Repair or Replace? The Numbers Owners Actually Ask For

No client has ever asked which option produces the better structural mechanism. They ask what it costs, how long it takes, whether the building can stay open, and what happens if they do nothing. Those are the right questions, and the engineering case has to be translated into them before any decision gets made.

The comparison is also asymmetric in a way that is rarely presented honestly. Replacement has a known cost, a known programme and a clear liability boundary. Repair has a survey cost, an uncertain extent, and a residual-life estimate that depends on assumptions the client cannot check. That asymmetry systematically favours demolition, even where repair is technically better.

1. Cost per square metre, and what is in the number

A like-for-like comparison has to include the things that never appear in the construction rate:

Cost line	Repair	Replace
Investigation and design	Higher — existing structure must be characterised	Lower — standard design
Temporary works and shoring	Often significant	Minimal
Construction	Moderate	High
Finishes, services, fit-out	Largely retained	Complete replacement
Permits and approvals	Retrofit permissions, possibly heritage consent	New-build consent, demolition permission
Financing during works	Carried by owner	Carried by owner
Demolition and disposal	Localised	Full, plus waste transport

The line that changes conclusions more than any other is finishes and fit-out. On a residential building with basic finishes, replacement can be competitive. On a hospital, a laboratory or a hotel with bespoke interiors and services, the replacement cost is dominated by reinstating what already exists, and repair wins on arithmetic before engineering is discussed.

2. Downtime and business interruption

For commercial assets this is usually the largest number in the model, and the one most often estimated badly. A retrofit that keeps 70% of the building in use for nine months can be worth more than a replacement that empties it for fourteen, even at a higher construction cost.

That advantage is not automatic. It depends on a contractor who can sequence around occupancy, on a structural scheme that avoids interrupting every column, and on a client who will accept phasing. Where those three are absent, the “retrofit keeps you open” argument is a promise rather than a plan.

3. Residual life and the next intervention

A repair buys a period, not a permanent state. A carbonation repair with a good cover replacement and a protective coating might give 25 to 40 years of additional life in a sheltered exposure; a seismic retrofit brings the structure to a defined performance level for a defined design life. Both need a maintenance regime, and both need to be honest about the next intervention.

Replacement resets the clock, at a cost that includes demolition and the loss of everything the building embodied. Which is better depends on how much life the repair actually buys relative to the capital it consumes — a comparison that requires a service-life estimate, not a materials warranty.

4. Embodied carbon and the regulatory direction

A reinforced concrete frame carries roughly 200 to 400 kg CO₂e per square metre of floor area, depending on the structure and the cement content; the finishes and services add substantially more. Retaining that frame avoids the majority of those emissions, and the structural strengthening itself is a small fraction of the total.

That is now a procurement factor rather than an ethical footnote. Whole-life carbon assessments are increasingly required for public projects and for large private developments, and a retention option usually scores well. Expect this to move from a scoring advantage to a condition of approval in the next code cycle.

5. Risk, uncertainty and who carries it

Repair concentrates uncertainty in the investigation phase and in the contractor’s scope. A structural survey that finds 15% of the columns deficient may find 30% once the cladding is off, and every additional column is a variation.

Three contractual devices manage this without pretending the uncertainty is gone:

• A defined investigation phase, separately procured and paid, with an agreed deliverable: extent, mechanism, and a repair strategy with a quantified range.
• Provisional quantities for the likely repair extent, with a re-measurement mechanism and unit rates agreed before work starts.
• A trigger clause — if the as-built condition differs from the survey by more than a stated threshold, the project returns to the engineer for re-design rather than absorbing the difference in the field.

Replacement has its own uncertainty, mostly in ground conditions and in the time to obtain approvals. It is not a risk-free option; it is simply a better-understood one. The honest comparison puts both on the same footing rather than treating construction risk as inherent to repair and absent from new build.

When replacement genuinely wins

There are situations where the arithmetic is not close, and arguing for repair is a disservice:

• The seismic demand exceeds what the existing geometry can be strengthened to resist without effectively building a new structure inside the old envelope.
• The use is changing — residential to office, warehouse to data centre — to the point that floor loads, services and floor-to-floor heights all need to change.
• The foundation is the governing deficiency in a way that requires underpinning every column.
• The building has no residual value beyond the site, and the site is the asset.
• The cost of phasing around occupancy exceeds the construction savings — common in buildings with a single critical process.

Conversely, repair tends to win where the frame is sound and the deficiency is local, where the finishes and services are valuable, where occupancy matters, where heritage or planning constrains demolition, and where carbon or approval requirements penalise new construction.

A decision table to take into the meeting

Question	Repair favoured	Replacement favoured
Is the frame’s geometry adequate for the target performance?	Yes	No
Can the work be phased around occupancy?	Yes, with a plan	No realistic phasing
Value of existing finishes, services, fit-out	High	Low
Carbon or planning constraints	Present	Absent
Foundation adequacy	Adequate or locally improvable	Globally deficient
Client’s tolerance for scope uncertainty	Managed by contract	Low

The output of that table is not a verdict; it is an agenda. The engineering then answers the rows that are currently answered by assumption.

What to tell an owner in one paragraph

If the frame can be brought to the required performance level with local interventions, and the building needs to keep working while it happens, repair is almost always the better use of capital — provided the investigation is honest, the scope is contracted with a mechanism for change, and the residual life is stated in years rather than implied. If the geometry cannot deliver the performance, or the foundation is the problem, no amount of composite will change the answer, and the sooner that is said the cheaper the project becomes.