Reality capture covers four capture methods that feed a delivery platform: 360 photography, drone photogrammetry, LiDAR and Matterport scanning. In the UK these are often sold as a scanning service by the day, but a scan on its own is not the product. The value comes from what you build on top of the capture, and each method has a clear job and a clear point where it is the wrong tool.
We treat capture as an input, not a deliverable. The question we start with is never "which scanner", it is "what does someone need to do inside this space once it is captured". That answer decides the method, the accuracy you pay for, and the platform the capture lives in.
What are Matterport scanning services in the UK actually for?
Matterport scanning services in the UK produce a navigable 3D walkthrough of an interior with rough dimensions attached, and that is exactly the right tool for orientation, familiarisation and remote inspection. It is the wrong tool for engineering-grade measurement. Matterport's own guidance puts the Pro3 at a dimensional accuracy typically within 1% to 2% of reality at working distances.
In plain numbers, independent write-ups put the Pro3 at approximately 20mm at 10 meters. That is genuinely useful. For workflows like space planning, lease audits, or pre-renovation scoping, that's often enough to confirm a dimension or catch a discrepancy without dispatching someone to re-measure in person.
Where it stops being the right answer is measurement-critical construction. Matterport is clear about this itself: the Pro3 isn't designed for measurement-critical construction workflows like clash detection, prefabrication layout, or tolerance verification—survey-grade scanners still lead in those contexts. The confusion comes from the Pro3 carrying a LiDAR depth sensor, which invites people to treat a communication tool as a survey instrument.
“A scan on its own is not the product. What someone does inside it is.”
So we do not lead with Matterport as a standalone scanning service. We use it where the job is people understanding a space: induction, remote walkthroughs, asset familiarisation, handover records. Then we put it inside a branded platform the client actually runs their process through, rather than handing over a raw model and a login.
When is 360 photography the right capture, and when is it not?
360 photography is the right capture when you need a fast, cheap, photographic record of a space that people navigate visually, and it is the wrong tool the moment anyone needs to measure from it. A 360 tour is imagery stitched into panoramas. It carries context and detail beautifully. It does not carry reliable dimensions.
That is not a weakness, it is a different job. If the task is "show a distributed team what the plant room looks like", "brief a contractor before they arrive", or "capture site conditions on a given date", 360 is quick to shoot and quick to publish. It suits progress records and visual as-builts, where the useful question is whether something has been installed, not whether it sits at an exact height.
We reach for 360 when speed and frequency matter more than precision, for example a weekly visual record across a programme. We stop reaching for it the moment the deliverable is a floor plan, a volume or a coordinated model.
When should you use a drone, and photogrammetry or LiDAR on it?
Use a drone when the subject is large, external or hard to reach safely from the ground, and then choose between photogrammetry and LiDAR by what you need to see and how much vegetation is in the way. Both can deliver centimetre-level results when flown properly, so the choice is about method, not a simple accuracy league table.
Photogrammetry builds a model from overlapping photographs. It is the more affordable route and excels at colour, texture and fine surface detail, which makes it strong for orthomosaics, marketing imagery and progress tracking. Its weakness is that it requires consistent daylight. Its accuracy is sensitive to poor lighting, shadows, or overexposure that can distort the final model. It also cannot see through vegetation to the ground.
LiDAR fires laser pulses and, as an active sensor, works independently of natural light and can operate in complete darkness. It tends to produce greater detail and accuracy on elevation and complex structures, and crucially it can find the ground beneath tree cover where photogrammetry cannot. The industry consensus is even-handed: photogrammetry and LiDAR are equally accurate—assuming that you're using them in the right way.
So the honest split is this. Photogrammetry for surface detail, colour and cost. LiDAR for vertical accuracy, complex or partly shrouded terrain, and low light. LiDAR is the pricier setup and its main drawback has always been cost, though that gap has narrowed as sensors have shrunk onto standard drone payloads.
A drone is the wrong tool where the space is a tightly cluttered interior, where flight is restricted, or where the answer is a simple ground-level walkthrough. Do not fly something you could capture faster from a tripod.
When do you actually need survey-grade LiDAR?
You need survey-grade LiDAR when the measurement itself is the deliverable and a real tolerance is riding on it: clash detection, prefabrication, MEP coordination, as-built drawings or anything that will be manufactured or legally defensible. This is a different class of instrument from a Matterport.
The gap is large and worth naming. Professional terrestrial scanners operate at plus or minus 2mm or better across comparable distances. That is a 10x difference in dimensional precision against a Matterport Pro3. Independent testing bears this out, with one study placing the Pro3 at ±20mm accuracy at 10 meters and the earlier Pro2 far coarser at ±50–70mm accuracy.
The practical rule is the one Matterport itself lands on: match your accuracy requirement to your workflow, not the other way around. If the question is "has the drywall gone in on the second floor", a walkthrough answers it. If the question is "will this duct clear that beam", it does not, and paying for survey-grade capture is the cheaper mistake to make.
Why treat capture as an input to a platform, not a service sold by the day?
Because the scan is where most of the market stops, and that is where the value actually starts. Plenty of suppliers will sell you a day of scanning and hand back a model. Fewer will build the branded platform your teams and clients run their process through, which is where a walkthrough becomes an induction, a marketing tour, an inspection record or a handover system.
That distinction decides everything upstream. If the platform is an induction, we might capture in 360 or Matterport and never touch survey-grade kit. If it is property visualisation, photogrammetry and interactive rendering carry the work. If it feeds a coordinated model, survey-grade LiDAR is non-negotiable and Matterport would be the wrong input entirely.
So the useful conversation is not "how much per scan day". It is what the space needs to do once captured, who uses it, and to what tolerance. Get that right and the capture method chooses itself. Get it wrong and you either overpay for precision no one uses, or you build on a measurement that was never accurate enough to carry the weight.