Information reviewed: 16 November 2026
“Accurate to ±2 mm” is not a complete survey specification. It might describe an instrument under stated test conditions, a registration statistic or a contractual requirement for a finished measurement. Those are different things. For a measured building survey, define the required accuracy of the delivered information, its reference and the checks used to demonstrate it.

The RICS plan accuracy table
The table below summarises the plan-position bands in the RICS third-edition measured survey standard. These are positional accuracy values relative to survey control, not permitted wall-thickness errors or a universal promise about every model element. Height requirements have a separate table and must be specified separately.
| Plan band | 1 sigma | 2 sigma |
|---|---|---|
| A | ±2 mm | ±4 mm |
| B | ±4 mm | ±8 mm |
| C | ±5 mm | ±10 mm |
| D | ±10 mm | ±20 mm |
| E | ±25 mm | ±50 mm |
| F | ±50 mm | ±100 mm |
Under a normal-error assumption, one and two sigma correspond approximately to 68% and 95% confidence. The standard includes further bands beyond F. Selecting a band is part of the project specification; a drawing scale alone does not establish it.
Why the scanner’s number is not the final answer
The finished drawing or model passes through several steps: instrument measurement, positioning between scans, control, interpretation and drafting or modelling. Access and surface conditions also affect what can be observed. A small scanner specification does not cancel errors introduced elsewhere.
Consider a window opening that will receive a close-fitting component. The critical question is whether its relevant edges have been captured and verified to the agreed requirement. A dense cloud of a distant façade, an impressive registration screenshot and a smooth Revit wall do not independently answer that question.
Set requirements by decision and feature
| Decision | Better briefing question |
|---|---|
| Early layout feasibility | Which dimensions and levels could change the design decision? |
| Façade coordination | Which openings, projections and adjacent surfaces must be resolved? |
| Heritage repair | Which irregular profiles must remain visible in the output? |
| Steel or joinery interface | Which local dimensions require a separate check before manufacture? |
These examples deliberately do not assign a universal band to a project type. A kitchen layout and a finished stone worktop can belong to the same project yet require different measurement decisions. The responsible designer or fabricator should set critical interface requirements and the timing of final verification.

A QA record worth asking for
Ask how the survey reference is established, where independent checks are made and what will happen if the checks do not meet the agreed limits. A useful handover identifies the areas tested, the nature of the checks and any unresolved limitations.
Registration residuals show how well the selected observations fit the registration solution. They are not automatically independent evidence of absolute accuracy. Checks that were not used to create the solution can provide a different and important test. Our point-cloud QA guide explains this distinction.
Drafting should also be checked. Confirm that floor plans align with the agreed sections, that levels use the same datum and that simplified surfaces are identified. A good point cloud can still produce a misleading drawing if an irregular wall is replaced without explanation by a convenient straight line.
Copy this into a survey brief
“Please state the proposed plan and height accuracy requirements, confidence basis and reference system. Identify critical features separately, describe independent verification, and record inaccessible or assumed geometry. Confirm whether the requirement applies to the point cloud, extracted drawings, model geometry or all three.”
Complete that statement with the project team’s actual numerical requirements. Do not leave “survey grade” or “high accuracy” as the only acceptance criterion. Compare our survey samples with the file handover checklist when deciding what evidence to request.

Frequently asked questions
Is Band A always the best choice?
It is the tightest plan band in this table, but suitability depends on the task and achievable site conditions. Unnecessary precision can increase effort without improving the decision; inadequate precision can make the output unusable.
Does more point density mean better accuracy?
No. More points may help describe a surface, but density and positional accuracy are separate properties. Coverage, control, surface behaviour and verification still matter.
Can I use survey drawings directly for fabrication?
Only where the appointment and responsible designer’s requirements support that use. Critical dimensions may need a later local check, particularly after alterations. Send your accuracy and deliverable brief for a scoped quotation.
References and review date
References checked on 16 November 2026. Requirements can change; confirm the applicable edition and project-specific requirements when appointing the team.

