Picture an owner across the table from a banker.
Revenue is up. Margin is fine. The banker turns a page and asks:
"Is the company getting safer?"
The owner says, "Our premium went up again."
And then stops. Because that doesn't answer the question. It might even suggest the wrong answer.
Why is premium the wrong scoreboard?
Premium is a fine number. It's just the wrong one to judge yourself by.
When your company grows, premium grows with it. More payroll, more trucks, more revenue, more exposure. A bigger premium is often a sign that business is good.
So "my premium went up" can mean risk got worse. Or it can mean the company got bigger. Premium can't tell you which.
That's the frustrating part. You're spending on training, gear and insurance every year, and you have no number that tells you if any of it is working.
You shouldn't need a risk department to know whether you're winning.
What is cost of risk per $1,000 of revenue?
Total cost of risk divided by revenue, times 1,000.
That's what you spend on risk for every $1,000 you earn.
Because it's measured against revenue, it takes growth out of the picture. Grow with risk under control and the number holds steady or comes down, even while premium goes up.
It also connects straight to net operating income. Every dollar that comes off this number is a dollar of revenue that stays as operating income instead of leaking out as risk cost.
What does it look like with real numbers?
Here's a sample case. It's an illustration, not a client.
A $12M remodeler paid $357,000 in premium. When everything risk cost was added up, the total was $617,950.
$617,950 divided by $12,000,000, times 1,000, equals $51.50.
So that company spent about $51.50 on risk for every $1,000 it earned.
Now picture that owner a year later. Revenue grew. Premium went up too. But total cost of risk grew slower than revenue, so the $51.50 came down.
Now the answer to the banker is one sentence: "We're spending less on risk for every dollar we earn, and here's the number."
In IT, the first thing you do with any system is pick the measure that tells you whether it's healthy. Risk is a system. This is its measure.
How do you figure yours, roughly?
A rough number is better than no number.
Start with what you pay to transfer risk:
- All your insurance premiums for the year, from final audit statements if you have them
- Taxes and fees on those policies
Add what you pay yourself:
- Deductibles you paid on claims
- Losses insurance didn't cover
- Audit charges and surprise bills
Add what you spend to control risk:
- Safety training, gear and programs
- Outside safety or consulting help
- Technology for risk, like cameras, telematics or certificate tracking
- A rough share of the time your people spend on safety, claims and insurance paperwork
Divide by revenue and multiply by 1,000.
What will a first pass miss?
A full total cost of risk also counts things that are hard to see on your own, like the hidden cost of injuries. A small injury (under $3,000) can cost about 4.5 times the medical bill in lost time and slowed-down crews.
So your rough number is a floor. The real number is almost always higher.
A Frost Baseline covers six areas, including the hidden injury costs, and locks the starting number so every later year is measured against the same line.
How does this number connect to profit?
Think of it as a share of every revenue dollar that goes out the door as risk cost. In the sample case, about 5 cents of every dollar went to risk. Whatever you take off that number stays as operating income instead.
Here's an illustration, not a client result. At a 5% net margin, $1,000,000 of new revenue produces about $50,000 of profit. Lower the cost of risk by $50,000 and you get the same profit without the new revenue, and without the new crews, trucks and subs that come with it.
That's why the number belongs next to margin on your monthly review, not in a folder you open once a year at renewal.
It also gives you a better way to judge safety spending. If you spend on training or gear and the number comes down over time, the spending is working. If it doesn't, you know where to look.
Are there other ways to measure it?
Same total, different lens:
- Per $1,000 of revenue, for the owner and the banker
- Per $100 of payroll, for workers' comp conversations
- Per hour worked, for safety conversations
Why track it over three years?
One year is a snapshot. Three years is a story.
Parts of your cost of risk move on different clocks. The hidden cost of an injury hits the month it happens. Premium moves once a year. Your experience mod reflects work you did two and three years ago.
So one good quarter won't show up everywhere at once. Over three years, the trend tells the truth.
Without a scoreboard, every renewal is a mood. With one, you lead differently.
One number.
Three years.
A trend you can defend.
Try this
Do the rough math with last year's numbers. Don't chase perfect. Use what you can find in an hour.
Write the number down, with the date and what you included. Next year, do it the same way. That's your scoreboard. Put it at the top of your one-page risk plan.
Where to start
Ask your broker or risk advisor to find your full number across six areas and lock it as your starting line. From there, and a 90-day plan puts an owner and a date on every task, with signed proof of every saving. Approved Frost agents run this kind of program for contractors, start to finish.
The reference entries behind this page: definitions, the math and the sources.
Cost of risk rates and benchmarksBrowse the whole Library →Quick answers
What is a good cost of risk per $1,000 of revenue for a contractor?
There isn't a reliable trade-by-trade benchmark to compare against. The most useful comparison is your own number over time, measured the same way each year.
Why not just track my premium?
Premium grows when your company grows, so it can't tell you whether risk got better or worse. Measuring total cost of risk against revenue takes growth out of the picture.
How accurate is the rough calculation?
It's a floor. It usually misses hidden costs like lost time and slowed crews after injuries, so the real number is almost always higher.