By the time a construction project's design is finalised, many of the decisions that determine its long-term resilience have already been made. Nearly 200 years of property losses have shown insurers where things most commonly go wrong, from water escaping inside pharmaceutical facilities to natural catastrophe exposures tied to a project's location.
Historical data explains where losses have occurred. Understanding tomorrow's risks requires something else. According to Mary Rieck, principal underwriter for renewable energy at FM, that means physical testing, including deliberately setting fires at the company’s research facilities, to understand how emerging risks behave before they show up as claims.
Liquid cooling systems now appearing inside data centre "white space" are one example of a technology the industry is still learning how to underwrite.
"We take our historical data and our more forward-looking research, and we put that together to come up with the risks that we want to worry about for your site, based on what you're going to be doing there and where it is in the world," she said.
When asked which design decisions matter most over a project's lifetime, Rieck did not point to anything dramatic. Equipment placement, she said, is a surprisingly common source of avoidable loss. Installing a spare transformer immediately beside the one it is intended to back up largely defeats the purpose of having redundancy in the first place.
"The same event that takes out your in-service transformer might take out your spare," she said.
If identified during design, the solution can be as simple as relocating the equipment. Once construction is complete, the same change could mean moving walls, bringing in specialist contractors and repeating testing.
"It's really pretty minimal" if identified early, she said, noting that it can become costly if it is not.
Many of the issues she sees are surprisingly unglamorous. Better drainage, correctly sized sprinkler heads and building materials matched to the wind zone a site actually sits in are recurring examples - decisions Rieck described as "pretty simple stuff" that only become expensive once construction has moved beyond the point where they can be changed easily.
Renewable energy, she believes, offers a useful model for the wider construction industry because of how readily the sector has adopted new loss-prevention thinking. She pointed to solar panels being redesigned to better withstand hail, alongside improved methods of stowing equipment, as examples of innovation the industry has absorbed quickly rather than resisted. Battery management systems, she added, are evolving just as rapidly.
That openness also shapes how renewable projects are designed from the outset, with room deliberately built in for whatever comes next. The same principle, she said, applies across construction: leaving space for future expansion or building in fire separation ahead of a second production line, rather than designing only for a site's immediate needs.
The harder challenge is timing. Many of the decisions that shape a building's long-term resilience are made far earlier than clients or insurers often realise. Once those choices are embedded in the design, they become progressively more expensive to change.
That thinking underpins FM's decision to expand further into construction insurance rather than waiting until projects are already under way.
"The majority of all loss is preventable," Rieck said.
"We want to make sure we can get to the table before those design decisions are locked in and really incorporate loss prevention as early as possible."
Rieck believes the single biggest determinant of future property risk is not a new technology or building material but where a project is built. Developers balance numerous commercial considerations when selecting a site, including access to power, customers, labour and tax incentives. Once that decision has been made, however, the natural hazard profile is largely fixed for the life of the asset.
She pointed to large solar farms built in severe hail zones in Texas as an example.
"You're married to that natural hazard profile," she said.
Asked whether that risk differs significantly between the UK and North America, Rieck said the physical hazards themselves are often less important than the way insurance markets respond to them.
"There's a bigger difference in the market insurance norms than there are in the risks," she said.
Policy structures, contract wordings and procurement practices vary considerably between jurisdictions, even when the underlying engineering risks remain much the same. The risks themselves rarely change much from market to market. The insurance built around them still does.
