How the two differ in kind
There is a point at which adding insulation to a wall stops making much difference, and understanding why changes what you do next.
Heat crosses a wall by several PARALLEL paths: through the insulated field between the studs, through the studs themselves, through the plates at top and bottom, through lintels and jambs at every opening, through the junctions at floors and roofs, and through anything structural that passes from inside to outside. Parallel paths do not average — the total is dominated by the best conductor available, in the same way that current takes the path of least resistance.
The consequence is that an assembly's EFFECTIVE performance is below the figure printed on the insulation, and the gap WIDENS as the insulation improves. A wall with modest insulation loses a modest proportion of its heat through the framing. Double the insulation between the studs and the framing has not changed at all, so it now carries a much larger share of a smaller total — and the third doubling buys less again. The declared R-value of the material rises in a straight line; the assembly's performance flattens out.
So the question after a certain point is not how much more to add but WHERE the remaining heat is going. The lever that works then is continuity: a layer of insulation running unbroken across the outside of the structure interrupts every parallel path simultaneously, including the studs, the plates and, detailed properly, the junctions. That is why a modest thickness of continuous insulation frequently outperforms a much larger thickness added to a cavity, and why performance standards increasingly specify continuous insulation rather than a cavity figure.
The junctions are the other half of it, and they are where the biggest single bridges usually live: a balcony slab cantilevering through a wall, a steel column passing from inside to out, a party wall that opens into a cold cavity, or a slab edge with nothing between it and the weather.
The factors that actually differ
Show
| More insulation in the field | Interrupting the bridges | |
|---|---|---|
| What limits it | Diminishing returns. Each added layer improves one parallel path, which carries a shrinking share of the loss. | Nothing comparable. Interrupting a bridge acts on the path that was dominating. |
| Effect on the effective figure | Raises the field's performance and widens the gap between nominal and effective. | Closes the gap, which is what the effective figure measures. |
| What it does not touch | Studs, plates, lintels, jambs, slab edges, balconies, structural penetrations, junctions. | Nothing — a continuous layer crosses all of them at once. |
| Cost per unit of improvement | Rises steeply once the field is already good. | Often far better at that point, because it acts where the heat actually is. |
| Where it is easy | A cavity being filled, or a loft where depth is free. | New construction and full re-clads, where a continuous layer can be added outside the structure. |
| Where it is hard | Nowhere much; this is the easy option and that is why it is the default. | Retrofit without re-cladding, and any junction already built — a balcony slab cannot be interrupted afterwards. |
| Condensation risk | A cold bridge in a well-insulated wall becomes relatively colder, so the surface condensation and mould risk at that point RISES. | Reduced, since the bridge is no longer a cold spot. |
| Consequence of ignoring it | An assembly that performs well below its declared figure, and mould in the corners. | Not applicable — this is the fix. |
| How it is verified | By calculation from the material's declared value, which is the optimistic number. | By an area-weighted or effective calculation across all paths, and by thermal imaging on a real building. |
| Which to do first | Up to a sensible level, and in a poorly insulated building it is unambiguously first. | Once the field is reasonable — and always at the junctions being designed now, since they cannot be fixed later. |
Which one, and when
Choose more insulation in the field when…
- The building is poorly insulated, where the field is genuinely the dominant loss.
- There is free depth — a loft, an accessible floor void, an empty cavity.
- The cost per unit of improvement is still favourable, which it is at low levels.
- Nothing structural can be changed, so the bridges are fixed and the field is the only lever.
Choose interrupting the bridges when…
- The field is already well insulated and further layers are showing little return.
- The building is being re-clad or newly built, where a continuous layer can go outside the structure.
- There is a large discrete bridge — a balcony slab, a steel frame, a slab edge, an uninsulated lintel.
- There is mould or condensation in corners and at junctions while the walls themselves are dry.
Now run your own numbers
This page holds no prices on purpose — a national average is wrong for almost every real project. Quantify both options with your dimensions and your local quotes.
Frequently asked questions
- Why does the gap between nominal and effective R-value grow?
- Because the bridge does not improve when the insulation does. Heat crosses the wall by parallel paths, and the total loss is the sum of what each path carries. Improving the insulated field reduces what that path carries, but the studs, plates and junctions carry exactly what they carried before — so their share of a now-smaller total is larger, and the assembly's effective performance falls further behind the material's declared value. Numerically, the first increment of insulation delivers most of its promise and each subsequent one delivers less, until the assembly is essentially as good as its bridges allow. This is also why two walls with the same nominal insulation can perform very differently: one framed at wide centres with insulated headers and one framed tightly with solid lintels and multiple studs at every opening are not the same wall.
