Chapter 4 — Construction Contracts: Lump Sum, GMP, Cost-Plus, Unit Price, and What Each Means for Risk

The Hook: "So this is the most we'll pay"

The conference room at Meridian Health System looks out on the surface parking lot where the Northgate Outpatient Pavilion is going to stand. Eleven months of preconstruction is sitting on the table in a three-ring binder four inches thick. Today we convert it into a contract.

Around the table: Priyanka Sethi — Pri — Meridian's owner's representative. Meridian's outside construction counsel, who has read every word and flagged forty-one of them. Our chief estimator, Tomás Reyes, with the GMP build-up printed at eight-point type because he refuses to round anything. And me.

We spent the morning walking the number. Direct cost of work, forty million. General conditions, two-nine. Insurance and bonds, nine hundred thousand. Construction contingency, one-three-two-zero. Fee at four percent. Escalation allowance. Bottom line: $47,500,000, and 565 calendar days from notice to proceed on March 3 to substantial completion on September 18 of the following year.

Pri put her pen down, looked at the last line, and said the sentence I have heard at every GMP conversion I have ever sat through.

"Okay. So this is the most we'll pay."

I could have said yes. It would have been easy, it would have felt collaborative, and it would have been the most expensive kindness of my career — because eighteen months later, when Meridian's imaging vendor selected a different MRI unit and the depressed slab had to get deeper, Pri would have been in that same chair holding a change order and a memory of me saying yes.

So I said: "It's the most we'll charge you for the work drawn in that binder. It is not the most you will spend."

Meridian's counsel wrote something down. Pri said, evenly, "Then what exactly is guaranteed?"

That is the best question anybody asked all year, and I want you to be able to answer it before you finish this chapter.

What is guaranteed is Kestrel's exposure. If the work in that binder costs us $49,200,000 to build, we eat $1,700,000. That is a real, enforceable, company-threatening promise, and it is why the fee is four percent and not two. What is not guaranteed is Meridian's total spend, because Meridian keeps the right to change its mind — about the MRI, the lobby finishes, whether the third-floor procedure rooms need medical gas — and every one of those decisions moves the number in exactly one direction.

I told Pri: "This document caps my downside on a defined scope. Your budget is capped by your own discipline. Those are two different sentences and they are both true."

She did not love it. She signed it anyway, because it was honest, and because I put a written change-management protocol in front of her in the same meeting — five pages, a decision log, and a rule that no cost above $25,000 gets committed without a signed directive. That protocol is why Meridian's final number was 4.4% above the guarantee instead of 15%.

Every construction contract is a machine for deciding who owns which risk and what that costs. The price is just the output.

🏃 Fast Track: If you already price work for a living, skim §4.1 and §4.3, then read §4.2 (the threshold concept), §4.5 (cost of the work vs. general conditions vs. fee — this is where money actually gets argued about), and §4.7 (time provisions and no-damage-for-delay) closely. The comparison tables in §4.4 are worth photographing.

🔬 Deep Dive: Appendix G decodes the twenty-five clauses that decide who pays, one at a time. Chapter 5 takes the legal framework — liens, bonds, insurance, disputes — much further. Chapter 6 turns contingency from a number into a managed reserve. Chapter 31 is where the clauses in §4.8 get used in anger.

4.1 What a construction contract actually is

A construction contract is a risk-allocation instrument with a price attached.

Most people have that backwards. They think a contract is a price with legal language wrapped around it to protect the price. It is the other way around: the legal language decides who absorbs the unknowns, and the price is what the market charges to absorb them. Change the clauses and the price changes, every time, whether or not anyone says so out loud.

Run the experiment yourself. Take the same drawings to the same three contractors twice. The first time, include a differing site conditions clause saying the owner pays for unforeseen subsurface conditions. The second time, delete it and add a sentence saying the contractor has satisfied itself as to all site conditions. Same building, same schedule, same subs. The second set of bids comes in higher — often by one to three percent of contract value on a job with real subsurface uncertainty — because you just handed the contractor a risk it has to price.

Risk does not disappear when you assign it. It gets priced.

The contract documents are a set, not a document

When people say "the contract," they mean the agreement — the ten or twenty pages with the price and the signatures. That is the smallest and least interesting part of what governs your job.

Those seven pieces are the contract documents. They are read together, and when they conflict — and they always conflict — there is an order of precedence that decides which one wins. Chapter 7 is entirely about reading this set and finding the conflicts before they find you.

The industry standard forms, by function

You will meet a handful of published contract families. Do not memorize clause numbers. Learn what each one is for.

