Chapter 37 — Residential Construction Management: The Unique Challenges of Building Homes

The Hook: Nine Houses Before Lunch

I drove out to Harbor Ridge on a Tuesday in March of Year 2 expecting to spend a morning looking at a small job.

That was the whole frame I brought with me. Northgate is 132,000 square feet, four stories, forty-seven and a half million dollars, thirty-four subcontracts, a Primavera schedule with eleven hundred activities, and a submittal log I have personally lost sleep over. Harbor Ridge is thirty-four lots of two-story houses off a county road, built by Tessa Bright Homes, a production builder that closes somewhere around ninety homes a year. I assumed I was going to see a smaller version of my own job — same problems, fewer zeros, less paperwork.

Colton Reyes met me at the model home at 7:15 with a coffee and a clipboard. He is the superintendent for the whole community. Not for a house. For all of it.

The clipboard had one page on it. Two columns. On the left, house numbers — lot 4, lot 7, lot 9, lot 11, lot 12, lot 14, lot 17, lot 19, lot 22, lot 23, lot 26. Eleven of them. On the right, a single word or two: rough, insulate, paint, frame, dig, drywall, trim, cabinets, dry-in, final, flatwork.

That was the schedule. That was the entire schedule.

"Where's the rest of it?" I asked.

"That is the rest of it."

We walked nine houses in three hours. He was in each one for eight to fifteen minutes. In lot 22 he looked at four things — the shower pan, the tub blocking, whether the tub protection was on, and whether the plumber had left the DWV test caps in — and then walked out. In lot 12 he stopped, put his hand flat against the sheathing at a window head, said "that flashing's lapped wrong, that's going to be a call," and sent a photo and one sentence to the framing foreman before we were back to the truck. In lot 26 he stood in the driveway for ninety seconds counting trucks in the street, then made a phone call I did not understand, which turned out to be moving an insulation crew from a house that would not be ready until Thursday to a house that had passed rough inspection Friday.

At 10:40 I asked him the question I had been carrying since 7:15. How do you keep eleven buildings straight in your head with a clipboard and no schedule?

He did not answer right away. He put the truck in park and turned it off, which I have learned is how a superintendent tells you the answer is going to take a minute.

"You're managing a project," he said. "I'm managing a line. That's the whole difference and everything else follows from it. My problem isn't that a house takes ninety-two days. Ninety-two days is fine. Ninety-two days takes care of itself. My problem is that a finished house has to come off the end of this line every eight days, forever, and that means a new one has to go on the front every eight days, forever, and there are eleven of them in between at all times and every single one of them is in a different stage."

Then he said the sentence I have quoted in every job interview I have conducted since.

"You've got one building and a critical path. I've got no critical path and eleven buildings. What I've got is a rate. And the only thing that can hurt me is anything that makes the rate go down."

I have been in commercial construction for twenty-two years. I had assumed for most of them that residential was the shallow end of the pool — the same work with fewer requirements, where people go when they cannot handle a real project. That is a common belief in my half of the industry and it is wrong in an interesting way. Harbor Ridge is not a smaller Northgate. It is a different business with a different unit of production, a different customer, a different risk profile, a different money structure, and a scheduling technique that has almost nothing to do with the one I spent four chapters of this book teaching you.

It is also, and I want to be careful here, the closest thing construction has to manufacturing — and the place where every lean idea in Chapter 27 lands most naturally, because the work actually repeats.

By the time I got back to my trailer that afternoon I had filled six pages of a notebook. This chapter is those six pages, worked out properly, with the arithmetic done.

🏃 Fast Track: If you already build houses, skip §37.1 and §37.4. Read §37.2 and §37.3 anyway — most production builders I have met run a line by feel and have never written down the throughput arithmetic, and the arithmetic is where the money is. Then go to §37.6 (what is missing and where the absence hurts) and §37.11.

🔬 Deep Dive: The scheduling technique here — line of balance — is the repetitive-work counterpart to the CPM you built in Chapter 14; read them together. Flow, variability, and the reliability chain are Chapter 27. Envelope and energy execution is Chapter 36. Offsite and modular get their proper treatment in Chapter 39. If you are a homeowner rather than a builder, Appendix E is written for you.

Every company, person, and project in this book is a Tier-3 illustrative composite. Harbor Ridge, Tessa Bright Homes, and everyone in this chapter are made up. The numbers are internally consistent and realistic; they are not a real community.

37.1 Five Businesses Wearing One Word

The word "residential" covers at least five genuinely different businesses. They share a building type and almost nothing else. If you take a job in one of them believing you are in another, you will make the specific mistakes that segment punishes.

Let me be honest about multifamily right now and then mostly set it aside. A 240-unit wood-frame apartment building over a concrete podium is a commercial project with residential finishes. It has an institutional owner, a real contract, a submittal log, RFIs, a schedule of values, retention, pay applications, an OAC meeting, and a lender who wants a monthly report. Almost everything in the first thirty-six chapters of this book applies to it directly. The one genuinely residential thing it borrows is unit repetition — the same 46 unit layouts repeated across five floors — which means the scheduling ideas in §37.3 apply to the interior buildout even though the contract is commercial. That is the whole of multifamily's special case, and pretending otherwise wastes your time.

The rest of this chapter is about the other four, and mostly about the first two, because production and custom are the poles that define the field.

37.1.1 The two poles

Production building and custom building are opposite in almost every respect, and a manager who is excellent at one is often actively bad at the other.

Hold that table. Almost every specific technique in this chapter is a consequence of one row of it.

🔄 Check your understanding. A builder who builds twelve houses a year, each one a different design for a different family, on the family's own lot, using cost-plus contracts — which segment is that, and which single risk should dominate their management system?

37.2 The Unit of Production: Cycle Time, WIP, and the Start Interval

Here is the single idea the chapter turns on.

On Northgate, the unit of production is the building, and the question is "how long does it take?" The answer is 565 calendar days and everything I do is aimed at not making that number bigger.

At Harbor Ridge, the unit of production is the house, and the question is not how long a house takes. It is "how often does one come out?" Colton's 92 calendar days per house is a fact about a single house, and by itself it tells you nothing about the business. The number that describes the business is the rate.

🧩 Productive struggle. Before you read the next paragraph, do this on paper. Colton has eleven houses in process at all times and each house takes 92 calendar days from the day the excavator opens the hole to the day it is finished and ready to close. Give me two numbers: (1) how many calendar days apart he must start houses, and (2) how many houses per year that line produces. Take three minutes. The arithmetic is easier than you expect and the interpretation is harder.

37.2.1 The arithmetic

There is one relationship underneath all of this and it is not complicated. In manufacturing it is called Little's Law; in construction nobody calls it anything, which is part of the problem.

Work in process = throughput × cycle time

WIP = T × C

Where WIP (work in process) is how many units are being worked on at once, throughput is units completed per unit of time, and cycle time is how long one unit takes from start to finish. Rearranged two ways, it gives you everything:

Throughput = WIP ÷ cycle time Start interval = cycle time ÷ WIP

Plug in Harbor Ridge.

(CD = calendar days. Residential builders schedule in calendar days almost exclusively, because the customer's closing date, the lender's interest clock, and the property tax bill all run on calendar days. See Chapter 14 §14 on calendar days versus work days.)

What those four numbers mean, in plain English.