- What counts as a thermal bridge?
- Any path that crosses the insulation layer with material that conducts better than the insulation does. The repeating ones are structural: studs, joists, rafters, top and bottom plates, and wall ties in a cavity — individually small, collectively a significant fraction of the wall area. The discrete ones are usually larger per item: lintels, jambs and sills at openings; a slab edge exposed at a floor line; a balcony cantilevering straight through the envelope; a steel column or beam passing from inside to out; a party wall opening into a ventilated cavity. And the geometric ones are the corners and junctions themselves, where the inside surface area is smaller than the outside and heat converges — which is why corners are the first place condensation and mould appear even when no material is bridging.
- Does fixing a bridge help condensation as well as heat loss?
- Yes, and in a well-insulated building the condensation argument is frequently the stronger of the two. Surface condensation and mould depend on the temperature of the SURFACE rather than on the room's average, and a bridge is a cold spot on an otherwise warm wall. Improving the insulation around it without treating it makes that spot relatively colder in comparison — the warm parts got warmer and it did not — so a building can be insulated to a high standard and develop mould in exactly the corners and junctions it did not have before. That is why thermal bridging calculations in performance standards are concerned with a minimum surface temperature factor as well as with heat loss, and why a thermal camera on a cold morning is such an effective diagnostic.
- How do I know when more insulation has stopped paying?
- Calculate the assembly rather than the material, which is what an effective or area-weighted calculation does: take each path's area and its own resistance, combine them properly, and see what the wall actually achieves. Do it twice — once as built and once with an added layer — and the difference is the return on that layer. What the exercise usually reveals is that the curve has flattened long before intuition suggests, and that a modest continuous layer moves the number more than a large cavity addition. The other diagnostic is empirical: thermal imaging on a cold morning shows immediately whether the heat is leaving through the field or through the framing, and a wall whose studs are clearly visible on the image is telling you where the next pound should go.
- Why is continuous insulation so effective?
- Because it interrupts every parallel path at once rather than improving one of them. A layer running unbroken across the outside of the structure sits between the inside and the outside of every stud, every plate, every lintel and every wall tie, so each of those now has to conduct through the insulation as well. That is a qualitatively different intervention from adding depth to a cavity, which leaves all of those paths exactly as they were. It has two further benefits. It moves the structure to the warm side of the insulation, which raises its temperature and reduces condensation risk within the assembly. And it is much less sensitive to workmanship than cavity insulation, which performs badly when it is compressed, gapped or left with air paths around it.
- Can bridges be fixed in an existing building?
- The repeating ones, largely yes, by adding continuous insulation during a re-clad or an internal lining — which is why a deep retrofit is the moment to address them and why doing insulation work without considering them is a missed opportunity that does not return. The discrete structural ones are much harder: a balcony slab cast through a wall cannot be interrupted afterwards, and a steel column crossing the envelope can only be insulated along its length, which reduces the loss without removing the bridge. What can usually be improved is the detailing around them — insulating the reveal and the underside of a lintel, wrapping a slab edge, insulating a party wall cavity at the eaves. The honest position is that some bridges are permanent once built, which is the argument for getting them right at design stage.
- What about the junctions between elements?
- They are where the largest avoidable losses usually are, and they are systematically under-considered because each element is designed and checked on its own. A wall meets a roof, a wall meets a floor, a wall meets a window, a wall turns a corner — and the insulation has to be continuous through each of those, which requires somebody to have drawn it. Performance standards deal with this through junction-specific linear values, either from a published set of accredited details or calculated for the specific detail, and the totals they contribute to a building's heat loss are substantial. The practical rule when reviewing a drawing is to trace the insulation line with a pencil without lifting it: wherever it has to lift, there is a junction that needs designing rather than assuming.
- Is there a point where insulation is actively counterproductive?
- Not thermally — more insulation always loses less heat — but there are real trade-offs beyond the thermal one, and pretending otherwise is how insulation gets a bad name. The three that matter: cost, where the return on each added layer falls until it no longer justifies itself; space, since internal insulation takes room from a floor plan and external insulation changes reveals, eaves and boundary lines; and moisture, because a thicker insulation layer makes everything outboard of it colder, which changes the condensation behaviour of the assembly and, in the case of internal insulation on a solid wall, can put the structure at risk of interstitial condensation and frost. Those constraints, not diminishing thermal returns alone, are what make the bridging question the more productive one past a certain point.