I describe these by function on purpose. I will not quote clause text at you, and you should be suspicious of anyone who quotes it from memory. Read your actual contract. The form number tells you the starting point; the supplementary conditions tell you where the owner moved it.

🔄 Check your understanding. An owner deletes the differing site conditions clause from the general conditions and adds language saying the contractor accepts all subsurface risk. The owner expects to save money because "we're not paying for surprises anymore." What actually happens to the bids, and why?

4.2 🚪 What a "guaranteed" maximum price actually guarantees

🚪 Threshold concept. A guaranteed maximum price guarantees the contractor's exposure, not the owner's cost. Scope changes still move the number — and most owners learn this the expensive way.

This is the gateway idea of the chapter. Getting it wrong poisons the owner relationship for the entire job.

Before you understand this, a GMP reads like a ceiling on the owner's spending. The owner budgets $47,500,000, signs a document with the words guaranteed maximum price printed in capital letters, and files it. When change orders arrive, the owner feels defrauded. The conversation becomes "You guaranteed this," and the contractor becomes the villain in a story where nobody did anything wrong.

After you understand this, a GMP reads as a promise about one side of the ledger only: for the scope defined in these documents on this date, Kestrel will not charge Meridian more than $47,500,000, and if it costs Kestrel more, Kestrel pays the difference. That is an enormous, balance-sheet-threatening promise. It is also silent about everything that changes the scope.

Why the asymmetry exists

The GMP caps the price of a defined scope. Scope is defined by documents. Documents get changed by exactly one mechanism — the change order — and the change order is the tool the owner uses when it wants something different.

So ask: when does an owner issue a change order? Almost never to remove something it already paid for and wants. Owners issue changes to add a procedure room, upgrade the imaging suite, accept a condition the geotechnical report missed, fix a design coordination error, satisfy a building official's interpretation. Every one adds cost. The change mechanism is structurally asymmetric — a one-way valve pointed up — and there is no corresponding clause that says "if you change your mind less than expected, you get a rebate."

🔍 Why this works. A contractor can only guarantee what it can control and what it can count. It can count the drawings. It cannot count the owner's future decisions, the geology under the north property line, or the architect's coordination between the structural and mechanical models. If a contractor did guarantee those, it would price them — and the price would be so high that the owner would never take the deal. The GMP exists precisely because both parties agree to carve the uncontrollable things out of the guarantee. That carve-out is not a loophole. It is the whole reason a guarantee is affordable.

What Northgate actually cost

Here is Northgate's change-order register rolled up at final completion. This is what happened to a guarantee that everybody negotiated in good faith, on a well-run job, with a competent owner and a written change-management protocol.

Final contract sum: $47,500,000 + $2,094,200 = $49,594,200. That is 4.4% above the guarantee.

💰 Money check. Meridian's total project budget was $61,000,000, of which $47,500,000 was construction. The $2,094,200 in changes consumed a large share of Meridian's own owner contingency — the money it held outside the construction contract for exactly this. Meridian did not go over budget. It went over the guarantee, and it had planned for that, because Pri carried roughly five percent of construction value as owner contingency. The owner who carries none is the owner value-engineering the lobby in month fourteen.

Notice something else. Only $1,284,000 of the $2,094,200 — about 61% — was Meridian changing its mind. The rest was design errors, site conditions, and a code official's interpretation. None of those is the owner's fault in any moral sense. All of them are the owner's risk under a standard GMP, because the contractor did not draw the building, did not write the geotechnical report, and does not employ the building official.

⚖️ What the contract says. Under the standard forms, the GMP is adjusted for: owner-directed changes; changes required by a differing site condition covered by the clause; changes required by a change in law or code interpretation after the contract date; owner-caused delay where the contract allows time-related compensation; and reconciliation of allowances against actual cost. It is not adjusted for: the contractor's own estimating errors, its own productivity misses, its own subcontractor defaults, or market escalation beyond whatever the escalation provision covers. Those are the guarantee.

The one sentence to say out loud in the GMP meeting

I now say a version of this in every conversion meeting, and I say it before anyone asks:

"This number is a cap on what we can charge you for the work in these documents. Three things will move it: you changing your mind, the documents being wrong, and the ground being different than the report said. We will manage all three with you in writing, but I am not going to pretend the number can't move. What I am guaranteeing is that if we build this scope and it costs us more than this, that is our problem, not yours."

Owners do not resent that sentence. They resent finding out in month fourteen.

