A Knocked Sprinkler Head Costs $10,000 a Minute: How LiDAR Scanning Cuts Change Orders

A single knocked sprinkler head costs about $10,000 a minute — yet most Atlantic Canada building owners still walk into a renovation carrying roughly 14% in budget overages as if they were an act of God. LiDAR scanning and 3D documentation do not eliminate surprises. They move them: from the field, where fixing a surprise multiplies in cost at every stage, to the desktop, where it costs almost nothing. The 14% norm is not a law of physics. It is the price of choosing familiarity over information.

That image — a sprinkler head, sheared off by a crew that did not know what was above the ceiling, bleeding ten grand a minute while the insurance company's phone rings — is the controlling fact of this piece. It is small, it is specific, and it is exactly the kind of thing that nobody draws on a 2D plan. On the Atlantic Construction Podcast, Colin Gillis of Smarter Spaces put the number plainly: "a sprinkler head if you knock a sprinkler head off any type of damage is about 10,000 a minute" (Colin Gillis, EP 27). The discipline that prevents that minute from ever starting is the subject here — existing-conditions scanning, reality capture, scan-to-BIM — and the case for it in Atlantic Canada is not technological. It is arithmetic.

Why does the industry treat a 14% overage as normal?

The first thing to understand is that the overage is not an accident. It is budgeted. Ask any seasoned project manager what a $10-million renovation will actually cost, and the honest answer builds in a cushion that has nothing to do with the scope of work and everything to do with the unknown. Gillis names the figure: "typically in construction you're in for 14 percent in overages no matter what and it's just kind of crazy how it's just accepted" (Colin Gillis, EP 27). Read that sentence twice. The remarkable part is not the percentage — it is the phrase no matter what. An entire industry has agreed, in advance, to spend an extra one-seventh of every renovation budget and to call it the cost of doing business.

The independent data says the number is real, not regional bravado. Across commercial construction, change-order costs run roughly 8–14% of total project value, and on renovation-heavy or complex jobs they can climb toward 25% (Rhumbix). Canadian-specific tracking is bleaker still: somewhere between 70% and 85% of Canadian construction projects go over budget, and design changes alone account for more than half of cost overruns nationally (Graditi). Gillis's 14% is not an outlier he invented to sell scanning. It sits in the dead centre of the verified range.

What makes the norm so durable is that it feels like weather. You cannot control weather; you carry an umbrella. But the comparison breaks the moment you look at what causes the overage. The dominant trigger is not material price swings or labour markets — it is the gap between what the drawings say is in the building and what is actually there. Up to 70% of total rework in construction is design-induced, and roughly half of all rework on a site traces back to poor or incomplete project information (Autodesk/FMI). That is not weather. That is a choice about how much you bothered to learn before you started. The 14% is the receipt for that choice.

What does a LiDAR scan actually do?

Strip away the jargon and existing-conditions scanning is simple: it produces an exact, measurable record of a building as it really stands today, down to the millimetre, before anyone designs against it or demolishes into it. Smarter Spaces runs two families of tools. The mobile LiDAR unit is a backpack — the operator walks the space at a normal pace while a sensor "shoots about 40,000 points a second," building a dense cloud of measured points as it goes. The terrestrial scanner, a tripod-mounted Leica RTC360, trades mobility for precision: Gillis pegs its accuracy at one to three millimeters while it captures "like a million points a second."

The choice between them is a problem-solving decision, not a gadget preference — a distinction Gillis is insistent on. "It's really not about the technology it's about the value and the problems you solve" (Colin Gillis, EP 27). The clearest illustration is the occupied building. Mobile LiDAR captures geometry without high-definition photography, which makes it the right tool for the places where a renovation is hardest and a camera is forbidden. Smarter Spaces has used it in nursing homes and hospitals "where they actually don't want photos" for privacy reasons, walking the sensor through live wards and even, in one telling anecdote, through a boardroom mid-meeting while bewildered occupants wondered what the man with the backpack was doing.

