James Duffy on What 3D Scanning Really Delivers | IFTI Q&A

Meet the Expert: James Duffy on Documenting Physical Space at Scale

James Duffy is Vice President of Business Development at Independent Floor Testing & Inspection (IFTI), where he has spent more than six years working with multi-site retail, restaurant, and manufacturing brands on documentation programs for their physical locations, according to Duffy’s LinkedIn professional profile. His job sits at the intersection most operations leaders care about: not “what does a 3D scanner do” but “how do I get reliable, repeatable documentation across a portfolio of buildings without a different vendor and a different file format at every site.”

We sat down with Duffy to separate what independent, peer-reviewed research says 3D scanning actually delivers from the marketing language that tends to surround it. The goal here is not to sell you on scanning technology. Matterport-powered capture is widely available, and IFTI is not the only company that owns a 3D camera. What follows is a grounded look at what a scan produces, where its limits are, and why the harder problem for a 200-location chain isn’t the scanner, it’s the program around it.

What Does a 3D Scan Actually Capture?

A 3D scan produces a point cloud: a dense set of measured coordinates representing surfaces in a space, not a finished architectural model. Terrestrial laser scanning can capture millions of these points within minutes and, under favorable conditions, achieve single-point accuracy at the millimetre level, according to the 2021 ScienceDirect review “Review Planning for terrestrial laser scanning in construction: A review.”

“People hear ‘3D scan’ and picture a finished blueprint. What you actually get first is raw geometry, millions of measured points. Everything useful happens after that.”, James Duffy, VP of Business Development, IFTI

Duffy’s answer tracks with the literature. A 2021 review published by the National Institutes of Health/PMC, titled “Application of Terrestrial Laser Scanning (TLS) in the Architecture, Engineering and Construction Industry,” analyzed 209 publications and grouped terrestrial scanning uses into five categories: 3D model reconstruction, object recognition, deformation measurement, quality assessment, and progress tracking. Notice that “instant floor plan” isn’t one of the five. The plan, the CAD file, the walkthrough, all of it is derived work built on top of the raw capture, which is why the next question matters more than the scan itself.

From Point Cloud to Deliverable: Floor Plans, CAD, Measurements, and Walkthroughs

A raw point cloud becomes useful only after it’s processed into a specific deliverable, such as a measurable walkthrough, a 2D floor plan, a CAD file, or a full BIM model, each requiring different levels of cleanup, modeling, and human judgment. Point clouds support downstream uses including BIM, monitoring, cost estimation, quality assessment, visualization, and spatial analysis, particularly for reducing rework caused by incomplete existing-condition data, according to research titled “Reality Capture of Buildings Using 3D Laser Scanners,” published through the MDPI Buildings journal and the Oxford Brookes research repository.

Confusion between these three terms, point cloud, 3D model, and BIM model, is common enough that it’s worth laying out side by side.

For most multi-location operators, a measurable walkthrough paired with a 2D floor plan and CAD export covers the bulk of design, leasing, and facilities use cases without the added cost of full BIM modeling, which Duffy reserves for renovation or engineering-heavy projects. For a 30-unit restaurant chain updating kitchen layouts, that combination is usually enough to scope the work; a ground-up remodel or MEP-heavy retrofit is where the added cost of BIM starts paying for itself.

How Accurate Is 3D Scanning, And Accurate for What?

Scan accuracy depends on the equipment, distance, surface conditions, and how multiple scans are aligned, and it is always tied to a specific tested configuration rather than a single universal number. A 2022 Scan-to-BIM accuracy study, “Scan-to-BIM method in construction: assessment of the 3D model accuracy,” published in a Taylor & Francis journal, found roughly ±8 mm of uncertainty (about the thickness of two stacked credit cards) under the most unfavorable tested measurement configuration, an average inventory-model accuracy near ±6.8 mm, and dimensional identification potentially better than ±1 cm under tested conditions. Those figures describe that specific study’s equipment and setup, not a guaranteed outcome for every job site.

“Clients ask me for ‘the accuracy number’ like it’s one fixed spec. It isn’t. It’s a function of the equipment, the site, and how the scans get stitched together. That’s why we validate every project instead of assuming.”, James Duffy, VP of Business Development, IFTI

Terrestrial (fixed-position) systems tend toward millimetre-scale accuracy, while mobile scanning systems trade some precision for faster coverage, a trade-off documented in the 2021 Taylor & Francis literature review “Mobile 3D scan LiDAR: a literature review.” A comparative study titled “Point Cloud Registration and Accuracy for 3D,” published in the Journal of Information System and Technology Management, found six-scan terrestrial laser scanning reaching roughly ±4.8 mm versus ±6 to 8 mm for photogrammetry on the same subject, illustrating that method choice, not just brand of hardware, moves the number. On a 100,000-square-foot distribution center, that gap between methods can be the difference between measurements a design team trusts outright and measurements that need a spot check first.