🔄 Check your understanding. Kestrel's cost to build the defined Northgate scope comes in at $48,900,000 — nine hundred thousand dollars over the $47,500,000 GMP, with zero change orders. What does Meridian pay, and what does Kestrel earn?

4.3 One scope, four prices: the same $2 million package, priced four ways

Abstractions about risk transfer become obvious the moment you put numbers on them. So let's price a single, real package four different ways and watch what happens.

The package: the Northgate site development scope — mass excavation and grading, aggregate base, storm drainage, curb and gutter, and asphalt paving. Roughly $2 million of work, sixteen weeks of field duration.

The contractor's honest internal cost view, before any pricing structure is chosen:

That $1,917,000 is what the work is expected to cost. It is not a price. A price requires a decision about who owns the gap between "expected" and "actual."

4.3.1 Lump sum (stipulated sum)

The contractor names one number for the whole defined scope and delivers it for that number, whatever it costs.

Notice what the owner just bought: certainty, for a premium of $196,300 over expected cost. The contingency is invisible — no line item, no audit right, no refund if the work goes well. If the crew beats production and quantities come in light, the contractor keeps every dollar. That is a fair deal; the contractor took a real risk to make that money.

Where lump sum goes wrong. Two temptations show up immediately, both covered in §4.9: unbalanced bidding (moving money between line items to exploit what you believe the real quantities will be) and front-loading the schedule of values (inflating early activities so your billings run ahead of your costs).

The third thing that goes wrong is nobody's fault: the drawings are wrong. If the plans show a 5-inch slab and the structural notes require 6 inches, the contractor priced 5 and owes 6 — and whether the owner pays for the extra inch turns entirely on the order of precedence and the notice clause. That fight is Chapter 7 and Chapter 31.

4.3.2 Guaranteed maximum price

Cost of the work plus a fee, capped. The books are open. The owner can audit.

The GMP is $59,810 lower than the lump sum for identical scope. Not magic — it is the difference between a disclosed 3% contingency plus a 4% fee and a hidden 4% contingency plus a 6% markup. The owner bought a smaller premium in exchange for administrative burden (somebody at Meridian now has to audit an open book) and in exchange for accepting that the cap covers only the defined scope.

Here is the Northgate build-up itself, line by line, because you should be able to read one of these cold:

At 132,000 gross square feet that is $360 per square foot — $47,500,000 ÷ 132,000 SF. Every owner's rep in the country will benchmark you against that number in the first meeting.

The contingency question: whose money is it? This is the fight in every GMP negotiation. The $1,320,000 sits inside the GMP, so the owner has already agreed to pay it. But if it goes unspent, Northgate's savings split returns 75% to Meridian and 25% to Kestrel.

💰 Money check — what the contingency really costs the owner. Suppose the whole $1,320,000 goes unused. Meridian gets $990,000 back; Kestrel keeps $330,000. The owner's net cost of carrying that contingency, in the best case, is $330,000 — about seven-tenths of one percent of the GMP, and that is the price of the contractor being willing to sign a cap at all. Under the lump-sum alternative the equivalent contingency is invisible and the refund is zero. The savings split is the single most valuable thing an owner gets from an open-book GMP, and owners routinely trade it away for a slightly lower fee. That is usually a bad deal.

What is inside the guarantee, and what is not. Say this out loud in the meeting.

4.3.3 Cost-plus

The owner reimburses the actual cost of the work and pays a fee. There is no cap. Three flavors, and they are not equivalent.

Our package, cost-plus fixed fee: estimated cost $1,917,000 + fixed fee $95,850 = $2,012,850 estimated. The cheapest of the three prices — and the only one that is not a price at all. It is a forecast. If the work costs $2,300,000, the owner pays $2,300,000 + $95,850.

Why the percentage fee is genuinely dangerous. At 5%, if cost runs from $1,917,000 to $2,127,000, the fee goes from $95,850 to $106,350: the contractor earned $10,500 for the overrun. Nobody consciously decides to waste the owner's money. What happens is a thousand small decisions where the cheaper option requires a hard conversation and the expensive one does not, and the contractor's own economics never push back. Case Study 02 walks a real one, month by month.

How to control cost-plus. If you are the owner and must use it, you need all four of these in writing before the first invoice:

1. A fee in dollars, not a percentage. Kill the incentive at the root.
2. A written definition of the cost of the work — reimbursable and non-reimbursable schedules, an agreed labor burden rate, an agreed equipment rate schedule, and audit rights with a stated record-retention period.
3. A control budget by cost code with a monthly cost-to-complete forecast, not just cost-to-date. See Chapter 28.
4. A written authorization threshold — no cost above a stated amount committed outside the control budget without the owner's signature. On a small job, $10,000; on Northgate, $25,000.

