Chapter 39 — Key Takeaways

A one-page reference card. If you keep one thing from this chapter on your desk for ten years, make it this page.

Key Takeaways

• You do not need to be a technologist. You need to be a good evaluator. There will always be someone in the room who knows the tool better than you. There is rarely anyone who can say cleanly: who pays, who benefits, what breaks, and what has to be true here. That skill is scarce and it does not expire.
• Construction technology fails in four recognizable ways. The Simultaneity Trap (everybody must adopt at once), the Displaced Burden (it adds work at the point of use to produce benefit elsewhere), the Repeatability Assumption (it was engineered for a factory and your project is a prototype), and the Unread Dashboard (it produces information with no decision waiting for it). Learn the names and say them out loud in meetings.
• All four are faces of one economic fact: in construction, the party who bears the cost of adoption and the party who receives the benefit are frequently different parties, joined only by a contract that ends when the job ends. The corollary is the most useful screening rule in the chapter — the technologies that succeed here are overwhelmingly the ones where one party can adopt alone and capture the benefit alone.
• Reality capture is the clearest win in the chapter, and it is not close. On renovation work, one avoided surprise usually pays for the whole scan. Its only exposure is that somebody has to actually look at the deviation map.
• The cheapest insurance in this book is a 360-degree camera on a fixed weekly route. Northgate: $13,600 on a $47,500,000 job — 0.029% of the contract — used substantively eleven times in nineteen months. Capture every wall before it closes.
• The reality-capture loop only works if it is short. One bay caught before MEP was hung cost $2,100 to fix in the model; the same condition after the duct was hung would have cost about $18,400 plus three days. A scan compared six weeks later is an as-built record, not a control.
• A drone is a screen, not a scale. A $450 flight that tells you cheaply whether a claim is worth a $3,100 survey is a real and repeatable value. A drone that "replaces the surveyor" is a sales claim. And half of every volume calculation is the other surface — the existing ground you are differencing against.
• Prefabrication is the only thing in this chapter that changes how the building actually gets built. Everything else observes, measures, coordinates, or documents. That is why it is the hardest and why it is worth it: prefabrication converts assembly work into production work.
• On direct cost, multi-trade prefabrication is roughly a wash. Northgate: $1,845,000 stick-built against $1,820,000 prefabricated — 1.4%. It pays on schedule and safety, not on cost, and only when a coordinated model at fabrication level of development and a subcontractor with shop capacity are both already true.
• Hours moved are not hours saved. Northgate removed 6,800 MH from the site and added 4,300 MH in a shop. The honest productivity number is the 2,500 MH net, or 22%. The 6,800 is the safety story, and it is enormous — but it is not a cost saving, and calling it one is how prefabrication business cases lose their credibility.
• Days are only worth money when they are on the critical path. Northgate's maturity sensors compressed the foundation cycle by 9 calendar days, which looked like $46,350 at $5,150/CD and was worth $0 in extended general conditions, because foundations had 22 days of float. They were still worth $26,460 in labor and a disaster avoided on two cold placements. Always run the float test before you book a schedule saving.
• Check which rate you are entitled to apply. You save extended general conditions ($5,150/CD) by finishing earlier. You avoid liquidated damages ($5,500/CD) only if you were otherwise going to finish late.
• A platform does not create discipline. It makes the absence of discipline visible faster. Buying a field management platform and believing it is document control buys you an expensive way to be disorganized in public.
• Accountability does not transfer to a tool. If a system reviews a submittal and misses the rating, you still stamped it. If it generates a schedule and the sequence is wrong, you still owe the owner the date. Read the terms of service with the attention you would give an indemnity clause.
• These tools change what construction workers do far more than they reduce how many are needed — and in most markets the binding constraint is a shortage of skilled labor, not a surplus. Treat any claim that a technology will replace construction workers as marketing until proven otherwise.
• The most attractive thing on the table is often the one to kill. A technology that produces a beautiful visual is optimized to be bought. A technology that removes 6,800 man-hours from work at eleven feet in the air produces no picture at all.
• The winner is the one who can tell a tool that removes work from a tool that moves work. Both are sold with exactly the same vocabulary.

Action Items — What to Do on Your Job This Week

1. Buy a 360-degree camera and pick a route. One camera, one subscription, twenty minutes per floor per week, and one rule: capture every wall before it closes. You can start this on Monday for less than the cost of one opened and repaired rated wall.
2. Find your scanning program's reader. If your company scans anything, name the person who looks at the deviation map and the day of the week they do it. If you cannot, you own a very expensive archive, and the fix costs one calendar entry.
3. Run the four failure modes on one thing you are already paying for. Not a proposal — something live. Ask which mode it is exposed to and whether it has drifted into one since it was bought.
4. Answer gate question 5 for your most expensive running tool: who owns it on a Tuesday when the person who championed it is on vacation? If you cannot name them, you have found the next failure before it happens.
5. Ask one foreman what they get back from whatever you make them enter. If the answer is "nothing," you have located a Displaced Burden and roughly six weeks before the data goes bad.
6. Price the as-built. If your company hands owners a scanned as-built model at closeout for free, put it in the next two proposals as a priced deliverable with a written accuracy-and-reliance statement. You are giving away a product.
7. Before your next progress claim on a unit-price scope, find out what the contract actually designates as the measurement basis for payment. Decide it before there is money riding on the answer.
8. Write the neighbor letter before the first drone flight — what you fly, when, why, what the camera points at, who to call, and the commitment not to fly over their property. One afternoon, and it prevents the problem that nearly ended Kestrel's drone program.
9. Call one mechanical or electrical subcontractor and ask whether they have a fabrication shop, what percentage of their work goes through it, and what a general contractor has to do to make prefabrication work for them. Their answer is your project's first fix.

Common Mistakes — and the Fix

Decision Framework — Keep This on a Card

Step 1 — Five gates. Any one unanswered is a stop, not a discount.

1. What problem does it solve? Name the decision or the cost. Not "visibility." Not "insight."
2. What does it cost — all in? License, hardware, training, internal hours, three years, renewal price.
3. What does it return, and to whom? Dollars, days, or man-hours, landing on a named party.
4. What does it break? Workflow, contract, relationship, liability.
5. What has to be true here before it works at all? And who owns it on a Tuesday when the champion is on vacation.

Step 2 — Four scores, 1 to 5.

D is a gate, not a score. If you cannot name the person and the decision, kill it regardless of the total. The Repeatability Assumption is handled by gate 5, not by a score, because repeatability is binary in practice.

Step 3 — If it survives, design a pilot. All six elements, in writing, before it starts.

One scope, one project, one boundary · a named owner and a named executive sponsor · a fixed window · a pre-registered success criterion in numbers · a cost cap · a kill criterion you could act on the same afternoon.

Add the element almost every pilot skips: a control — a comparable conventional scope on the same project, measured the same way. Without it you have a demonstration, and demonstrations always succeed.

Step 4 — The four-question field screen, for when you have ninety seconds

1. How many parties must adopt before this produces value? (More than one is a warning; more than three is usually fatal.)
2. Who does the extra work, and do they get anything back on their own timescale?
3. What percentage of my actual scope is repetitive enough for this?
4. Name the person and the decision. If you cannot, stop.

And the one-line rule

Does it remove work, or does it move work?

Removing work: machine control, reality capture, a demolition robot, a rack built at a bench, current drawings on a device, a machine that reports itself. Moving work: a foreman typing for twenty minutes so somebody in an office can read a chart.

Both are sold with the same words. Everything in this chapter is a way of learning to hear the difference.