Why Are Existing Drawings So Inaccurate?

A plan marked “existing” can look authoritative while being decades out of date. When a proposed layout, heritage consent package or coordination model depends on it, that gap becomes expensive. Why are existing drawings inaccurate? Usually not because one person made a simple mistake, but because the drawing no longer represents the building as it stands.

For design teams, the question is less about assigning fault and more about deciding whether the available information is dependable enough for the next decision. A quick feasibility study may tolerate a degree of uncertainty. Detailed design, structural coordination, fabrication and work to a listed building generally cannot.

Why are existing drawings inaccurate in practice?

Existing drawings are records of a moment in time. Buildings change continuously through refurbishment, maintenance, tenant fit-outs, repairs and informal alterations. A set produced for planning approval may have recorded an intended design rather than the completed work. A contractor’s drawing may show what was instructed, not what was installed.

Over time, these documents are copied, redrawn, cropped and issued without their original notes, scale bars or survey control. Dimensions can be rounded, assumptions can become embedded, and uncertainty can disappear from the record altogether. The result is a drawing that appears precise because it uses clean CAD linework, while its relationship to the physical building is unclear.

This is especially common on older properties, multi-phase commercial buildings and buildings with complex roof forms or irregular structure. In those settings, even a small discrepancy at ground level can become a significant coordination issue across several storeys.

The most common causes of unreliable existing-condition information

Drawings were based on design intent, not as-built conditions

Many drawings originate before construction. They may accurately show the approved proposal but not site changes made during delivery. Substituted materials, adjusted partitions, altered service routes and revised structural openings are all routine responses to construction constraints.

Even formal as-built sets need careful interpretation. Their reliability depends on how they were compiled, whether changes were checked on site, and which elements were actually updated. A title block stating “as built” is useful context, not proof of measured accuracy.

The building has changed since the drawings were issued

Alterations are often piecemeal. A doorway is moved during a fit-out. A riser is boxed out. A ceiling is lowered to accommodate services. A roof void is adapted, or a historic wall is repaired with a slightly different build-up. Each change may be minor in isolation, but together they can make an inherited plan unsuitable for current design work.

The risk is not limited to visible changes. Floor levels, soffit heights, beam depths and service zones are frequently absent from older records. These are often the dimensions that determine whether a new intervention can be delivered.

Previous surveys used a different purpose and tolerance

Not all surveys are meant to support the same decisions. A marketing floor plan, estate record or early-area schedule can be entirely appropriate for its original purpose, yet unsuitable for detailed architectural design. Likewise, a survey undertaken to locate boundaries may not provide the internal geometry required for a Revit model.

Accuracy is therefore not a single number. It includes the capture method, coverage, control, level of detail and the way the information is translated into CAD or BIM. A drawing may be broadly correct overall while still failing where exact wall thicknesses, opening positions or ceiling geometry matter.

Complex geometry is simplified

Traditional measured surveys often rely on selected dimensions and manual interpretation. This can work well in straightforward spaces, but it has limits when walls are out of plumb, rooms are non-orthogonal, or surfaces are curved, stepped or irregular.

Historic buildings make this particularly clear. A nominally rectangular room may have bowed walls, uneven floor levels and openings that do not align vertically. Forcing that geometry into neat right angles can make a drawing easier to read, but less useful for design, conservation or fabrication.

Scaling, conversion and coordination introduce errors

A PDF scaled from an uncertain original is not a measured drawing. Neither is a CAD file automatically dependable because it contains dimensions. Rescaling, importing between software platforms, tracing from raster scans and using inconsistent units can all create errors.

Coordination also suffers when different disciplines work from different baselines. If the architect, structural engineer and services consultant each receive a slightly different version of the existing building, conflicts can be designed in before the project has properly started.

The cost of treating approximate information as fact

Inaccurate existing drawings create more than an awkward site query. They affect decisions at every stage.

At feasibility stage, an overstated floor area or misunderstood core arrangement can distort a development option. During design, a misplaced structural wall or underestimated roof slope can force late revisions. On site, clashes between proposed work and concealed conditions can lead to delays, variations and strained client conversations.

For heritage projects, the consequences can be more sensitive still. A model that regularises an uneven historic structure may obscure the relationships that need to be protected. If interventions are designed around assumed geometry rather than captured conditions, the team may need to revisit proposals after intrusive checks or opening-up works.

There is a practical trade-off. Commissioning a detailed measured survey and model has an upfront cost, and not every project needs LOD400 information. But the right level of documentation should be chosen against the cost of getting a key decision wrong. A small interior reconfiguration may need accurate plans, elevations and ceiling heights. A complex retrofit involving structure, services and heritage fabric may justify a registered point cloud, detailed sections and a coordinated Revit model.

How to establish a dependable starting point

The first step is to review inherited information critically. Check drawing dates, stated scales, revision histories and whether the documents identify a survey method or accuracy standard. Compare a small number of known dimensions on site before relying on the files for layout or coordination.

Then define what the project team genuinely needs. This should cover the areas to be captured, the required outputs, the intended software environment and the critical geometry. For example, a team planning a rooftop extension may need roof plans, parapet levels, neighbouring relationships and detailed sections through the existing structure. A retail fit-out may prioritise internal walls, reflected ceiling information, access routes and service constraints.

A precision-first documentation workflow normally begins with 3D laser scanning and measured survey control. The resulting point cloud provides a dense record of the visible site condition, allowing plans, elevations, sections and models to be checked against captured geometry rather than reconstructed from sparse notes.

The point cloud is not, by itself, the final design document. It needs to be translated into structured outputs that suit the project. Clean 2D CAD drawings may be the right answer for a focused package of work. A Revit model can provide a stronger basis for multidisciplinary coordination, provided its level of development is defined clearly and the modelled elements reflect the agreed scope.

At Space Captures, this means agreeing the deliverable before capture begins, rather than treating every building as a standard survey. Irregular geometry, listed fabric and difficult-to-access areas need to be identified early, so that the capture approach and final documentation answer the real design questions.

Questions worth asking before you rely on old drawings

Before issuing inherited information to the wider team, establish whether it records design intent or verified site conditions, when it was last checked, and which parts of the building it omits. Ask whether dimensions were measured, scaled or assumed, and whether floor levels, roof geometry, structural elements and services have been captured at the detail your project requires.

It is also worth agreeing a clear convention for uncertainty. If an area cannot be accessed or a concealed element has not been verified, that should be recorded rather than guessed. Honest limitations allow the team to plan targeted investigations. Hidden assumptions tend to emerge later, when the programme is less forgiving.

Reliable existing-condition information does not eliminate every site surprise. Buildings retain concealed conditions, particularly where fabric is historic or previous work is poorly recorded. It does, however, give the design team a dependable geometric baseline from which to make decisions, coordinate proposals and identify what still needs investigation.

When the existing drawing is central to the project, treat it as evidence that needs validation, not a background file to be accepted on trust. The earlier that validation happens, the more freedom the team has to design with confidence.