Then there is the part nobody outside the field appreciates: the work is mostly after the visit. The scan in the building is the fast part. Converting raw point clouds into a usable drawing set, a CAD plan, or a BIM model is the slow part, and the ratio is steep — by Gillis's account, "for every hour we're in the field and there are a lot of variables there are probably three to four hours of production" to turn it into something a client can build from. That production happens in Nova Scotia, on Revit, by graduates of NSCC and the regional colleges. It is why the regional scan-to-BIM market exists at all, and it is why the deliverable — not the scan — is what the owner is actually buying.

The renovation math: what does half a change-order budget look like?

Here the argument stops being abstract. Smarter Spaces' core renovation pitch is a single before-and-after number, and Gillis states it without hedging: "just by engaging us bringing us in to update existing conditions just on that aspect alone we've reduced change orders by about 50 so on a 10 million project you're talking about 1.4 million" (Colin Gillis, EP 27). Run that out. A $10-million project carrying the industry-standard 14% is carrying $1.4 million in change orders. Cut that in half by walking a scanner through the building first, and roughly $700,000 stops being spent — for the cost of a scan.

That $700,000 is the figure to hold against the price of the scan itself. North American scan-to-BIM runs $0.50–$10.00 per square foot depending on detail, with basic architectural modelling at the low end and dense MEP documentation at the high end; point-cloud-only capture is cheaper still, $0.20–$0.70 per square foot (iSCANO; Existing Conditions). On a mid-sized Atlantic Canada renovation the scan is a fraction of one percent of contract value. The change-order budget it attacks is 14%. The ratio is not close.

The worst-case lives above the ceiling. Gillis describes hospital above-ceiling renovations as "about the most complicated, awful thing to experience" — a tangle of mechanical, electrical, and infection-control infrastructure that the traditional method assessed by having a tradesperson "pop his head up and try to guess what was up there." Guessing is expensive. On those jobs the change orders ran far past the norm: "above ceiling can be even worse, like 75 percent over" (Colin Gillis, EP 27) — and as Gillis put it, it was ridiculous how much got pushed through because you don't know what's there. The phrase because you don't know what's there is the whole thesis in five words.

New builds carry the same logic underground. Scanning geothermal pipe runs as they go in, Smarter Spaces has found that "from the original drawings to what we scanned and where they're actually located could be off by as much as 10 feet" (Colin Gillis, EP 27). A ten-foot error is invisible the day the slab is poured and catastrophic the day, five years on, a backhoe goes looking for a pipe the drawing misplaced — turning, as Gillis puts it, a $10,000 parking-lot fix into a $500,000 one. The scan does not change the pipe. It changes whether anyone knows where it is.

Why is catching a defect in design so much cheaper than catching it on site?

There is a corollary to the renovation math that applies to every project, scanned or not: the cost of a defect rises at every stage it survives undetected. The cheapest place to fix anything is on a screen, before it is built. Charlene Cormier, working the building-envelope side of the business, frames the whole discipline of pre-construction review around exactly this escalation: "you want to catch this into design if you're catching this during construction it's going to cost you more" (Charlene Cormier, EP 49). Once a wall is built wrong, the money to build it is already spent — by the general, the owner, the sub — and now you are paying a second time to tear it out.

This is precisely why a scanned existing-conditions model pays off upstream. Accurate as-built data at the design stage means every downstream decision rests on what is real rather than on an unverified assumption inherited from a decades-old drawing set. The national data backs the mechanism: design changes drive 56.5% of cost overruns in Canada (Graditi), and a BIM-led design process reduces most of the root causes of construction claims — missing drawings, design errors, late changes — before the project breaks ground (PMC/NIH).

The clash-detection record makes the point in hard numbers. When Pomerleau's team built a coordinated 3D model before construction, the conflicts it surfaced were not theoretical. Ashwin Rajendran recounts that "before we even had the first bucket in the ground" the team modelled the building and "we identified i don't know probably two dozen issues" (Ashwin Rajendran, EP 14). Two dozen conflicts resolved on a screen are two dozen change orders that never reached the field, where each unresolved MEP clash averages $1,500-plus and the rework and waste can consume up to 30% of cost on dense jobs (Budlong). Rajendran is careful, though, to locate the value in people rather than software: "it's not just having the technology but also having the knowledge and everybody understanding what we have" (Ashwin Rajendran, EP 14).