Colton starts a house roughly every eight and a half days and finishes one roughly every eight and a half days, permanently. His line produces about 43.6 houses a year. Harbor Ridge's thirty-four lots will take about twelve months from the first excavator to the last closing — the 92-day cycle for the first house, plus 33 more starts at 8.36 days apart.

And Tessa Bright Homes closes around ninety homes a year, which means Harbor Ridge is roughly half the company. Run the same law backward on the company: at 90 homes a year and a 92-day cycle, the company is carrying

WIP = throughput × cycle time = (90 ÷ 365) × 92 = 22.7 houses in process company-wide

about twenty-three houses under construction at any moment, of which Colton is personally responsible for eleven. At the company level a house closes about every four days (365 ÷ 90 = 4.06). At Colton's level one closes every 8.36. Both numbers are true; they describe different lines.

💡 Aha moment. On Northgate I can tell you what a day costs: $10,650 — $5,150 of extended general conditions plus $5,500 of liquidated damages — and it costs it once, when I lose it. At Harbor Ridge a day of cycle time does not cost anything once. It costs something every year, forever, until somebody takes it back out of the line. That is not a difference of scale. It is a difference of kind, and it is why production builders behave the way they do.

37.2.2 💰 Money check: what one day of cycle time is worth

This is the arithmetic that makes the chapter. Do it once and you will never look at a residential schedule the same way.

First, the economics of one Harbor Ridge house. These are Tessa Bright's actual numbers for the community's blended average, and the split between base and options is the part most people have never seen.

Look at that middle column before you go on. Options are 10.5 percent of revenue and 19.5 percent of gross margin. The design center — the room where the buyer picks the quartz and the wood floor and the third-car bay — is not a customer-service amenity. It is where roughly one dollar in five of the builder's margin is made. I will come back to what that does to the schedule in §37.7.

Now perturb the cycle time by one day and hold WIP constant at 11.

0.47 houses/year × $438,000 = $205,400 of annual revenue 0.47 houses/year × $63,500 = $29,800 of annual gross margin

One calendar day of cycle time on Colton's line is worth about $29,800 a year, every year. Across all twenty-three houses Tessa Bright has in process, the same day is worth about $61,500 a year to the company.

Say it the other way, because that is how you will actually use it. If Colton can find five days in his 92 — a day out of framing, a day out of the rough-inspection wait, three days out of the finish sequence — the community produces

11 ÷ 87 × 365 = 46.15 houses/year — up 2.5 houses, worth $159,000 a year in margin

for exactly zero additional lots, zero additional sales staff, and zero additional overhead. That is why a good production superintendent walks nine houses before lunch and looks irritated at a house with nobody in it. He is not being a hard case. He is standing in front of thirty thousand dollars a year.

⚠️ Safety alert. Read that paragraph again and then read this one. Everything I just showed you is a machine for generating schedule pressure, and schedule pressure is a hazard exactly like an unguarded edge — canon theme #4 and the lesson of Chapter 24. The residential sector's fall-injury record is genuinely worse than commercial's, and the mechanism is not mystery: small crews, piece-rate pay, multiple tiers of subcontracting, and a superintendent doing exactly the arithmetic above. A builder who shows the rate math to the field without also showing §37.9 has built an incentive to work a roof without fall protection. Do both or neither.

🔄 Check your understanding. Colton's boss asks him to raise output by starting more houses — go from 11 in process to 14 — without changing anything else. What actually happens to his three numbers?

37.3 Line of Balance: How You Schedule a Line

Now the technical core.

You spent Chapter 14 learning the critical path method, and CPM is the right tool for Northgate. It is close to useless at Harbor Ridge, and understanding why teaches you something real about both.

37.3.1 Why CPM does not fit

Build a CPM network for one Harbor Ridge house. You will get maybe 45 activities, a critical path that runs straight down the middle of the sequence with essentially no parallel paths worth mentioning, and total float of approximately zero everywhere because the sequence is almost purely serial. Fine. Now do it for all thirty-four houses. You have a 1,530-activity network in which the same logic repeats 34 times, and the thing that actually governs — the framing crew cannot be in two houses at once — is not in the network at all unless you resource-load it, at which point the "critical path" becomes a resource artifact that jumps between houses on every update and tells you nothing you can act on.

🔍 Why this works. CPM answers the question "what is the longest chain of dependent work through one pass of this network?" That is exactly the right question when you build a thing once. In repetitive work it is the wrong question, because the binding constraint is not a path — it is a rate. The interesting question is not "which activities are critical" but "which trade's production rate is lowest, and is it lower than the rate the line needs?" CPM has no vocabulary for that question. Line of balance is built out of nothing else.

There is a second reason, and it is about people, not math. In repetitive work you must give each crew continuous work — house after house without gaps — or they leave. A CPM schedule optimized for shortest total duration will happily produce a plan in which the framers work for six days, sit idle for four, and work six more. That plan is arithmetically valid and operationally suicidal. Line of balance treats crew continuity as a hard constraint and duration as the output. That is the correct priority when the crews do not work for you.

37.3.2 The technique

Line of balance (LOB, sometimes called linear scheduling or the repetitive scheduling method) charts unit number on one axis and time on the other. Each trade becomes a sloped line. The slope is the trade's production rate — how many units per week that crew completes. The rules are short:

1. Each trade's line must have a slope at least as steep as the line's required rate (one house per 8.36 days at Harbor Ridge). A shallower slope means that trade cannot keep up, and it will govern everything behind it.
2. Trade lines must not cross. A crossing means a trade catches the trade in front of it and is standing in a house that is not ready.
3. Between adjacent trades you leave a buffer — in time, in units, or both — so ordinary variability does not turn into a collision.
4. The slowest trade sets the rate of the entire line. Not the average. Not the total. The slowest.

That last one is the whole discipline. Everything else is bookkeeping.

37.3.3 📊 Diagram (described): the balanced line

Here is Harbor Ridge running correctly. House number runs up the left; time runs left to right. Each column is 4 calendar days. Five trade stages are plotted at the point where they start in each house: E = excavation and foundation (cycle day 0), F = framing (day 19), M = MEP rough-in (day 37), D = drywall (day 50), T = trim and finish (day 59).

 House                              time  →   (each column = 4 calendar days)
   │
H11│                      E    F   M  D  T
H10│                    E    F   M  D  T
H09│                  E   F    M  D T
H08│                E   F    M  D T
H07│              E   F    M  D T
H06│           E    F    M  D T
H05│         E    F    M  D T
H04│       E    F    M  D T
H03│     E    F   M   D T
H02│   E    F   M   D T
H01│ E    F   M   D T
   └────────────────────────────────────────────────
     0   20   40   60   80  100  120  140  (calendar days)

     E = excavation + foundation    F = framing    M = MEP rough-in
     D = insulation + drywall       T = trim through final

Read it as five parallel diagonal lines marching up and to the right at the same slope. That parallelism is balance. Every trade is advancing one house every 8.36 days. The horizontal distance between two adjacent letters on the same row is that stage's position in the 92-day cycle; the vertical distance between the same letter on adjacent rows is the start interval. Nobody catches anybody. Every crew has a house waiting when it finishes the last one.

Notice something that is easy to miss: the line is not "faster" than a one-off house. House 1 still takes 92 days. What the line buys you is that houses 2 through 34 come out 8.36 days apart instead of 92 days apart. Production is not about making one thing quickly. It is about making the next thing without waiting.

