The information decides the answer
Every cost figure is derived from something. At feasibility it is derived from area schedules and a description of intent. At later stages it is derived from drawings and specifications. The quality of the figure follows the quality of that input directly.
This sounds obvious and it is routinely ignored, because a surveyor asked for a cost plan will usually produce one from whatever is available rather than declining. The document then carries an authority its inputs do not support.
The honest alternative is a cost plan that states what it was measured from, lists what was assumed in the absence of information, and puts a value against each assumption.
The relationship runs both ways, which is worth saying. A design team asked for information at the right moment, in a form that is useful, will generally supply it. Asked at the wrong moment, in a form that means redoing work, they will supply what they have, and the cost document inherits the gap.
What is needed at each stage
At feasibility: an area schedule with gross internal and net saleable areas, a description of use and storey heights, and a statement of the intended specification level. That is enough for a benchmarked estimate and not enough for anything else.
At the design stages: general arrangement drawings with dimensions, a structural strategy, an outline services strategy, an outline specification and an elevation strategy. That supports an elemental cost plan measured to NRM1.
At tender: coordinated drawings, a full specification, and enough detail for the elements that carry most of the value. That supports measured quantities to NRM2, or a schedule of works where a bill is not being prepared.
The gaps that cost most
Four recur. The services strategy, which is frequently the last to arrive and which can carry a large proportion of the cost on residential and commercial work. The structural solution below ground, which depends on a site investigation that may not have happened. The extent of external works, which is often shown as a boundary line and priced as an allowance. And the specification level, where a schedule saying high quality supports no measurement at all.
Each of those can be handled by an explicit assumption with a value against it. What cannot be handled is a gap nobody names, because it then sits inside the total as though it were measured.
We wrote about the general discipline in measuring an incomplete design.
Revisions and what they do
A cost plan measured from one revision and issued after two more have been produced is describing a building that no longer exists. That happens constantly, because design and cost run on different cycles and nobody synchronises them.
The fix is procedural. Every cost document lists the drawings and specifications it was measured from, by number and revision, and every reissue includes a reconciliation showing what changed and what that change cost.
That reconciliation is also the only reliable way to distinguish design development, which is the design becoming what it was always going to be, from scope creep, which is the design becoming something else. The two look identical in a total and are entirely different commercially.
Models, and what they do not answer
Where a model exists, quantities can be extracted from it and the exercise is faster and more consistent. What a model does not supply is the specification behind the geometry, the site constraints, the sequence, or anything that has not been modelled because it was not yet decided.
A quantity taken from a model is only as reliable as the modelling conventions behind it, and those vary between practices and between packages within one project. Auditing what the model actually contains is part of using it, which we covered in BIM quantity take-off limits.
The model is a method rather than a substitute for information.
What this means for you
Ask what any cost document was measured from before you rely on it, and check the revisions against what the design team has issued since.
Where information is missing, ask for the assumption and the value rather than accepting a figure that implies certainty nobody has. A surveyor who lists what they did not have is telling you something useful.
That discipline runs through cost planning and estimating and quantities and take-off, and it is what makes a figure defensible to a board or a funder rather than merely available.
One practical arrangement helps more than any process document. Agree at the start of each design stage what information will be issued for costing, on what date and at what level of completeness, and hold that as a short list rather than as an assumption. It converts a recurring source of friction into a diary entry. A client side quantity surveyor who lists their information sources on every document is giving you a figure you can defend rather than one you have to trust.