The return on doing this work early is the most lopsided number in the whole field. Terry Hussey states it as a flat multiple: "every dollar you spend in early stage planning you'll save a hundred times that back in construction" (Terry Hussey, EP 43). That is a practitioner's headline figure; the documented industry range for pre-construction coordination ROI is a more conservative $5–$10 returned per dollar, with one detailed case study netting $2.55 million in savings against a $200,000 coordination investment (DBIA). Whether the multiple is 10x or 100x, the direction is never in dispute: information bought early is the cheapest information you will ever buy.

How does a 3D render become an insurance policy?

The same principle that governs structure governs finishes and design intent, and this is where the visualization layer earns its place. Luminous Labs builds photorealistic renders and visualizations of buildings before they exist, and Nick LeBlanc frames the deliverable not as a sales tool but as risk transfer. "it kind of is an insurance policy like if you actually give us all the information we need and we do things the right way" (Nick LeBlanc, EP 77). The render is where you discover that a material reads wrong, a proportion fails, or a finish clashes — while the only thing it costs to change is a few hours at a workstation.

The economics are the renovation math repeated at the level of the finish. Catching a problem on a render and swapping a material on paper is trivially cheap; catching it after the work is built means tearing out finished work and paying for it twice. LeBlanc puts the trade in plain figures: "if we're saving 100,000 or 200,000 or whatever that amount is well that's a valuable tradeoff" (Nick LeBlanc, EP 77). And the failure mode he describes is the familiar one — the owner who skips the upfront visualization, builds on assumption, and only later sees the problem that a render would have caught. Had they done the work first, "they would have been able to say you know that's not flat" before it was ever framed.

There is a clean pipeline hiding in these two businesses. Smarter Spaces captures what is real; Luminous Labs visualizes what is proposed against it. Scan the existing building to millimetre accuracy, design and render the new work on top of that verified geometry, and the owner is no longer making finish decisions against an imagined room — they are deciding against a measured one. The scan removes the guesswork from the geometry; the render removes the guesswork from the look. Both move the surprise from the field to the desktop. That is the entire game.

What is the point cloud worth after the ribbon is cut?

A scan's value does not end when the project closes. The point cloud and the 360-degree imagery captured behind the walls and above the ceiling — Smarter Spaces shoots them before they're sealed up — become a permanent, queryable record of the building. Gillis's example is the maintenance call years out: instead of opening drywall to chase a leak, a facilities staffer can "pull up on your phone unit 302 jump into that unit and see where all the pipes are located exactly what's above the ceiling" (Colin Gillis, EP 27). The sprinkler head that costs $10,000 a minute is far less likely to get knocked when the crew can see where it is before they cut.

This lifecycle value is the answer to the most common objection, which is that the spend is front-loaded. It is — but the payback runs for the life of the asset, against the building's ongoing maintenance cost. The contrast with current practice is stark. The information advantage a live point cloud gives a facilities team is precisely what the rest of the trade still lacks: as Karsh observes from the estimating side, "95 of the trades are estimating using PDFs" — and the moment one of them lands a job, "the second they get a job they're like what did I miss" (Karsh, EP 56). What did I miss is the entire problem expressed as a question. A flat PDF cannot answer it. A measured 3D model can.

Why do Atlantic Canada owners still resist the upfront spend?

If the math is this lopsided, why is adoption not universal? Gillis is candid that the hard part of the sale has nothing to do with the technology and everything to do with time preference. "that's probably the toughest sell to spend upfront to save five years from now" (Colin Gillis, EP 27). The cost is visible and immediate; the saving is invisible and deferred. A building owner who has never had a $500,000 surprise has no felt reason to spend on preventing one — and so, in Gillis's experience, the first-time builder often passes, and only becomes a customer after the second project, once the lesson has been paid for the expensive way.

