CAD Versus BIM Documentation: Which Fits?

A refurbishment project can lose days before the design work has properly begun when a ceiling height is assumed, a historic wall is shown as straight, or an old plan bears little resemblance to site conditions. The choice between CAD versus BIM documentation affects how reliably a team can assess the existing building, develop proposals and coordinate decisions from that point onward.

Neither format is automatically better. CAD drawings and BIM models answer different project questions, require different levels of information, and carry different production requirements. The right choice depends on the building, the proposed work, the consultant team and how the information will be used after survey.

CAD versus BIM documentation: the practical distinction

CAD documentation presents measured information as two-dimensional drawings. Typical deliverables include floor plans, roof plans, elevations, sections and reflected ceiling plans, produced in DWG and PDF formats. It is a clear, efficient way to communicate existing geometry where the team needs dependable reference drawings rather than a data-rich digital model.

BIM documentation represents the building as a structured three-dimensional model, most commonly in Revit. Walls, floors, roofs, doors, windows and other elements are modelled as objects with location, dimensions and, where required, defined properties. Views, schedules and drawings can then be generated from the same coordinated model.

The difference is not simply 2D versus 3D. A well-produced CAD survey drawing is a precise record of what has been measured and agreed within the scope. A BIM model is a managed interpretation of that information, built to support particular design, coordination or asset-information tasks. Both need reliable site capture. Neither can compensate for uncertain survey control or incomplete scope definition.

When CAD documentation is the better fit

CAD is often the most proportionate choice for early feasibility work, planning submissions, measured building records and straightforward alterations. An architect may need accurate plans, elevations and key sections to test layouts, establish areas or prepare a planning package. In those cases, a clean set of existing-condition drawings can provide exactly the required foundation without paying for modelled objects that will not be used.

It is also effective where a project team has an established 2D workflow. Some heritage, conservation and planning processes rely heavily on carefully annotated drawings, with the emphasis on recording fabric, levels and visible character rather than building a fully structured model. Detailed CAD elevations can be particularly valuable for façades, ornate interiors and listed features where linework, notes and graphic clarity matter.

CAD has practical advantages on projects with limited intervention. If the proposal concerns a small extension, a change of use or internal reconfiguration, the critical need may be an accurate measured base drawing delivered quickly. Provided the drawings include the agreed levels, dimensions and sections, a model may offer little additional value.

That does not make CAD a lower-standard service. Accuracy comes from the capture method, survey control, checking process and the experience applied when translating complex conditions into usable drawings. For irregular buildings, a carefully considered section or elevation can reveal far more than a generic modelled form.

When a BIM model earns its place

BIM becomes more valuable when several disciplines must work from the same existing-condition reference, or when design decisions depend on understanding relationships in three dimensions. Complex refurbishments, major fit-outs, adaptive reuse projects and constrained extensions are common examples.

A Revit model can help an architectural team develop options without repeatedly rebuilding the base building. It can support coordination with structural and building services consultants, provide a basis for clash reviews, and make it easier to extract consistent plans, sections and elevations as the design evolves. Where client standards or consultant workflows require Revit, starting with a dependable existing model can save significant internal production time.

The value is strongest where the building itself is difficult. Sloping floors, irregular roof geometry, vaulted spaces, non-orthogonal layouts and layered historic alterations are hard to understand from partial records. A model does not remove that complexity, but it can organise it into a usable spatial reference when the underlying capture and modelling are handled with care.

BIM is not always necessary for a listed building simply because it is historic. The deciding factor is how the model will be used. A conservation statement supported by accurate 2D survey drawings may be appropriate. A phased renovation involving multiple disciplines, detailed interventions and frequent design coordination may justify a model, often alongside tailored CAD drawing outputs.

LOD is a scope decision, not a quality label

One of the most common misunderstandings is that a higher Level of Development automatically means a better BIM deliverable. It does not. LOD describes the intended definition and reliability of model elements at a particular stage. It should be selected according to the decisions the model needs to support.

An LOD100 or LOD200 model may be suitable for massing, feasibility, area studies and broad spatial coordination. LOD300 is often appropriate where building elements need dependable size, location and form for design development. LOD400 is more detailed and may be required for fabrication, installation planning or specialist coordination, but it requires a clearly defined purpose and substantially more effort.

For existing buildings, the requested LOD must also reflect what can be observed, measured and reasonably represented. Concealed structure, hidden services and inaccessible roof zones should not be presented with false certainty. A dependable model makes its assumptions clear, records exclusions and focuses effort where the project risk is greatest.

A useful brief states which elements need modelling, the required LOD, the expected tolerances, the software version, the preferred origin and coordinates, and the outputs needed alongside the model. Without this, a request for an “accurate Revit model” leaves too much open to interpretation.

Start with the decisions the information must support

The most effective way to choose between CAD and BIM is to work backwards from the project’s next decisions. Ask whether the team needs to test spatial options, coordinate multiple disciplines, prepare planning drawings, record heritage fabric, calculate areas, or issue a base model to consultants. The answer usually clarifies the format.

It is equally important to consider who will maintain and use the information. If the architectural team will work solely in AutoCAD, issuing a BIM model may introduce an unnecessary conversion step. If consultants will build their design packages in Revit, issuing coordinated CAD files alone can shift substantial modelling work back onto the design team.

The programme matters too. CAD can often be the fastest route to an accurate design base, especially when a limited drawing set is required. BIM production may take longer because each element must be interpreted, structured and checked in three dimensions. That additional time is worthwhile when it prevents repeated remodelling and coordination effort later. It is less worthwhile when the model will be used once and set aside.

The capture method determines what either format can achieve

Whether the output is a DWG drawing set or a Revit model, dependable documentation starts on site. High-accuracy 3D laser scanning provides dense spatial data that can be registered, checked and used to understand difficult geometry. The point cloud becomes an evidence base for drawing and modelling, rather than relying on sparse dimensions or assumptions taken from outdated records.

That evidence is particularly useful in complex and heritage settings. Historic buildings often contain uneven floors, settling walls, irregular openings and roof structures that do not follow a modern grid. A measured survey needs to capture meaningful departures from ideal geometry while avoiding unnecessary detail that obscures the information designers actually need.

The documentation team should also establish the project’s datum, access constraints, required coverage and expected tolerances before work begins. If the client needs floor levels, roof details, internal elevations or external context, these requirements should be agreed at quotation stage. Late additions can affect both programme and the completeness of the output.

A combined deliverable is often the strongest option

The decision does not always have to be CAD or BIM. Many projects benefit from a Revit existing-condition model supported by issued CAD and PDF drawings. This gives the design team a coordinated spatial base while retaining familiar, easily reviewable drawing outputs for planning, heritage consultation and site discussion.

A hybrid approach is especially useful when different stakeholders work in different ways. The architect may coordinate proposals in Revit, the conservation consultant may mark up 2D elevations, and the client may need a straightforward PDF set for review. Producing these outputs from a controlled documentation workflow reduces the risk that each party works from a different version of the building.

For projects across England and Scotland, Space Captures scopes the required outputs around the building and the intended workflow, from measured CAD drawing packages to Revit models at the appropriate level of development. The aim is not to prescribe a format, but to provide dependable geometry that design teams can use immediately.

The useful question is not which format sounds more advanced. It is which documentation will remove uncertainty from the next decision, without adding cost or detail that the project will never use.