37.3.4 📊 Diagram (described): what a shallow slope does

Now break one line. Drywall — hang, tape, finish, texture — was taking Colton's crew 6 days per house, comfortably faster than the 8.36 the line needs. In February it went to 9 days per house, because Coastline Drywall split its crew between Harbor Ridge and another builder's community. I will tell you why they split it in §37.5 and in Case Study 1, and the reason is not what you would guess.

Nine days per house is a slope shallower than the line's. Watch what it does across the whole community. Below, r marks the day a house is genuinely ready for drywall (insulation complete and inspected), and D marks the day the drywall crew can actually get to it. Every column is 6 calendar days; only every third house is shown.

 House                                                      time  →
   │
H34│                                                 r  D        wait = 21.0 days
H31│                                            r   D            wait = 19.1
H28│                                        r  D                 wait = 17.2
H25│                                    r  D                     wait = 15.3
H22│                                r D                          wait = 13.4
H19│                            r D                              wait = 11.5
H16│                        rD                                   wait =  9.5
H13│                   r D                                       wait =  7.6
H10│               rD                                            wait =  5.7
H07│           rD                                                wait =  3.8
H04│       D                                                     wait =  1.9
H01│   D                                                         wait =  0.0
   └──────────────────────────────────────────────────────────────
     40      100     160     220     280     340  (calendar days)

     r = house ready for drywall (insulation inspected)
     D = drywall crew actually starts

The two lines diverge by 0.64 days per house — the difference between 9.00 and 8.36 — and the gap never closes. It is not a delay. A delay is a fixed quantity you can recover. This is a rate mismatch, and a rate mismatch compounds: house 4 waits two days, house 10 waits six, house 22 waits thirteen, house 34 waits twenty-one. By the end of the community the last house sits finished-and-insulated for three weeks with nobody in it.

The arithmetic of what that costs:

Two things in that table are worth stopping on.

First, the naive answer is wrong. If you reasoned "drywall went from 6 days to 9, so the cycle went from 92 to 95, so output goes from 43.6 to 42.3," you lost 1.4 houses. The real loss is 3.1 houses — more than double — because once a trade's capacity falls below the line rate, that trade stops being a duration and starts being a ceiling. Output is no longer 11 ÷ 92. It is 365 ÷ 9.

Second, look at the cycle time: 99 days, not 95. Where did the extra four days come from? From Little's Law running backward. WIP is still 11, throughput fell to 0.1111 houses/day, so cycle time = 11 ÷ 0.1111 = 99 days. Those four days are houses standing in a queue in front of the drywall crew. You did not schedule them and they are not in anybody's plan, and they are physically visible on the ground as finished-and-insulated houses with the doors locked. Learn to read that. On a production site, a row of houses all sitting at the same stage is not a coincidence; it is a queue, and a queue means you have found your constraint.

37.3.5 Buffers, even flow, and the takt connection

Three practical consequences.

Buffers are not slack; they are the shock absorber. In a line, variability in one trade propagates to every trade behind it. The cheapest protection is a small, deliberate buffer between adjacent trades — typically one house or two or three days. Colton runs about a two-day inter-trade buffer in the finish stages and a one-house buffer ahead of drywall (there is always one insulated, inspected house waiting). That buffer is the reason a plumber who is a day late on one house does not shut the community down. It is exactly the reliability chain idea from Chapter 27, applied to a line instead of a look-ahead.

Even flow is the central discipline. Even flow means starting houses at a constant interval, matched to what the trades can absorb and what sales can sell, and not deviating — not for a good month, not for a bad month, not because a sales manager wants inventory for a spring promotion. The production builders who survive downturns are the ones who held even flow when it was tempting to gun it. Starting too many houses feels like ambition and produces inventory, longer cycles, angry trades, and a balance sheet full of unsold drywall.

This is takt, and it is the same idea as Chapter 27's. Takt time is the interval at which a unit must be completed to meet demand. Colton's takt is 8.36 days per house. The takt planning that commercial lean practitioners apply to a hospital floor — divide the floor into zones, size the work so each trade finishes a zone in the same time, march the trades through in a train — is exactly what a production builder does with houses. Residential got there first, without the vocabulary, because the repetition was handed to them by the product.

🔄 Check your understanding. Colton's framer, Ardent Framing, completes a house in 12 days with one crew. Is Ardent a constraint on the 8.36-day line?

37.4 The 92 Days: Stages, Gates, and the Inspector's Calendar

Here is Colton's actual stage list. Nobody at Tessa Bright Homes calls it a schedule; they call it the sequence, and every superintendent in the company has it memorized. The durations are calendar days for one house, and they include the weekends the stage spans, because that is how a builder thinks about a closing date.

Two stages run in parallel and are therefore not in the running total: exterior finish (siding or masonry veneer, soffit, fascia, gutters) starts around day 37 once the house is dried-in and wrapped, and site work (driveway, walks, patio, final grade, landscape and irrigation) starts around day 73. Both must land before the finals. Either one, forgotten, holds a certificate of occupancy on a house that is otherwise done — and I have seen a closing move a week because nobody ordered the driveway.

Notice what the stage table does not have: any float. Three plus two plus four plus … sums to exactly 92. There is no contingency in it anywhere. That is deliberate, and it is why the drywall slip in §37.3.4 turned into a cycle-time slip with nothing to absorb it. A production builder's protection is not float inside the house. It is the buffer between trades across eleven houses, plus the superintendent's ability to move a crew from a house that is not ready to a house that is. Colton's protection is optionality across the line, not slack within the unit.

37.4.1 The inspector is on your critical path and does not work for you

In commercial work the AHJ (authority having jurisdiction) matters, but on a $47.5M hospital job the building official is one constraint among a hundred and the general contractor has full-time staff whose job is to manage the permit and inspection process.

In residential the inspector is a genuine, first-order schedule driver, and there is exactly one person managing it: the superintendent, on his phone, at 6:40 in the morning.

Jurisdictional variation here is enormous and you must check yours. Some building departments take requests until 3:00 p.m. and inspect the next morning. Some are four to seven business days out in the spring. Some combine the four rough inspections into one visit and some require four separate visits with four separate correction cycles. Some jurisdictions do not inspect insulation at all; others require a third-party energy rater who is not a municipal employee and books two weeks ahead. Some allow partial inspections of a demarcated area; some do not. None of that is a national rule and none of it is stable — departments change staffing, adopt new code cycles, and change their scheduling systems.

Here is the management consequence, and it is the whole reason I put this table in the chapter. A one-day-per-inspection turnaround difference, across seven inspections, is seven days of cycle time — worth about $209,000 a year to Colton's line (7 × $29,800). Which means the following are not administrative chores; they are line-rate decisions:

• Knowing your department's request cutoff and hitting it every time.
• Never calling an inspection on work that is not finished. A failed inspection is not a neutral event — it costs the trip, the correction, and a second trip, and in a busy season the second trip is four days out. Colton's rule is that he personally walks the house before he calls the four-way, and it is the highest-return hour in his week.
• Asking, once, whether your jurisdiction allows combined or partial inspections, and building your sequence around the answer.
• Keeping a relationship with the inspector that is professional, prepared, and boring. An inspector who trusts your houses inspects them faster. An inspector who has failed you three times reads every joist hanger.

