Digital Construction

The work behind the drawing: why digital estimates need manufacturing information

An accurately measured component is not necessarily an accurately priced one. Digital estimating needs to recognise that what has to happen to a material is just as important as how much material is to be used, argues Robert Streather

Shape London's joinery facility. Image: Shape London
Image: Shape London

An accurate digital take-off can still produce a misleading estimate. Software can count and measure joinery components precisely. But an accurately measured component is not necessarily an accurately priced one. The missing information is often not how much material exists, but what has to happen to it.

Harry Hussey recently argued in CM that traditional measurement rules capture materials more readily than the work needed to build them. I see this from both sides of the drawing. I trained as an architect and now run a specialist joinery manufacturing business.

Take two lacquered cabinet fronts of similar dimensions, one flat and one curved. My brother Richard, our production director, and I reviewed the likely production operations. We estimated around 1.7 hours of combined CAD/CAM and workshop work for the flat front, against 10.5 hours for the curved: more than six times as much. Most of the difference lies in forming, programming and machining, not material quantities. These are method-specific estimates, not measured production records or a universal multiplier.

We saw the same issue in a recent tender. The cost plan carried a rate for straight staircase mouldings, but the design included around 32 curved transitions over three floors. To price them properly, we needed to establish the tightest radii, direction of curvature and whether any profiles repeated. One repeatable geometry might use a reusable CNC programme; 32 different geometries could demand far more drawing and programming time. Until we resolved the details, our allowance had to remain provisional.

Measurement rules go some way. Where the relevant measurement rules apply, RICS’s NRM2 requires radii and descriptions of whether curves are concave or convex. But a radius alone cannot tell a workshop whether 32 curved components share one geometry or need 32 separate programmes.

Auditing past work to avoid future blind spots

We have started auditing drawings and project records from our own past work, mapping recurring joinery elements to the manufacturing assumptions behind their estimates. Where production records allow, we can test those assumptions against actual fabrication hours. An estimating tool trained only on historic quotations risks inheriting their blind spots.

That distinction becomes more important as AI enters estimating. RICS’ own guidance on AI in cost estimation notes that take-off tools can misread incomplete drawings and lack understanding of alternative construction methods. Training systems on labelled examples and verified workshop knowledge can help them recognise likely methods and flag uncertainties. It does not make an unconfirmed construction detail a fact.

Architects need not produce fabrication drawings at tender. But they should make build-ups, radii, repetition, junctions, interfaces and unresolved choices explicit, and involve specialists early enough to test the implications. A useful check is to ask: could two competent manufacturers price the same requirements without inventing different assumptions? Otherwise, contractors risk comparing interpretations of the drawing, not like-for-like prices.

The next advance in digital estimating is not simply measuring drawings faster. It is teaching our tools to recognise the work behind them.

Robert Streather trained as an architect and is managing director of Shape London, which designs and manufactures bespoke joinery. AI tools were used to support research, fact-checking and editing for clarity. The argument, evidence and examples are the author’s own.

Story for CM Digital? Get in touch via email: [email protected]

Leave a comment

Your email address will not be published. Required fields are marked *

Latest articles in Digital Construction