Why your NatHERS or HERS rating never touches your thermal bridges
I was on site in Elsternwick a few months back, running a thermal camera over a near-finished extension, when the owner asked why there was a cold stripe running down the wall exactly where the steel portal frame sat behind the plasterboard. The house had a reasonable NatHERS rating. The certificate said nothing about that stripe. It couldn't have, because the software never saw the steel.
That's the gap I want to walk through today: thermal bridging, and why it sits almost entirely outside what a NatHERS or HERS rating actually models. Not a flaw exactly. More a boundary of the tool that almost nobody explains before the build starts.
What thermal bridging actually is
A thermal bridge is any point in the building envelope where a material with much higher conductivity punches through the insulation layer and gives heat a shortcut. Steel studs through a batt-insulated wall. A concrete balcony slab that cantilevers straight out from an internal floor into the open air. Window and door reveals where the frame meets masonry. Roof battens, structural steel portals, even the humble nail or screw fixing through a layer of rigid foam, multiplied by the hundreds across a roof.
None of these are rare or exotic. They're in nearly every house built in the last twenty years, especially anything with a steel frame, a cantilevered balcony, or a lot of window area. The physics isn't controversial: conductive heat flow runs happily through metal and concrete at many times the rate it moves through timber, let alone insulation batts, so a relatively small area of bridging can account for a disproportionate share of total heat loss.
Why the rating software can't see it
NatHERS (via its accredited software, mainly AccuRate or BERS Pro) and the US HERS Index (via RESNET-accredited software like REM/Rate or Ekotrope) both build their energy model from a set of standardised assemblies. You tell the software "brick veneer wall, R2.5 batts, 90mm timber stud" and it pulls a pre-calculated total system R-value for that assembly type from a library. That library value is usually a clear-wall calculation, meaning it assumes an unbroken, uniform layer of insulation across the whole wall area, with maybe a notional allowance for standard timber framing factor baked into the assumption.
What it doesn't do, in the vast majority of standard residential assessments, is ask you to draw every individual steel stud, every slab penetration, every window reveal detail and calculate a bridge-specific heat flow path through each one. That's a much more detailed exercise, closer to what you'd see in a Passive House PHPP model or a bespoke thermal bridging calculation done by a building physicist for a commercial project. For a project home or a standard renovation going through a NatHERS or HERS assessment, it simply isn't part of the brief.
So the assessor enters "steel frame" as a construction type, the software applies a generic correction factor for steel (it does acknowledge steel conducts more than timber, to be fair), and that's roughly where the detail stops. The actual geometry of your specific portal frame, your specific balcony cantilever, your specific window reveal, doesn't get modelled as a discrete bridge with its own calculated heat loss.
The clear-wall assumption and why it matters
This is the technical crux, and it's worth being precise about it because the term gets thrown around loosely. A "clear-wall R-value" describes the thermal performance of a representative, uninterrupted section of wall assembly, away from corners, openings, and structural penetrations. It's a genuinely useful number for comparing insulation products and assembly types on a like-for-like basis. The problem is when that clear-wall number gets treated as if it represents the whole wall, corners, penetrations and all.
Building science research, including work published through CSIRO and the kind of thermal bridging guidance found in the Australian Institute of Refrigeration, Air Conditioning and Heating's technical resources, has shown that whole-wall R-values, the real-world average once you account for every stud, noggin, lintel and bridge, can run noticeably below the clear-wall figure quoted on a product datasheet. Steel-framed walls suffer this more than timber, because steel is such a better conductor than wood. A wall that's notionally R2.5 in the clear cavity might behave, averaged across its whole area including the steel, meaningfully worse.
None of that discrepancy shows up on your NatHERS certificate. The star rating is calculated from the clear-wall assumption, assembly by assembly, and the certificate doesn't carry a line item for "thermal bridging penalty." If you want to know how a building rating treats hot water appliances or ductwork, both similarly absent, I've covered that ground in why your NatHERS or HERS rating never asks about your hot water and why your NatHERS or HERS rating never mentions your ductwork. Bridging sits in the same category: real, measurable, and outside the model's scope.
Where bridges actually hide on a typical build
Balcony and deck slabs that cantilever from an internal floor are probably the worst offenders I see, because the thermal path is so direct. Concrete is a poor insulator and an excellent conductor, and a cantilevered slab physically connects your conditioned internal floor to the outside air with no break unless the builder has specifically detailed a structural thermal break into the slab, which does exist as a product category but gets value-engineered out more often than it should.
Steel portal frames and structural steel beams that run through the external wall line are next. They're increasingly common in open-plan renovations and anything with a big glazed opening needing lintel support, and they create a continuous metal path from inside to outside unless someone's specifically wrapped or broken that steel.
Window and door reveals matter too, particularly the junction between the frame and the surrounding wall. I wrote about the install-quality side of this in why your NatHERS or HERS rating never checks your windows' install, and bridging is the thermal cousin of that same gap: the rating credits the window's whole-of-unit U-value, but the reveal detail around it, where the frame meets masonry or cladding, is rarely modelled as its own bridge.
Roof battens and the sheer number of fixings through continuous insulation are a smaller individual effect but they add up across a whole roof plane. And balconies aside, exposed concrete slab edges on any suspended floor are a classic bridge that a lot of auditors flag when they bring thermal cameras onto a finished site, which is a service I've covered from the vetting side in vetting a blower door and thermal imaging auditor.
Does this mean the star rating is wrong?