⚖️ What the contract says. In residential you are usually the permit holder as owner-builder or licensed contractor of record, which means code compliance is legally yours in a way it may not be on a commercial job where a design professional of record carries part of it. Residential contractor licensing is a state and sometimes local matter — some states license residential builders separately from commercial contractors, some have a home-improvement contractor registration with its own written-contract and disclosure requirements, and some do not license general residential building at all. Working unlicensed where a license is required can, in some states, void your right to be paid at all, lien or no lien. Check your state before you sign anything, and do not assume the rule in the state next door.

37.5 The Trade Model: Paid by the Unit, Scheduled by Phone

This is where a commercial manager gets the biggest shock, and it is worth spending real time on because the entire labor and procurement machinery you learned in Chapter 16 and Chapter 19 is replaced by something that looks, at first, like nothing at all.

37.5.1 Priced by the unit, not by the job

On Northgate, Cardinal Mechanical holds a $6,400,000 lump-sum subcontract for one building, with a schedule of values, monthly billing, retention, and a scope sheet I spent two weeks writing.

At Harbor Ridge, Fairwater Plumbing has a price list. So does everybody else. Here is the trade cost structure of one 2,180-square-foot Harbor Ridge house.

Read the "how it is priced" column. Almost nothing is a lump sum for a scope of work. Almost everything is a unit rate against a measurable quantity, agreed once for the community and applied to every house.

Why does this matter? Because it changes what estimating is. On Northgate, Chapter 12's takeoff-and-unit-cost machinery produces a number for a building nobody has built. At Harbor Ridge, the estimate for plan 2180 is a known quantity — you have built it eleven times — and estimating collapses into two much narrower questions: did the unit rates move? and did the options mix change? The whole apparatus of bid tabs, scope sheets, and buyout that occupies weeks on a commercial job is replaced by a spring conversation about whether framing labor goes from $8.10 to $8.35 a foot.

37.5.2 Scheduled by phone, the day before

There is no six-week look-ahead at Harbor Ridge. There is Colton, at 6:40 a.m., making eleven phone calls.

"Lot 19 passes rough today, you're insulating Thursday." "Lot 12's board is stocked, Ines can start Wednesday morning." "Don't come to lot 23, the meter's not set. Go to 26."

That is the entire coordination system, and commercial managers find it appalling until they understand what is holding it together. It works because of three things:

1. The sequence never changes. The same twenty-one stages in the same order, thirty-four times. Nobody needs a plan telling them what comes next, because what comes next is what came next last time.
2. The product never changes. Four plans. The plumber has roughed plan 2180 forty times. There is nothing to coordinate about it.
3. The crews are the same crews. The same six framers, the same four-person drywall crew, house after house. There is no mobilization, no orientation, no learning curve, no submittal, and no meeting.

Repetition and standardization are doing the work that planning and coordination do on a commercial project. Take away the repetition — a new plan, a revised plan, a custom lot, an odd elevation — and the system has no reserve, because the reserve was never built. I will come back to that in §37.6.

37.5.3 💡 Aha moment: you are competing for a slot in their line

Chapter 19's threshold concept was that you do not manage the work; you manage the people who contracted to do the work, and your leverage is the subcontract, the schedule, and the coordination — not authority. Residential has its own version of that idea, and it is sharper, because in production housing you often have no meaningful subcontract leverage at all.

Ines Duarte's drywall crew hangs and finishes about sixty houses a year. Roughly eleven of those are Tessa Bright's. The other forty-nine belong to two other builders. Coastline Drywall is not a subcontractor on your project. Coastline is running its own line, and you are one input to it.

Now ask the question that reorganizes residential trade management: what happens when Ines's crew shows up at your house and the house is not ready?

On Northgate, if a subcontractor arrives at an unready area, we have a conversation, they go somewhere else on the job, they document it, and eight weeks later we argue about a disruption claim. Annoying, expensive, survivable.

At Harbor Ridge, Ines's crew has already been paid for nothing that morning — six people, a truck, a day. She has one house of yours and forty-nine houses of somebody else's. Tomorrow she does not come back and try again. Tomorrow she goes to the other builder, whose house was ready, and you do not get another slot until she has a gap. And because she is running her own line, "a gap" is not tomorrow. It is next week.

In production housing, your single most valuable asset with a trade is not price and not the contract. It is being the builder whose houses are always ready when the crew shows up.

That is not a soft skill. It is the mechanism by which a builder either holds his rate or loses it, and it is why the drywall slip at Harbor Ridge happened. Four times in nine weeks, Coastline sent a crew to a Tessa Bright house that was not ready:

Four wasted trips. Each one cost Coastline roughly six hangers × 8 hours × $34/hour loaded = $1,632 of payroll with zero production. Four trips: $6,528.

Coastline's gross margin on a Tessa Bright house is about $1,352 on $9,656 of revenue. So four wasted trips wiped out the margin on nearly five houses. Reggie Prosser, who owns Coastline, did not fire Tessa Bright. He did something more rational and much worse for Colton: he split the crew, so that Coastline always had a second builder's house to fall back on when a Tessa Bright house was not ready.

A split crew hangs and finishes a house in nine days instead of six. And that — not laziness, not greed, not another builder paying more — is why the slope of the drywall line went shallow and why Tessa Bright is losing $196,000 a year in gross margin.

The counterintuitive finding, which you will work through properly in Case Study 1: the constraint is not the drywall crew. The constraint is the builder's own readiness. Colton created the bottleneck and then spent six weeks trying to solve it as a subcontractor problem.

🔄 Check your understanding. Why is "pay Coastline more" a weak response to this problem, even though it is the first thing most managers reach for?

37.6 What Residential Does Not Have, and Where That Hurts

This is the most useful comparison in the chapter, and I want to do both halves of it honestly: what is genuinely absent, why it is absent, and — the half people skip — where the absence hurts.

🔍 Why this works. Look down the "why it is absent" column and you will see one mechanism repeated: repetition substitutes for planning, and standardization substitutes for coordination. A submittal exists to answer the question "is the thing you intend to install the thing the documents require?" When you have installed that exact thing thirty-three times and it passed inspection thirty-three times, the question has already been answered and the log is pure cost. An RFI exists to resolve an ambiguity in a document. When the document has been successfully built thirty-three times, the ambiguities are gone — not because anybody resolved them formally, but because the field resolved them once and remembered.

This is not sloppiness. It is a rational allocation of management effort to a production system with very low uncertainty. And it means the entire system's reserve is the repetition itself. Take that away and you have a project management system with no project management in it.

37.6.1 The three places it breaks

The first house of a new plan. Tessa Bright added plan 2650 to Harbor Ridge in month five. The first one took 117 days instead of 92 — 25 days over — and generated eleven field decisions that nobody wrote down. Every trade's price was wrong because the quantities were estimated instead of measured. The framer found a beam that did not bear on anything and fixed it with a phone call, and the fix is now in one house and not in the plan, and nobody will remember in three years. That is a prototype, and the professional answer is to treat it like one: schedule the first house of a new plan with 25 to 30 percent more time, walk it deliberately at every stage with the trade leads, write down every field change, and update the plan set and the price list before house two. Almost nobody does this. It is the single highest-return process improvement available to a production builder and it costs about eight hours of somebody's attention.

A plan revision mid-community. The window supplier discontinues a unit. The truss manufacturer changes a bearing detail. The energy code cycle changes and the insulation package has to change with it — see Chapter 36. Now you have houses 1 through 14 built one way and houses 15 through 34 built another way, with no submittal log, no revision control, and no record of which house got which. Five years later, in a warranty claim, someone will ask you which houses had the old flashing detail and you will not be able to answer. The fix is a one-page revision log per community — date, what changed, which lots are affected, who was told. Twenty minutes per revision. It is the cheapest insurance in residential construction.

