The information is the opposite way round
New build cost planning works because the building is entirely defined by documents. Every element is drawn or specified, and where it is not, the gap is visible.
Refurbishment inverts that. The building is standing, so everybody assumes it is known, but what is actually known is the surface. What sits behind the plaster, above the ceiling and under the floor is unrecorded, and it is where the cost is.
This is why a rate per square metre, which is a reasonable early tool on new build, produces figures on refurbishment that bear no relationship to what the work costs. The rate describes an average of buildings whose hidden conditions differed from each other more than their floor areas did.
The gap between the two shows up in the spread of tender returns. On a well documented new build, competitive tenders tend to cluster. On refurbishment they scatter, and the scatter is not a sign that some contractors are cheap. It is a sign that they have each made different assumptions about the same unknowns, and the lowest bid is usually the one that assumed most optimistically.
What actually drives the cost
Five things, none of which correlate with floor area. The condition of the structure and whether it takes the new loading. The state of existing services and whether anything can be retained. The presence of deleterious materials and the regime for dealing with them. Access and sequencing, because working inside a standing building is slower than working on an open site. And whether the building is occupied during the works.
The last of those is the one most often underestimated. A phased refurbishment around occupiers changes working hours, deliveries, noise windows, protection, welfare and the number of times a trade has to return to a room. None of it appears in a floor area.
Surveys are the cost plan
On refurbishment the survey work is not a precursor to the cost plan, it is the substance of it. A structural appraisal, an asbestos survey, a services condition survey and a measured survey of what is actually there convert unknowns into either costs or defined risks.
The temptation is to defer that spend until the scheme is committed, because surveys cost money on a project that may not proceed. That is understandable and it is the wrong order, because the decision the surveys inform is the decision to proceed.
Where a survey cannot be done before purchase, because access is not available, the correct response is to price the risk explicitly rather than to assume it away. [VERIFY: any statutory survey obligations applicable to the specific building type and its intended use.]
How the contract should handle what is left
Some unknowns survive every survey, and pretending otherwise produces a contract that allocates risk to whoever is least able to see it. The choices are provisional sums, which keep the risk with the employer and are valued when the work is known, or a fixed price with the risk transferred, which the contractor will price for.
Neither is wrong. What is wrong is transferring a risk nobody has quantified and assuming it has been priced sensibly, because the contractor prices what they can see too. We covered the mechanics in defined and undefined provisional sums.
Contingency behaves differently here
A percentage contingency on new build is a blunt but tolerable instrument. On refurbishment it is close to meaningless, because the distribution of outcomes is not smooth. Either the structure is adequate or it is not, and the two answers are far apart.
The better approach is a schedule of identified risks, each with a value and a trigger, so the allowance can be released deliberately as surveys and opening up close each one out. That is more work to set up and considerably easier to defend to a funder.
What this means for you
Treat the survey budget as part of the cost plan rather than as a cost before the cost plan. It is the only spend that reduces the range of outcomes rather than describing it.
If you are holding a refurbishment figure derived from a rate per square metre, it is worth testing against measured scope before it hardens into a commitment. That is exactly the work described on our cost planning and estimating and quantities and take-off pages.
The commercial habit that helps most is refusing to give a single figure where the honest answer is a range. A cost plan showing a base position, an identified risk schedule and a stated worst case tells a board something they can act on. A single number for a building nobody has opened up tells them something they will act on and should not. Testing a refurbishment figure against measured scope, rather than against a rate, is the core of what a client side quantity surveyor does before a scheme is committed.