Refurbish or replace: set a fair comparison before choosing

The two prerequisites from any option comparison apply here unchanged, and both are commonly got wrong in a refurbish-or-replace decision.

Functional equivalence means the refurbished building must deliver what the client needs, over the same period, to the same standard as the replacement. If the refurbishment cannot achieve the required floor area, ceiling heights, servicing strategy, accessibility or performance, then the options are not equivalent and the comparison must either price the shortfall or acknowledge that the refurbishment does not meet the brief.

The study period must be long enough to capture the consequences of the choice. A refurbishment with a significantly shorter remaining life will require further intervention within a period long enough to capture it, and a comparison that ends before that intervention flatters the refurbishment. Conversely, a replacement carries the impact of demolition and of new construction at the start, which a short study period will not allow a longer-lived option to amortise.

The options matrix below sets out the lines a defensible comparison carries, and the evidence needed to populate it. Much of the difficulty in this comparison comes from the fact that the information required is not all available at the point the decision is taken, which is why the exercise is usually iterative and why the evidence gaps are themselves a finding.

The evidence gaps above share a feature: most of them can only be closed by investigating the existing building, and the investigation itself has a cost that must be committed before the decision is made.

This is the practical reason why refurbish-or-replace decisions are frequently made on incomplete information. The survey that would establish the condition of the structure, the extent of contamination or the state of the services is a real cost, and committing it before knowing the answer is uncomfortable. The alternative — deciding on assumption — usually costs more, because an optimistic assumption about the existing building is discovered during construction, when the design is committed and the programme is under pressure.

A staged approach manages this. The first stage is a review of existing information: drawings, condition reports, maintenance records, any previous surveys. It is inexpensive and often eliminates options quickly. The second stage is a targeted investigation of the specific items the comparison is most sensitive to — usually the structure and the services, because they are the largest determinants of refurbishment cost and feasibility. The third stage is the full survey, commissioned once the comparison has narrowed to options that justify it.

The output of this process is a comparison with its uncertainty stated, which is more useful than one that appears precise because its assumptions were never tested. Where an evidence gap cannot be closed within the available budget, the honest course is to state it and to identify what the decision's sensitivity to it is, so that the risk is visible to whoever takes the decision.

Where the comparison carries an environmental dimension, the boundary question returns. Retained material and new material must be assessed on the same basis, over the same period, or the two options are not being compared at all. The Commission's recommended Environmental Footprint methods [B] are relevant here because they set out a consistent way to quantify environmental performance across a life cycle, which is what makes a retained-versus-new comparison legible rather than rhetorical.

Rebound deserves a specific mention because it is systematically omitted and it can reverse a conclusion.

When a building's performance improves, the improvement is frequently taken partly as service rather than wholly as saving: a warmer or cooler space than before, more extensive lighting, additional equipment. The resource consumption therefore falls by less than the improvement predicts. The effect is well documented in other contexts and applies here.

In a refurbish-or-replace comparison, the significance is that the operating performance advantage assumed for one option may not be realised in full. Where the comparison is sensitive to that advantage, the sensitivity should be tested — and the assumption about how much of the improvement becomes saving should be stated explicitly rather than left at one hundred per cent by default.