Any deviation from the standard. A buyer wants the kitchen island moved 14 inches. There is no change-order process because the system was built on the assumption that nothing changes. So it gets handled verbally, it does not get priced, the electrician does not hear about it, and the pendant lights land 14 inches off center — which is a callback, a drywall patch, a paint patch, and an unhappy customer who tells nine people. This is the residential echo of CO #14 at Northgate: the change that got built before it was priced. Same failure, three orders of magnitude smaller, roughly the same relative damage. Chapter 31's discipline scales all the way down.

⚖️ What the contract says. The absence of a submittal log does not mean the absence of an obligation. You are still building to an approved permit set, an energy compliance report, a manufacturer's installation instructions (which the code generally makes mandatory for listed products), and whatever your purchase agreement or warranty promised the buyer. A substitution nobody logged is still a deviation from the permit set, and if it shows up in a defect claim, "we always do it that way" is not a defense. Keep a revision log even though nobody makes you.

37.7 The Money: Draws, Purchase Agreements, Options, Allowances, and Carry

Residential money is structured completely differently from commercial money, and this section is the one commercial managers most consistently underestimate.

37.7.1 The construction loan and the draw schedule

On Northgate, Kestrel bills Meridian Health System monthly on an AIA G702/G703 pay application against a schedule of values, retention is withheld at 10 percent until 50 percent complete and then 5 percent, and Meridian pays in 30 days. That entire machinery — Chapter 32 — assumes there is an owner on the other side of the transaction.

In production and spec building the builder is the owner. There is no owner to bill. There is a construction loan, and the bank releases money in draws when a stage is complete and the bank's inspector verifies it.

Harbor Ridge's vertical construction loan on a plan 2180 is $246,000, released in six draws:

Three things about that table are the whole lesson.

The builder finances the gap. Between draw 3 on day 37 and draw 4 on day 53, Colton spends $49,200 on rough MEP and insulation and receives nothing. Then the bank inspector takes three to seven business days and the funding takes two more. So the money he spent on day 40 arrives around day 60 — call it 20 days of float on roughly a draw's worth of cost, per house.

Multiply by WIP. Take the average unfunded balance per house at about $45,000. With eleven houses in process, Tessa Bright is standing in roughly $495,000 of its own working capital at any moment on this community alone — money that is not profit, not available, and not visible on an income statement. This is exactly the point Chapter 32 makes about cash flow not being profit, in a form where the number is easy to compute and easy to underestimate.

Cycle time and cash are the same variable. Every day of cycle time is a day of unfunded balance across every house in process. This is the second, quieter half of the $29,800-per-day figure from §37.2.2: the margin loss is the visible half, and the working-capital drag is the half that decides whether you can start the next community.

37.7.2 The purchase agreement, options, and the closing

The contract in production housing is not a construction contract at all. It is a real-estate purchase agreement: a promise to convey a completed house and its lot on a date, in exchange for a price, with earnest money as the buyer's consideration and a financing contingency as their escape.

This has consequences that surprise commercial people:

• There is no "owner" during construction. The buyer does not direct the work, does not approve submittals, does not run a punch list, and has no contractual right to be on the site — though most builders allow scheduled walks, because a buyer who feels shut out becomes a difficult closing.
• There is no retention. Nobody withholds anything from the builder. The builder gets paid once, in full, at closing.
• There are no liquidated damages for lateness, usually. What there is instead is an outside date, after which the buyer may typically walk and recover their earnest money — which is far worse than LDs, because you lose the sale and the house becomes inventory.
• Options are handled at a cutoff, not as change orders. Structural options (that third bay, the extended patio, the optional bath) lock before permit or before framing. Finish options lock before rough MEP if they affect wiring or plumbing, and before the material order otherwise.

And options are where the money is. Recall §37.2.2: options are 10.5 percent of revenue and 19.5 percent of gross margin, at a 27.0 percent margin rate against the base house's 13.0 percent. That single fact drives behavior you will otherwise find inexplicable — the design center staffed with a designer, the incentive structure for the sales team, the willingness to hold a lot open another two weeks for a buyer who is "still deciding."

It also creates the schedule risk. Every option that is not selected by its cutoff is a house that cannot start its next stage. A production builder's selections process is therefore not a customer-service function; it is a schedule gate with a dollar sign on it. If Colton's option cutoffs slip by an average of four days per house, his cycle time goes to 96 and his line drops to 41.8 houses a year — a $117,000 annual margin loss created entirely inside a sales office.

⚖️ What the contract says. Purchase agreements for new homes are governed by real-estate law and by consumer-protection statutes that vary a great deal by state. Common provisions worth understanding before you sign or draft one: the earnest-money and default terms; who bears the risk of loss before closing; the outside completion date and the buyer's remedy; the express warranty and whether it disclaims implied warranties (many states recognize an implied warranty of habitability or workmanlike construction for new homes, and whether and how it can be waived varies); any binding-arbitration clause; and any statutory right-to-cure procedure the state requires a homeowner to follow before filing a construction-defect claim. None of this is uniform, several states have amended it in the last decade, and nothing in this book is legal advice. Read your state's statute and have a lawyer read your form.

37.7.3 Custom-home contracts and the allowance problem

Custom building runs on one of two structures, and the choice is the same risk-allocation decision you studied in Chapter 4, with a client who has never made it before.

And then there are allowances — a dollar amount included in the contract for a scope the client has not yet chosen. Cabinets, countertops, plumbing fixtures, lighting, flooring, tile, appliances, landscape. Allowances exist because you cannot price a kitchen before the client picks the kitchen.

Allowance overruns are the single most common source of custom-home disputes. Not defects. Not delay. Allowances. The mechanism is worth stating precisely because it is so avoidable:

1. The builder sets allowances from a standard sheet, at a level that produces an attractive total contract price.
2. The client hears "$38,000 for cabinets" and pictures the cabinets in the magazine.
3. The client goes to the showroom and picks something that costs $71,400.
4. Everyone is surprised, and the surprise arrives ten months into an eleven-month relationship, when there is no money left and no goodwill left either.

There are two structural traps inside that sequence, and both are contract-drafting failures rather than character failures.

Trap one: material-only allowances. If the contract says the flooring allowance is $28,000 and is silent on whether that includes labor, the client will assume it is the finished floor and the builder will assume it is the material. Then the client upgrades from carpet to tile — which costs more in material and three times as much to install — and the labor overrun is not covered by any allowance at all. Fix: state every allowance as an installed allowance, material plus labor plus fee, or state explicitly and in bold that it is material only.

Trap two: silence on the fee. If the contract says overruns are billed "at cost" and does not say whether the builder's fee applies, you have written a lawsuit. On the $129,450 of allowance overruns in Case Study 2, an 18 percent fee is $23,301 — and both readings of the contract are honestly arguable. Fix: one sentence saying whether the fee applies to allowance adjustments, and whether underruns are credited to the client.

37.7.4 Spec building and the cost of thirty days

A spec house is a house built on the builder's own money with no buyer. Every day it stands unfinished is a day of carry, and the arithmetic is not intuitive.

Take one Harbor Ridge spec: lot $92,000 plus $246,000 of vertical cost, drawn over the cycle. The average outstanding loan balance during construction is roughly $215,000 — the lot from day one, plus about half the vertical draws on average. What does 30 extra days of cycle time cost?