Underneath the time-preference problem is an older inertia: the conviction that the way it has always been done is the way it should be done. The pattern recurs across the show — the operator who resists a better method because the old method built the building that is still standing. It is a deeply human instinct and a deeply expensive one, because the building industry's defaults were set before millimetre-accurate scanning existed. The contrarian point is not that anyone is foolish for following the convention; it is that the convention itself is now obsolete, and the 14% overage is the bill the convention quietly mails every year.

The structural barriers are real and worth naming honestly, because pretending the spend is frictionless helps no one. Canadian BIM-adoption research lists the same impediments repeatedly: software and training cost, unclear business value, small firm size, trade fragmentation, and — decisively for this region — the absence of any mandate (IAARC/ISARC). No Atlantic province has issued a BIM or 3D-scanning requirement for public projects; nationally, Canada is still at the roadmap stage through buildingSMART Canada (NRC). Every institution that scans today — and the regional client list runs to health authorities, universities, and municipalities — does so voluntarily, on a project manager's judgment, not under compulsion.

Yet the forcing functions are arriving from the side. Fire safety is legislated through the National Building Code, and Gillis notes the city's auditor general recently spotlighted buildings that had gone years without inspection — which sends owners scrambling to update fire-safety plans on three weeks' notice, plans that are far cheaper to produce when an accurate model already exists. Accessibility is the next wave: Smarter Spaces has certified with the Rick Hansen Foundation ahead of Accessibility 2030, anticipating the day when verifying door widths and switch heights against a desktop model beats repeated site visits. Regulation, in other words, is quietly converting the voluntary scan into a near-necessity — not because anyone mandated LiDAR, but because the evidentiary bar for compliance now rewards owners who already know exactly what they are working with.

What should an owner actually do with this?

The practical takeaway is a sequence, not a sentiment. The scan earns its keep at four moments, and an owner planning any significant work should know which one they are at. Bring a scanner in pre-design, so the architect designs against measured reality rather than legacy drawings. Bring one in pre-demolition, especially on heritage or historic buildings, before anything that cannot be put back is removed. Bring one in during construction, on a cadence matched to the project's complexity, so clashes between the as-built and the model surface in days rather than months. And bring one in above the ceiling and behind the walls before seal-up, capturing the lifecycle record that turns a future leak from a demolition job into a phone lookup.

The questions an owner should put to a general contractor follow directly. Ask how existing conditions were documented — by tape and flashlight, or by scan. Ask whether there is a coordinated 3D model and a clash-detection workflow before the first bucket breaks ground, the way Pomerleau ran two dozen conflicts to ground before mobilizing. Ask whether the MEP sub-trades are delivering models or, as is still common on Canadian jobs, flat 2D drawings that quietly reintroduce the very gaps the GC's model was meant to close (Mars BIM). And weigh the cost of a basic existing-conditions scan — a fraction of one percent of contract value — against the 14% the project is otherwise budgeting to lose. For the deeper logic of building against verified rather than assumed conditions, the same discipline runs through the air-sealing and radon guide and the broader construction technology hub.

The 14% is not the cost of construction

Return to the sprinkler head. The reason it costs $10,000 a minute is not that sprinkler heads are fragile or that crews are careless. It is that someone went into a ceiling they could not see, against a drawing that did not match, carrying an assumption instead of a measurement. The entire case for LiDAR scanning, scan-to-BIM, and reality capture reduces to closing that gap — to making sure the crew, the designer, the estimator, and the facilities team years later all know what is actually there before they act on it.

The 14% overage that the industry treats as fixed is nothing of the kind. It is the running tab on a single decision, repeated project after project: the decision to value the comfort of the familiar over the cost of the unknown. The operators who have done the arithmetic — on hospital ceilings, on geothermal fields, on rendered finishes, on the maintenance call five years out — keep arriving at the same on-the-record conclusion. The surprises in a renovation are real and they are not going away. The only thing you get to choose is where they happen: in the field, where they multiply, or on the desktop, where they cost almost nothing. The 14% is not the cost of construction. It is the cost of not knowing what you are working with before you start.