Knowledge centerUSIBD Level of Accuracy (LOA)
USIBD Level of Accuracy is the US specification for stating how accurate building documentation is. It exists because “accurate” on its own is not a specification, and two firms can both be telling the truth while delivering very different things.
What LOA separates
The framework distinguishes measured accuracy — how closely the captured data matches reality — from represented accuracy, how closely the delivered drawing or model matches the captured data.
That distinction is the useful part. A survey can be captured to a high standard and then modeled loosely, and without separating the two nobody can tell where the tolerance was lost.
Why it is worth specifying
Most US firms quote against square footage and deliverable type, and say nothing about accuracy at all. Specifying an LOA turns a vague expectation into something both parties can check on handover.
It also stops the common failure where a client pays for high-accuracy capture and receives a model built to whatever tolerance was convenient.
How we work with it
We document capture method, control, and verified deviation per project — under 0.4 inches in point cloud comparison for terrestrial work. Where a project specifies a particular LOA level, we confirm in writing which one applies to which area before mobilizing. That agreement is part of scoping an existing conditions survey, and the neighbouring terms are in the knowledge center.
The specific LOA levels and their tolerance bands are defined in the USIBD Level of Accuracy Specification. Which level Metrika360 warrants for a given deliverable is agreed per project.
Where the specification comes from
The Level of Accuracy framework is published by the U.S. Institute of Building Documentation, a body formed by the people who produce and consume building documentation rather than by a software vendor. It exists because the word accurate carries no obligation on its own, and because two firms can both be honest and deliver very different things.
It sits alongside the level of development framework used for BIM, and the two answer different questions. Confusing them is the most common error in a brief that mentions either.
LOA, LOD, LOI
| Framework | Question it answers | Applies to |
|---|---|---|
| LOA, level of accuracy | How closely does this match the real building? | The survey and the deliverable produced from it. |
| LOD, level of development | How much is modeled, and in what detail? | The BIM model. See scan to BIM. |
| LOI, level of information | What non-geometric data is attached to each element? | Asset and operational data. See facility management data. |
A model can be highly developed and inaccurate, or accurate and sparse. Specifying one and assuming the others follow is how a client ends up with a beautifully detailed model built to whatever tolerance was convenient.
How to put it in a brief
Four lines are usually enough, and they save the argument that otherwise happens at handover.
- Which areas. Accuracy per area, not per project. A plant room and a warehouse do not need the same tolerance, and paying for the higher one across a whole building is the most common overspend in this industry.
- Measured or represented. State both. Capture can be excellent and the drawing produced from it loose, and without separating the two nobody can tell where the tolerance was lost.
- How it will be evidenced. Ask for the registration report and the deviation figure for your project, not a brochure number.
- What happens if it is missed. Rare, and worth agreeing anyway, because it forces both sides to be specific about what was promised.
The specific LOA levels and their tolerance bands are defined in the USIBD Level of Accuracy Specification. Which level applies to which area of your project is agreed in writing before mobilizing rather than asserted afterwards.
What gets assumed
- “Laser scanning is millimeter-accurate, so the drawing is too.” The capture is. What survives into a drawing or model depends on how it was produced, which is the entire reason represented accuracy is a separate idea.
- “Higher accuracy is always better.” It is always more expensive. Accuracy that nothing downstream consumes is money spent on a number in a report.
- “Everyone works to a standard.” Most US firms quote against square footage and deliverable type and say nothing about accuracy at all. That is not dishonesty, it is convention, and it is why specifying is worth the four lines above.
How the underlying measurement works is covered in LiDAR; the method difference that most affects accuracy is in 3D laser scanning. The rest of the vocabulary is in the knowledge center.
Frequently asked questions
Is LOA the same as LOD?
No, and they are frequently confused. LOD describes how much detail a BIM model contains. LOA describes how accurate it is. A highly detailed model can be inaccurate, and vice versa.
Do we have to specify an LOA?
Not formally, but the conversation is worth having. Even without citing the standard, agreeing what accuracy applies where prevents the most common dispute on handover.
Who publishes the LOA specification?
The U.S. Institute of Building Documentation. It is a US framework, which is one reason it is the right reference here rather than a European standard.
Can you state an accuracy figure before the survey?
We can state the method and what it typically achieves: terrestrial scanning millimeter-accurate, deviation held under 0.4 inches across a registered project. The figure for your building comes out of its own registration report, and that is the one we stand behind.
Documenting a building?
Tell us what you need measured.