Owners who need to specify a tolerance before a project starts have a vocabulary for it: industry frameworks such as the U.S. Institute of Building Documentation’s Level of Accuracy (LOA) specification define five tiers ranging from LOA10 (50 mm, roughly 2 inches) to LOA50 (1 mm, roughly 0.04 inches), alongside guidance referenced in the GSA BIM Guide. Most facilities-documentation work for retail and restaurant portfolios lands around LOA20 to LOA30 (15 mm, roughly 0.6 inches), which is precise enough for space planning and leasing decisions without the cost of survey-grade capture. Naming the LOA tier you need up front, rather than assuming “3D scan” implies one fixed precision, is the single best way to avoid a mismatch between expectation and deliverable.

Where Scans Run Into Trouble: Limitations Every Owner Should Know

3D scanning has known failure points, including line-of-sight gaps behind obstructions, distortion from highly reflective or glossy surfaces, and registration error that accumulates when multiple scans are aligned incorrectly. In a restaurant kitchen, stainless steel prep tables and hood systems are common culprits; in a retail location, floor-to-ceiling glass storefronts can produce the same effect. Understanding these limits before a project starts prevents surprises in the final deliverable.

Registration, the process of stitching individual scan positions into one unified coordinate system, is a documented risk area. The 2018 MDPI Sensors review “Registration of Laser Scanning Point Clouds: A Review” notes that alignment error can accumulate across scans and distort the final dataset if not properly controlled. A single mis-registered scan in a 40,000-square-foot facility can throw off measurements in a way that isn’t obvious until someone tries to build from it, sometimes only surfacing weeks later when a contractor’s field measurements don’t match the file.

Scan quality should be validated, not assumed, through checks like stitching reports, dimensional-accuracy spot checks, point-cloud thickness review, and model-layering inspection, according to 2024 research titled “Application of 3D laser scanning technology for mapping,” published in Nature’s Heritage Science journal. That validation step is where a managed program earns its keep over a one-off freelance scan: someone is actually checking the data before it reaches your design team.

Scan, Survey, or Both? Knowing When You Need More

A 3D scan is not a certified survey or stamped engineering document, and it should not be treated as one for legal, permitting, or structural decisions. Scans are best used for planning-reference measurements, existing-condition documentation, and remote decision-making, while a licensed surveyor or engineer should still be engaged for any work requiring certified accuracy or legal sign-off.

Duffy is direct about this boundary: scanning removes the need for repeated site visits during the design and planning phase, but it doesn’t replace the professional of record on a project that requires one. Where a scan genuinely changes the workflow is upstream of that step, giving a design or construction team enough reliable existing-condition data to scope a project correctly before a surveyor or engineer is ever engaged, which is one of the reasons incomplete existing-condition data is a documented driver of construction rework.

Standardizing Documentation Across a Multi-Site Portfolio

The real value of 3D scanning for an operator with dozens or hundreds of locations isn’t the camera, it’s having one process, one data format, and one point of accountability across every site instead of a patchwork of local vendors and inconsistent files. Anyone can buy a 3D camera. Few organizations can run a professionally managed capture program at portfolio scale.

“Our clients place a lot of trust in our expertise and expect us to accurately and efficiently document their spaces.”, James Duffy, VP of Business Development, IFTI (Matterport case study, “IFTI emerges as a leading construction surveyor by embracing 3D technology,” 2019)

This is the piece that generic “book a scan” services can’t offer: consistent field-service quality across geographies, standardized deliverable formats so a facilities director in Ohio and one in Arizona get the same file structure, and a single point of contact instead of chasing down a dozen local freelancers when a lease renewal or renovation timeline is tight. That operational layer, coordinating technicians, quality-checking capture, and delivering data through one shared portal, is what turns “we did a cool scan once” into “we always know what our buildings look like.” If your portfolio is still running on that patchwork model, seeing a coordinated program in action is the fastest way to spot the gap.

The Bottom Line for Multi-Location Teams

3D scanning gives operations, design, and facilities teams a shared, remote-accessible record of a building’s existing conditions, cutting down on redundant site visits and giving multiple teams the same starting data. It is not a substitute for certified survey or engineering work, and its accuracy depends on the equipment, site conditions, and validation process used, not on the word “3D” alone.

For a single-location renovation, a one-off scan might be enough. For an organization managing 50, 200, or 800 sites, the deciding factor isn’t which company owns the best camera. It’s which partner can deploy trained technicians consistently, validate the data the same way every time, and deliver it through one system your design, construction, and facilities teams can all pull from without waiting on a local vendor’s schedule.