No, and I want to be careful here because it's easy to overstate this. The rating isn't wrong, it's doing exactly what it's designed to do: compare the thermal performance of a design against a standardised climate file and a standardised set of assembly assumptions, consistently, across every house that gets assessed. That consistency is the whole point. If every assessment suddenly had to model bespoke bridging geometry, the system would become slower, more expensive, and arguably less comparable between projects, because the modelling choices would vary assessor to assessor.
What it does mean is that two houses with the same star rating can perform differently in practice if one has a lot of exposed structural steel and uninsulated balcony slabs and the other doesn't. I'd argue this is the single biggest reason project-home display villages and architect-designed steel-and-glass houses with identical NatHERS ratings can feel completely different to live in, something I touched on more generally in why a NatHERS or HERS rating can't tell you two houses feel different. Bridging is one of the concrete, physical reasons behind that comfort gap, not just a vague "it's complicated" answer.
What you can actually do about it
If you're designing or renovating, ask your architect or draftsperson directly whether structural thermal breaks are specified anywhere steel or concrete crosses the insulation line. These are proprietary products, essentially low-conductivity structural connectors, and they're standard practice in Passive House and high-performance builds but still optional extras on a standard project home. They cost money and get cut in value engineering more often than anyone would like.
For balconies specifically, ask whether the slab is thermally broken or whether it's a continuous pour straight through the external wall line. It's a yes/no question a structural engineer or builder should be able to answer on the spot.
If you're buying an existing house rather than building, a thermal imaging walkthrough on a cold morning, ideally with a decent temperature differential between inside and out, will show up bridging as cold stripes and patches on an interior wall surface almost immediately. It won't change the NatHERS certificate, which is fixed at design stage anyway (more on that distinction in why your NatHERS or HERS rating is a design tool, not a build receipt), but it tells you something the certificate structurally can't.
Honestly, I think bridging deserves more airtime in pre-purchase conversations than orientation does these days, and that's a slightly contrarian thing to say given how much attention orientation gets. Most display homes are oriented reasonably well because project builders iterate the same few floor plans across thousands of lots. Bridging is far more variable, far less visible, and far less discussed.
Reading the envelope, not just the certificate
The US side runs on similar logic. RESNET's HERS Index methodology, documented through the RESNET National Standard, uses assembly-level U-factors in much the same way NatHERS does, and RESNET's own technical guidance acknowledges that framing factors and thermal bridging can meaningfully affect whole-wall performance relative to the clear-cavity number used in standard modelling. It's not an Australian quirk. It's how assembly-based energy modelling works everywhere, because full bridge-by-bridge calculation is genuinely a different, heavier modelling exercise, the kind you'd commission separately if you were chasing a Passive House certification rather than a standard compliance rating.
None of this is a reason to distrust your rating. It's a reason to treat the certificate as one input, not the whole picture, and to ask a few pointed questions about steel, concrete and window reveals before the walls go up. By the time plasterboard's on, as that Elsternwick owner found out, the cold stripe is a lot easier to photograph than it is to fix.
— Ben Sorensen, Green Building Writer, IceWarm
Common questions
- Does NatHERS or HERS modelling ever account for thermal bridging at all?
- Partially. Both systems apply generic correction factors for framing type (steel versus timber, for example), but they don't model the specific geometry of your balcony slab, portal frame or window reveal as an individual calculated bridge. That level of detail sits outside standard assembly-based modelling.
- Is a steel-framed house always worse for thermal bridging than a timber-framed one?
- Steel conducts heat far more readily than timber, so steel-framed walls generally have a bigger gap between their clear-cavity R-value and their real whole-wall performance. It doesn't mean steel framing is a bad choice, just that bridging deserves more attention when steel's involved, particularly at junctions like portal frames and lintels.
- Can I ask for a more detailed thermal bridging assessment than the standard NatHERS or HERS rating gives me?
- Yes. It's not part of a standard compliance rating, but building physicists and some high-performance building consultants can run bridge-specific calculations, which is standard practice for Passive House certification. It's an extra service, not something your NatHERS assessor does by default.
- Will fixing thermal bridges change my star rating?
- Generally no, because the rating software doesn't score bridging in the first place, so eliminating a bridge won't move the number it's already not counting. It will change how the house actually performs and feels, which is a separate (and arguably more important) thing from the certificate.
Ben covers insulation, glazing and home energy ratings — the building-envelope work that decides how hard every other system has to work.
Building-certification background (NatHERS-adjacent).
More from Ben Sorensen
- Why a NatHERS or HERS rating can change without touching the buildA NatHERS or HERS rating can shift star bands when software updates, climate files get revised, or rules change — even if your house hasn't.
- Why your NatHERS or HERS rating ignores your floor coveringsCarpet, tile and polished concrete all score the same in a NatHERS or HERS model, even though they feel wildly different underfoot and change how your heating actually performs.
- Vetting an insulation installer: the R-value receipt nobody asks forInsulation installers rarely need a specific trade licence, which is exactly why the paperwork trail matters more than the handshake.
- Why your NatHERS or HERS rating never checks your orientation on siteYour NatHERS or HERS star rating locks in orientation at design stage, but nobody re-checks it once the slab is poured the wrong way round.
- Vetting a window glazier for new-build glazing: what to checkBefore a glazier touches your new-build windows, check the licence, the AS 2047 compliance paperwork and the install method. Here's what actually matters.
- Why a NatHERS or HERS rating can't tell you two houses feel differentA NatHERS or HERS rating scores energy load, not comfort. Two homes with identical stars can feel completely different to live in, and here's why.