(Use your own interest rate. Construction-loan rates move constantly and vary by builder, lender, and credit; 9.5 percent is an illustrative all-in rate for this example and nothing more.)

$2,542 of hard carry is 4.0 percent of the $63,500 gross margin, spent to accomplish nothing. That alone should get your attention. But the hard carry is not the real risk, and this is the part that separates people who have built specs from people who have not:

• The slot. Those 30 days occupy a lot and a position on the line. At Colton's rate, 30 days of a slot is roughly 3.6 house-slots' worth of the community's total capacity consumed by one house standing still.
• The market. Thirty more days is thirty more days of exposure to a market you do not control. A 1.5 percent softening over a quarter is $6,570 on a $438,000 house — two and a half times the hard carry, and you cannot see it coming. Spec builders do not usually go broke on interest. They go broke on timing.

🔄 Check your understanding. A production builder and a spec builder both lose 30 days on one house. Why is the loss structurally different?

37.8 The Homeowner Is the Customer

Everything above is production. This section is the part commercial managers underestimate most badly, and it applies to custom and remodel work above all.

Pri Sethi, Meridian's owner's rep, is a professional. She has a budget, a board, a hospital operations committee, a construction background, and a career that will continue after this building. When I tell her a number moved, she asks why, checks it, and decides. She is spending the institution's money.

A homeowner is none of those things. A homeowner is:

• Spending the largest sum of their life, often their entire savings plus thirty years of future income.
• Emotionally invested in a way no institutional owner ever is. It is not a building. It is where their kids will grow up.
• Untrained. They cannot read the drawings. They do not know what a rough-in is. They think a wall is a wall.
• Present. They will visit. They will visit in the evening. They will bring their mother.
• Slow to decide, and prone to changing decisions, because every decision is permanent, unfamiliar, and expensive.

None of that is a character flaw. It is the predictable behavior of a rational person making irreversible decisions in a domain they do not understand with money they cannot replace. Design the process for that person, not for the person you wish they were. This is canon theme #6 — you build with people, not with materials — in its purest form.

37.8.1 The selections schedule is a deliverable, not a courtesy

Late selections are the single biggest schedule risk in custom work. Not weather. Not trades. Not permits. Selections.

The professional answer is a selections schedule issued as a contract exhibit at signing, with three columns that most builders leave out:

The three columns people leave out are the last three, and they are the ones that make it work. A deadline with no stated consequence is a suggestion. A deadline with a consequence — "the builder will select the base item and any later change is a change order at cost plus fee plus lead time" — is a schedule.

Two more rules that cost nothing:

• Never ask a client to decide something they cannot yet picture. Take them to the framed house and walk the electrical plan with a roll of blue tape. Put the flooring sample on the actual floor in the actual daylight. Decisions made from a catalogue get changed; decisions made in the space stick.
• Bring the deadline forward once, in writing, before it is missed. "Your cabinet decision is due next Friday; here is what happens if it moves." A client who blows a deadline they were warned about twice is in a different conversation from one who was never told.

37.8.2 Communication cadence and the walkthrough

The cadence that works on a custom job is boring and unglamorous and it prevents most disputes:

That last row deserves emphasis. A homeowner who tells a trade partner to move something has just created work nobody priced, nobody scheduled, and nobody will remember agreeing to. It is not malice; they do not know it is not allowed. Say it out loud at the pre-construction meeting: "If you want something changed, tell me, not the crew — and I will get you a price before anybody touches it."

37.8.3 Renovation with the family in residence

Remodel work has the highest margins in residential and the highest variance, and both come from the same source: you are working inside someone's home while they live in it.

The technical risks — unknown conditions behind the finishes, undersized existing structure, aluminum wiring, knob-and-tube, unpermitted prior work, asbestos in flooring or duct wrap, lead paint in pre-1978 housing (which triggers real federal requirements for renovators, plus state and local rules), a wall you assumed was non-bearing — are real and belong in your contract as an allowance or an explicit unknown-conditions clause. Price them or exclude them, but never assume them away.

But the thing that decides whether the client refers you is not the technical work. It is:

In renovation, the customer experience is the product. Two contractors can install the identical kitchen for the identical price, and one gets three referrals and the other gets a complaint to the licensing board, and the difference is entirely in the list above. I have watched a technically flawless job end badly because the crew used the client's guest bathroom for six weeks without asking.

37.8.4 Warranty, callbacks, and why referrals are the marketing budget

Residential warranty structures vary enormously. A widely used express structure in U.S. new-home building is one year on workmanship and materials, two years on mechanical, electrical, and plumbing systems, and ten years on major structural elements — often called "1-2-10." That is a common commercial practice, not a national legal standard. Statutory warranty obligations, implied warranties, statutes of limitation and repose, and mandatory right-to-cure procedures differ substantially by state and change with legislation. Do not represent any warranty period to a buyer without confirming what your state requires and what your own written warranty says.

What is universal is the economics of the callback.

In residential, referrals and online reviews are the marketing budget. A production builder with a good callback reputation buys fewer leads. A custom builder with three referrals a year does not advertise at all. A remodeler with a one-star review answering the phone at hour 48 instead of hour 4 has just paid more for that defect than any repair would have cost.

So the operational rule is not "fix everything." It is:

1. Answer fast. The response time matters more than the repair time. An acknowledgement within a day converts an angry customer into a patient one.
2. Send one person who can decide. A warranty tech who has to "check with the office" doubles the trip count and halves the credibility.
3. Track callbacks by cause and feed them back to the line. Colton's warranty manager, Kelvin Boatwright, sorts every callback by stage and trade. Three consecutive houses with the same nail-pop pattern is not three callbacks — it is a drywall-timing problem in the sequence, and fixing it upstream is free compared with fixing it thirty-four times downstream. This is Chapter 23's cost-of-quality argument with an unusually fast feedback loop, because you build the same house again next week.

37.9 Quality, Inspections, and Safety on a Residential Site

37.9.1 Quality is one person's eye

On Northgate we have an inspection and test plan, hold points, witness points, mockups, benchmark installations, a third-party testing agency, a commissioning agent, and a nonconformance procedure with four dispositions and named approvers. Read Chapter 23 again if that is fuzzy.

At Harbor Ridge there is Colton, and a checklist, and eight to fifteen minutes per house.

That is not a criticism; it is a design. The quality system in production housing is the superintendent's stage walk, and it works when three conditions hold: the checklist is written down by stage, the walk happens before the work is covered, and the same person walks every house so that deviation is visible against a memory of thirty-three previous houses.

Here is the shape of Colton's checklist. It is not exhaustive — it is the list of things that are expensive or impossible to fix later.

Two of those rows carry most of the money. Dry-in flashing decides whether the house leaks, and a leak found in year three is a callback, a repair, a mold conversation, and possibly a defect claim. Pre-drywall air sealing and insulation contact decides whether the house performs, and it is the row that ties directly to Chapter 36: the energy code applies to this house exactly as it applies to Northgate, and envelope execution — not envelope specification — is what determines whether the blower-door number passes. A house can be insulated exactly to code and fail the test, because insulation is not an air barrier and a batt stuffed behind a wire is not R-13.

