A small fraction of mining sites worldwide account for a disproportionate share of the industry's biggest losses, and having a more accurate understanding of which ones are most likely through predictive analytics rather than after-the-fact review, according to Michael Beaumont (pictured), account engineering group manager and senior advisor at FM.
FM's loss data shows that roughly 2% of locations are responsible for close to 30% of total losses, a pattern Beaumont said has held consistently across the years the company has been running the analysis. Beaumont noted that roughly 60% of the underlying loss data involves Canadian-owned or Canadian-located mining assets, reflecting how much of FM's mining business runs through Canada rather than pointing to anything distinct about Canadian risk itself.
"It's been lining up that about 2% of locations create 30% of loss," Beaumont said. "That's more an observation of the way loss is being distributed in our book as opposed to a reason why."
The finding comes from combining FM's global loss history with the detailed engineering data its own teams collect on site, feeding both into a statistical model built to flag which locations carry the highest predisposition to loss.
"We take our loss data and all the data that we collect from the site," Beaumont said. "Our engineers go out to these mining sites. At a minimum, they'll collect 700 pieces of discrete data to do their reporting and their risk analysis."
According to FM's report, that approach has been applied across more than 60,000 visited locations globally, allowing the company to identify the 1,000 sites carrying the highest loss exposure, including those most likely to produce losses greater than $3 million.
FM's data shows a 27% overlap between the locations flagged as highest-risk and where its largest actual losses have occurred, a consistency Beaumont said gives the model real predictive weight rather than functioning as a purely retrospective exercise. He added that the underlying incident data is likely conservative in some categories, since minor events resolved on-site, such as a small fire put out with an extinguisher, don't always get formally recorded.
What separates a high-risk location from a lower-risk one isn't necessarily poor risk management, Beaumont said. It often comes down to the underlying complexity and throughput of the operation itself.
"Do they have more higher-risk issues? Yes. They tend to be locations that may be in need of improvement," Beaumont said. "But the other really important component is that these are very complex locations, very high business-interruption numbers. We have some very well risk-managed locations in there. They're just more predisposed to loss because there's more throughput, the complexity of the processes is higher."
For companies that aren't FM clients and don't have access to the underlying model, Beaumont said the loss report itself is designed to serve as a practical substitute, a way to stress-test a business against real-world scenarios rather than assuming a given failure couldn't happen on-site.
"We're not opening up all our IP, but we're opening up some of it and saying, this is where the losses are happening," Beaumont said. "What someone can do is take a look at that loss history and go, what if this happened at my site? What would that do to my business?"
That exercise matters most, he said, when it forces operators to confront costs insurance was never designed to cover in the first place, particularly the value of lost time in a high-price commodity environment.
"There's nothing we can do to give you back time in 2025 if you're a gold miner," Beaumont said. "I can't give you back time to make gold whilst it's $5,000 an ounce. You've lost that opportunity. That's uninsurable."
Beaumont said the shift he's encouraging clients toward isn't really about adopting a new tool, but about changing the underlying posture from reactive to anticipatory, a mindset he traces back to a much older, low-tech version of the same idea from earlier in his career.
"When I started in this industry 36 years ago, we had a book called 'What Has Happened Can Happen,'" Beaumont said. "It's this little binder of stories you could tell the clients so they could start overlaying that on their business. That's really what we're talking about here: take the loss history, overlay it on your business, and ask, what if that happened here?"
The alternative, he said, is a riskier bet than most operators realize.
"You could be an ostrich and stick your head in the sand and say, 'Oh no, it's not going to happen here'," Beaumont said. "But that's probably not a great way to manage risk, at a mine site or any site."