And the fifth, which is really the first: convert to a GMP as soon as the design supports it, usually around 70% construction documents. Cost-plus should be a bridge, not a destination.

4.3.4 Unit price

The owner and contractor agree on a price per unit of measured work. The owner supplies estimated quantities for bidding. Payment is for actual measured quantities at the bid unit prices.

Our package, priced as unit price:

Those unit prices are direct costs loaded with general conditions, insurance and bond, a 2% productivity contingency (not 4% — the contractor no longer carries quantity risk), and 7% overhead and profit (higher than the lump-sum markup, because unit-price work carries measurement disputes and mobilization risk). The bid lands about $19,500 below the lump sum. That is the market pricing the transfer of quantity risk back to the owner.

How unit price actually gets administered — this is where new project engineers get eaten alive:

• Somebody has to measure. On a DOT job the owner's inspector measures and your engineer measures, and you both sign a joint quantity sheet daily or weekly. If you are not in the field with a wheel, a rod, and a truck-ticket log, you are accepting the owner's number.
• The contract has a variation clause. If actual quantity on a major item varies from the estimate by more than a stated percentage — 15% and 25% are both common — either party can request renegotiation of that unit price, because fixed mobilization and setup costs were spread over the estimated quantity. Thresholds vary by agency; check yours.
• Truck tickets, weight tickets, and survey records are your money. Lose them and you cannot bill. Theme 5 with a dollar sign on it.

4.3.5 Target price and shared savings, briefly

Target price / target value delivery. The team sets a target cost during design and then designs to it rather than designing and then pricing. Overruns and underruns are shared on an agreed band — say, the first 3% of savings 50/50, everything beyond that 80% owner. It works when owner, designer, and contractor are in the room early; it fails when the "target" is really an owner's budget wearing a new name.

Incentive and shared-savings arrangements bolt onto any base structure. The design questions: what is the baseline, who validates it, is the sharing symmetric, is the contractor's upside capped, and — the one everyone forgets — does an incentive on cost create a disincentive on quality or safety? An early-finish bonus with no corresponding safety metric is a schedule-pressure machine. See §4.7 and Chapter 24.

4.4 Who owns which risk: the comparison that matters

Here is the whole chapter in two tables. If you photograph anything, photograph these.

📊 Diagram (described). The five pricing structures arranged on a risk-transfer spectrum. At the left end, the contractor absorbs nearly all execution risk and charges the most to do it. At the right end, the owner absorbs nearly all of it and pays actual cost plus a fee. The vertical dimension is the owner's administrative burden, which rises as you move right — the owner who transfers the least risk must do the most work.

  CONTRACTOR OWNS MORE RISK  <────────────────────────>  OWNER OWNS MORE RISK

   LUMP SUM        UNIT PRICE        GMP        COST-PLUS      COST-PLUS
  (stipulated)   (fixed units,    (cost + fee,  (fixed fee)   (% of cost)
                   open qty)        capped)
       │               │               │             │              │
   Highest          High            Medium          Low          Lowest
  contingency    contingency     contingency +   contingency   contingency
   in the bid     in the bid         fee          + fee          + fee
       │               │               │             │              │
   Owner needs    Owner needs     Owner needs   Owner needs    Owner needs
    complete       accurate        audit staff   tight scope     a miracle
    documents     quantities      + discipline     control
       │               │               │             │              │
   ────────────── OWNER'S ADMINISTRATIVE BURDEN RISES ──────────────►

Table 1 — Who owns each risk?

Table 2 — What each structure demands and rewards.

CHOOSING A PRICING STRUCTURE — start with the biggest unknown

  What is the biggest unknown on this project?
  │
  ├─ The DESIGN is unknown (scope not fixed)
  │    ├─ Need a price cap before design finishes ──────► GMP (CM at Risk)
  │    ├─ No time to define scope at all ───────────────► COST-PLUS, FIXED FEE
  │    │                                                  → convert to GMP at ~70% CD
  │    └─ Emergency / disaster response ────────────────► COST-PLUS + audit rights
  │
  ├─ The QUANTITIES are unknown, but the work is measurable
  │    └─ Earthwork, utilities, paving, drilled shafts ─► UNIT PRICE
  │
  └─ Nothing important is unknown (complete documents)
       ├─ Competitive market, public owner ────────────► LUMP SUM (hard bid)
       └─ Private owner, speed, relationship ──────────► NEGOTIATED LUMP SUM

Remember what Chapter 3 established: delivery method and contract type are one decision. You do not pick CM at Risk and then separately pick a GMP. CM at Risk is the two-phase arrangement that makes a GMP possible before the drawings are done. Design-bid-build on a public job is the lump-sum award to the low responsive bidder. Choosing the delivery method chose the pricing structure, and both of them chose who absorbs the unknown.