In most jurisdictions today, a new house must demonstrate airtightness by a blower door test and duct systems must demonstrate leakage limits by a duct leakage test, performed by a third-party rater. Both are pass/fail, both happen at day 90 of a 92-day cycle, and both are effectively un-fixable at that point without opening finished work. Which is why the real test is the pre-drywall walk on day 53, and why a builder who treats the rater as a formality eventually loses a closing.

37.9.2 ⚠️ Safety alert: the residential sector has a real problem

I am going to be direct here, because softening it would be dishonest.

Residential construction's safety performance is worse than commercial's, and the reasons are structural rather than accidental. Construction as a whole consistently accounts for roughly one in five workplace fatalities in the United States, and falls are consistently the leading cause of construction fatalities. Residential framing and roofing carry disproportionate exposure within that. I am not going to give you a precise statistic, because the numbers move year to year and I will not fabricate one — go to the Bureau of Labor Statistics and OSHA for current figures. But the direction of the finding is not in dispute among people who study it.

Here is why, mechanism by mechanism:

And here is the part builders get wrong most often: OSHA's construction standards, 29 CFR Part 1926, apply to residential construction. Project size does not create an exemption. Fall protection, ladders and stairways, excavations, electrical safety, respirable crystalline silica when cutting masonry or concrete, and hazard communication all apply on a house lot exactly as they apply on a hospital tower. OSHA withdrew its long-standing residential-construction exception to conventional fall protection in 2010; the current expectation in residential construction is conventional fall protection — guardrails, safety nets, or a personal fall arrest system — at the trigger height in the standard, with a written, site-specific fall-protection plan available only where an employer can demonstrate that conventional systems are infeasible or would create a greater hazard. Confirm the current standard and OSHA's current directives before you rely on any of that; the agency's interpretations in this area have changed and can change again.

What actually works on a residential site is short, cheap, and unglamorous:

• Anchors installed at truss set, so that fall arrest is available before anybody needs it and nobody has to stop work to create it.
• Guardrails at floor and stair openings, installed by the framer as part of the framing price — written into the price list so it is not a favor.
• Ladders inspected and the right length. More residential injuries come off ladders than off roofs. A ladder that does not extend three feet above the landing is the most common violation on a house lot.
• A one-page orientation in the languages your crews actually speak, given by the superintendent, taking four minutes.
• The superintendent stopping work. Colton has done it twice this year. Both times the crew was back at it within the hour with the right setup. Both times it cost less than an hour of production and roughly nothing compared with the alternative.

Remember the scaffold near-miss at Northgate in week 34: the third finding — a crew running behind after the acceleration, under unwritten "make it up" pressure — was the one nobody wanted to write down. Everything in §37.2.2 of this chapter is an engine for producing exactly that pressure. Safety is a property of the production system, and a production line is a production system. Read Chapter 24 with a house in your head and you will see it.

37.10 Panels, Trusses, Modules — and Why Industrialized Housing Keeps Slipping

Residential construction is, quietly, the most industrialized part of the U.S. building industry — and simultaneously the field with the longest record of over-promising about industrialization. Both things are true and the reasons are interesting.

What has genuinely succeeded, to the point that it is now simply how houses are built:

Look at the pattern: components industrialized completely. The pieces of a house are almost all factory products now. The house itself, mostly, is not.

Modular and volumetric offsite housing — building whole rooms or half-houses in a factory and setting them on a foundation — has been about to transform homebuilding for roughly seventy years. It works. It is real. It produces good houses. And it has never taken over, and the reasons are boring and structural rather than technological:

• Transport. A module is limited by road width, bridge height, turning radii, and route permits. That caps module size, which caps design freedom, and it means the factory's economical delivery radius is a few hundred miles — so a factory needs enough demand inside that circle.
• Financing and appraisal. Construction lending, draw schedules, and appraisal practice were all built around site-built progress. A factory wants a large deposit up front; a construction lender wants to fund verified work in place. Reconciling those is doable and is friction, and friction at the money layer kills more good ideas than engineering ever has.
• Code variation. Modular units are typically approved through a state modular program with third-party plan review and factory inspection, and the programs differ by state. A factory serving four states may be managing four approval regimes and four sets of local amendments on the site work.
• The factory needs continuous volume. This is the one that actually decides it. A factory is a fixed-cost machine. It is cheaper than site-built only above a utilization threshold, and homebuilding demand is violently cyclical. Offsite housing plants have a long history of opening in good markets and closing in the next downturn — not because the product failed but because the volume did.
• Site work does not go away. The foundation, utilities, driveway, grading, and the "buttoning up" of the module joints are all still site work, on the same weather-exposed schedule. A 60 percent reduction in field labor is not a 60 percent reduction in cycle time.

The honest summary: offsite construction moves labor and variability from the site into a factory, and moves risk from the schedule into the supply chain and the balance sheet. For a builder with steady volume in a tight labor market, that is a good trade. For a builder whose volume swings with the market, it is often a bad one. Chapter 39 takes prefabrication, reality capture, and the rest of the technology stack properly.

Production-home software deserves one honest paragraph. The category is real — builder ERP systems that tie the sales contract, the option list, the plan-and-option-specific budget, the purchase orders, the trade scheduling, the draw requests, and the warranty log into one record. Their genuine value is not scheduling; it is automatically generating the correct purchase orders for plan 2180 with this buyer's exact option set, which is a combinatorial problem a human does badly and a database does perfectly. Their genuine limitation is that they encode your sequence, so they are excellent when your product is standard and actively obstructive when it is not. Name the tool, understand the mechanism, and do not confuse a scheduling module with a production system.

🔄 Check your understanding. A regional builder closing 90 homes a year is considering a wall-panel supplier. Which single number should dominate the decision, and why is it not the price per square foot of panel?

37.11 Crossing Over: What Each Side Gets Wrong

People move between commercial and residential constantly, in both directions, and both directions have a signature failure mode.

What transfers cleanly in both directions is short and worth knowing:

• Risk allocation. Who owns which risk and what it costs is the same question on a $47.5M hospital and a $740,000 custom house. The clauses are shorter; the question is identical.
• The schedule and the budget are the same conversation. $10,650 a day on Northgate; $29,800 a year per day of cycle time at Harbor Ridge. Different arithmetic, same discipline.
• Contemporaneous documentation. Records created at the time are worth ten times records created after the argument. This is true in a trailer and in a truck.
• You build with people. Ines Duarte's decision to split her crew and Hank Duffy's mill slot at Ironbridge Steel are the same phenomenon: a company that does not work for you making a rational decision about its own business, which you can influence only by being worth working for.

📋 Try It: The Drywall Slip at Harbor Ridge

Colton calls you on a Tuesday in February. He has been carrying a problem for six weeks and has not told anybody, which is its own lesson.

The situation. Harbor Ridge runs 11 houses in process on a 92-calendar-day cycle. The drywall stage is budgeted at 6 days per house. Since the first week of January, Coastline Drywall has been taking 9 days per house, because they split their crew between Harbor Ridge and another builder's community. Coastline runs one crew on Colton's line. Nothing else in the sequence has changed.

The economics of one house: sales price $438,000, total cost $374,500, gross margin $63,500.

Your four tasks.

(a) Compute the annual output of the line before the slip and after it. Be careful — the obvious method gives the wrong answer.

(b) Quantify the lost revenue and the lost gross margin, per year.

(c) Decide whether the constraint is the drywall crew or the builder's own readiness, and say what evidence you would go get to prove it.

(d) Choose among these three responses and price your choice. Say what you would do first.

