat security measures, and some require certain protections as a condition of coverage. Our article on construction site security cameras and insurance covers what to ask. Do not assume a discount. Count one only if you have it in writing.
Estimate what changes
Cameras do not stop every incident, so the model has to estimate how much loss changes. Research gives a direction, not a promise. A 2019 systematic review of 76 studies found that CCTV was associated with a reduction in crime in target areas, with larger effects in parking lots and for actively monitored systems than for passive recording-only systems. Your site is not the average of those studies, so treat the finding as a reason to model scenarios.
Build three scenarios.
- Cautious: the program prevents or limits only a small share of losses.
- Expected: a moderate share.
- Strong: a large share.
Judge the investment against the cautious case. If it still pays back or is acceptable, the decision is safe. If it only works in the strong case, the program is a gamble. Read more about the evidence in our piece on whether security cameras deter crime.
The formula
Keep the math simple enough for anyone in the room to follow.
- Annual benefit = annual loss avoided + annual cost avoided in downtime and time.
- Annual net benefit = annual benefit minus annual program cost.
- ROI = annual net benefit divided by annual program cost.
- Payback period = upfront cost divided by annual net benefit, where upfront cost applies.
Use the same period for both sides. If the program cost is monthly, convert it to annual.
Worked example (illustrative only)
The loss figures below are hypothetical round numbers chosen to show the arithmetic. They are not VDS data, not a VDS price, and not a prediction for any site. Replace every number with your own.
Suppose a yard has the following yearly history:
| Item | Hypothetical yearly amount |
|---|---|
| Stolen or damaged property | $60,000 |
| Deductibles and uncovered loss | $10,000 |
| Downtime and rework | $20,000 |
| Staff time and cleanup | $10,000 |
| Total annual loss | $100,000 |
Suppose the cautious scenario assumes that a monitored program avoids half of that loss, and the program costs, for illustration, 30 percent of the yearly loss (your own quote goes here).
- Annual benefit: $50,000.
- Annual program cost: 30 percent of the loss.
- Annual net benefit: 50 percent minus 30 percent of the loss, or 20 percent.
- ROI: 20 divided by 30, about 67%.
If the expected scenario avoids three quarters of the loss, the net benefit is 45 percent of the loss and ROI is 150%. Change the assumptions and the answer moves, which is the point. The model shows which inputs drive the result, so you know where to gather better data.
What the model leaves out
Some benefits resist numbers. Safety of employees, protection of customers, reputation, contract eligibility and the ability to prove what happened all matter. List them next to the math without assigning dollars, so the financial case stays credible while decision makers still see the full picture.
Risks to the model matter too. If the program is poorly placed or unmonitored, it will not deliver the assumed reduction. Cost estimates should include setup, service and any monitoring, and the remote video monitoring cost guide breaks down the pieces to ask about.
Using the result
Present the model with its assumptions listed, the cautious case first and a note on what data would improve it. Offer a review date, for example after six months, to compare actual incidents to the forecast. That turns the ROI from a sales claim into a measurement. The VDS ROI page has additional material.
Common mistakes
Several errors undermine otherwise sound models.
- Using only the value of what was stolen and ignoring downtime and time.
- Assuming cameras will prevent every incident.
- Comparing a monthly cost on one side to an annual loss on the other.
- Counting an insurance discount that has not been confirmed.
- Leaving out setup, service or monitoring from the program cost.
- Never going back to compare the forecast to what actually happened.
Avoid these and your model will hold up in front of a skeptical finance reviewer. Keep a one-page version with the inputs, the three scenarios and the review date, so that the logic is easy to follow and easy to update.
How VDS fits
VDS builds solar-powered mobile surveillance trailers with a 5MP PTZ camera with 25x optical zoom and three 8MP wide-angle cameras on a 22 ft telescopic mast. A 24/7 US-based security operations center verifies events, uses live talk-down through a loudspeaker and then dispatches police with video. Rental, month-to-month and term subscriptions are available, so you can match the commitment to your model. We do not publish prices here, and we are happy to quote against the numbers you bring. See remote monitoring or talk to our team.
Sources
- Piza, E. L., Welsh, B. C., Farrington, D. P., & Thomas, A. L. (2019). CCTV surveillance for crime prevention: A 40-year systematic review with meta-analysis. Criminology & Public Policy, 18(1). onlinelibrary.wiley.com