There is also an option D that Colton's regional manager suggested: start three more houses to make up the lost volume, taking WIP from 11 to 14. Evaluate it.

Spaced Review

Three recalls. Answer each one before you read the response.

From Chapter 36: you learned that the energy code applies to every project — not only to the ones pursuing a green certification — and that envelope execution, not envelope specification, decides performance. Now say what that means for a 2,180-square-foot house on lot 19. Then read on. → It means the energy code governs this house exactly as it governs Northgate: a compliance path, an insulation and fenestration package, duct sealing requirements, and in most jurisdictions a blower door test and a duct leakage test performed by a third-party rater. And it means the test is decided on day 53, at the pre-drywall walk, not on day 90 when the rater arrives. A house insulated exactly to the specified R-value will still fail if the air barrier is discontinuous, because insulation is not an air barrier. The specification is the same on both projects. The execution is one superintendent's eye.

From Chapter 27: state the reliability chain in your own words, and what "flow" means when work moves between trades. Then read on. → Work flows only when each handoff is reliable, and reliability compounds: five sequential handoffs at 90 percent reliability each deliver about 59 percent, not 90. In this chapter you saw the same mathematics wearing different clothes. Colton's four unready houses in nine weeks were four broken handoffs. They did not merely delay four houses — they changed a trade's rate, and a rate change compounds across every house that follows. Variability in a line is not absorbed. It accumulates.

Deep callback — Chapter 14: where does a duration come from? Then read on. → Duration = quantity ÷ (crew production rate × crew size), and the critical path is a calculated result, not a management opinion. Both halves of that still hold at Harbor Ridge — the framing stage is 12 days because 2,180 square feet at a crew's known production rate takes 12 days. What changes is which question matters. On a one-off building, you compute durations, chain them, and find the longest path. In a line, durations set each trade's rate, and the governing calculation is no longer the longest path but the lowest rate — because when the same logic repeats 34 times and each crew must work continuously, the constraint is capacity, not sequence. Same input. Completely different governing question. That single shift is what makes production housing a different discipline rather than a smaller one.

Project Checkpoint: The Willow Street Residential Contrast Memo

Last chapter you built Willow Street's sustainability plan — the energy-code compliance path, the LEED scorecard estimate, and the waste-management plan. This chapter's deliverable is deliberately strange, and it is the best diagnostic in the notebook: it asks you to rebuild your own project as a different kind of business.

The deliverable: a 3-to-4-page memo titled "Willow Street as 34 Houses."

The premise. The City of Rivermont still has $6.8 million of construction money and 425 calendar days. But instead of a 24,000-square-foot two-story community center, they are building 34 single-family houses as a workforce-housing community on a larger site. Same money. Same contract time. Same you.

That is $6.8M ÷ 34 = $200,000 of construction cost per house — a real constraint, and the first number in your memo.

Cover these seven items. Be specific; a memo full of adjectives is worth nothing.

1. Cycle time and rate. Pick a target cycle time and a WIP, and defend them with arithmetic. The relationship you must satisfy is total duration = C + (34 − 1) × (C ÷ WIP) ≤ 425 CD. Show that your pair fits, and state your start interval and annual output. (Worked example to check your method: C = 100 CD with WIP = 11 gives an interval of 9.09 days and a total of 100 + 33 × 9.09 = 400 CD — it fits, with 25 days to spare.)
2. Scheduling technique. State that you are replacing CPM with line of balance and say why in two sentences — the rate-versus-path argument from §37.3.1. Then sketch the LOB chart with five trade stages, by hand, on paper.
3. The submittal log and the RFI process. What happens to them? Be honest: what genuinely disappears, what you keep anyway, and — the graded part — what you put in their place to protect yourself. A one-page-per-community revision log is the minimum defensible answer.
4. Trade contracts and payment. Convert at least five of your Willow Street trades from lump-sum subcontracts to unit pricing. Give the unit for each (per SF of board, per opening, per square, per house) and say how and how often you would pay. Note what happens to retention and to your lien-waiver process.
5. Quality and inspections. Write your stage-walk checklist — the five to seven walks and what you look at in each. Then list the municipal inspection sequence you would expect and identify which two inspections are most likely to cost you cycle time in your jurisdiction.
6. Keep three, abandon three. From your Willow Street notebook so far — the risk register, the CPM schedule, the site logistics plan, the buyout log, the submittal log, the daily reports, the cost codes, the quality plan, the safety plan, the look-ahead — name three practices you would keep and three you would abandon, with one sentence of reasoning each. Do not hedge. Commit.
7. One paragraph on what you would get wrong. You are a commercial manager. Read §37.11 and name your own most likely failure mode honestly.

Next chapter you write the other contrast memo — Willow Street's site utility and paving scope run as a unit-price heavy-civil contract — and you will find that these two memos are opposites. Residential removes contract machinery because repetition replaces it. Heavy civil adds contract machinery, because when you are paid by the measured cubic yard, measurement itself becomes the contract.

Chapter Summary

A reference card, not a recap.

The one sentence: Residential construction is not smaller commercial construction — it is a different business whose unit of production is the house, whose constraint is cycle time rather than duration, and whose scheduling question is "which trade's rate is lowest," not "which path is longest."

The arithmetic you must be able to do cold

The five segments and their governing risk

Where the money went in this chapter

The nine rules

1. Finishing houses is production. Starting houses is not. Raising WIP without raising capacity raises cycle time and inventory, and nothing else.
2. The slowest trade sets the rate of the entire line. Not the average, not the total.
3. A trade that falls below the line rate stops being a duration and becomes a ceiling. The naive cycle-time arithmetic will understate the loss by half or more.
4. A queue is visible on the ground. Several houses sitting at the same stage is your constraint, drawn to scale.
5. Be the builder whose houses are ready when the crew shows up. In production housing that is worth more than price and more than any contract you could write.
6. Repetition substitutes for planning; standardization substitutes for coordination. Which means the first house of a new plan, a plan revision, and any deviation from the standard have no reserve behind them. Treat the first house as a prototype and keep a one-page revision log per community.
7. Options are a fifth of the margin. So the selections cutoff is a schedule gate with a dollar sign on it, not a customer-service function.
8. A deadline with no stated consequence is a suggestion. Put the consequence in the selections schedule, in writing, at signing.
9. OSHA applies to a house. Project size creates no exemption, piece-rate pay is an incentive working against you, and everything in §37.2.2 is a machine for producing schedule pressure.

What is still true in both worlds: risk allocation drives price · the schedule and the budget are the same conversation · contemporaneous records beat reconstructed ones · you build with people who do not work for you.

And the standing caution: residential lien law, homeowner-protection statutes, notice and disclosure requirements, warranty periods, right-to-cure procedures, contractor licensing, and inspection sequences vary substantially by state and locality and change over time. Nothing in this chapter is legal advice. Read your state's statutes and your own contract, and ask a lawyer who practices where you build.

What's Next

Chapter 38 takes you to the other end of the spectrum: Cottonwood Creek, an $18.7 million bridge replacement for a state DOT on a unit-price contract with 210 working days. Where residential strips contract machinery away because repetition replaces it, heavy civil piles it on — because when you are paid $38.50 per cubic yard of structural excavation, the measurement becomes the contract, and a differing site condition is not an inconvenience but the central commercial event of the job. Then Chapter 39 picks up the prefabrication thread from §37.10 and asks what the next decade actually changes, and what it only promises to.