🔄 Check your understanding. A university is building a $30 million research laboratory. The program is 90% settled but the lab equipment vendors will not be selected for another eight months, and the equipment drives the mechanical, electrical, and structural design. The university has a two-person facilities team and no construction staff. Which structure, and what is the one thing they must add to make it work?

4.5 Cost of the work, general conditions, and the fee

This is the most misunderstood distinction in construction contracting and the source of more owner–contractor arguments than change orders. On a cost-reimbursable contract — GMP or cost-plus — every dollar the contractor spends lands in one of three buckets, and the bucket decides who pays.

The rule of thumb: if it would exist whether or not this project existed, it is fee. If it exists only because of this project, it is cost of the work or general conditions.

Northgate's $2,900,000 in general conditions, itemized

Owners always ask what general conditions buy. Here is the answer for a 565-calendar-day, $47.5 million medical building.

Two things jump off that table.

First: half your general conditions is people. The nine staffing lines total $1,446,950 — 49.9% of the GC budget. When an owner asks you to cut general conditions, it is asking you to take someone off the job. Sometimes that is right. Usually the person they want to cut is the safety manager or the project engineer — the two whose absence costs the most and shows up latest.

Second: the daily burn rate.

💰 Money check — deriving the $5,150/CD rate.

General conditions ÷ contract time = daily general-conditions cost $2,900,000 ÷ 565 CD = $5,132.74 per calendar day

Northgate's contract sets the agreed extended general-conditions rate at $5,150 per calendar day — the computed number, rounded. What it means: every calendar day the job runs past substantial completion costs Kestrel $5,150 in staff, trailer, temp utilities, hoist, and cleanup that nobody is paying for. Add the liquidated damages of $5,500 per calendar day and you get the number that should be tattooed on your forearm: $10,650 per calendar day.

⚖️ What the contract says. There is a real negotiation buried in that rate. Not all of the $2,900,000 is time-related — final cleaning, small tools, mobilization, and permits are production-related or fixed, and do not cost more because the job runs twenty days longer. Meridian's counsel ran that analysis and argued the honest extended-GC rate was closer to $3,560/CD. Kestrel argued for $5,150.

They settled at $5,150/CD, and Kestrel gave something real for it: the rate became the exclusive remedy for time-related overhead on compensable delay. No home-office overhead formula on top, no separate unabsorbed-overhead claim, no lost-productivity claim folded in. Meridian bought certainty and closed off a whole category of claim; Kestrel bought a generous, undisputed daily number it never has to prove. Both sides got something. That is what a good negotiation looks like — and why you should read Appendix G before you sit down for one.

The gray zone — the items that actually get argued about

🧩 Productive struggle. Five minutes on this before you read on. Meridian's auditor reviews Kestrel's month-nine billing and refuses to reimburse four items. Which is the auditor right about, and what is Kestrel's argument on the others?

1. $14,200 — the assistant superintendent's overtime during the three weeks of accelerated steel erection in Year 1.
2. $6,800 — a Kestrel estimator's time pricing owner change orders 7 through 11.
3. $22,400 — re-pouring a section of the level-2 deck where Kestrel's self-perform crew set the embeds four inches off.
4. $9,100 — Nadia Haddad's monthly executive review visits to the site.

4.6 The four escape valves inside a fixed price — and escalation

Every "fixed" price has doors in it. There are four, and each is a legitimate tool that gets abused about as often as it gets used well.

Allowances

An allowance is a stated dollar amount carried in the contract for scope known to exist but not yet defined. At the end it is reconciled against actual cost — the contract sum moves up or down by the difference.

Northgate carried one that matters. During the GMP negotiation Meridian wanted all subsurface risk inside the guarantee. The geotechnical report had fourteen borings across 6.2 sloping acres, and the north property line abuts an active clinic with utilities nobody has a good record of. Kestrel would not guarantee that. The compromise:

Unsuitable soils allowance: $285,000, covering up to 3,800 CY of undercut, export, and replacement with imported structural fill, at an agreed unit price of $75.00/CY. Quantities beyond 3,800 CY are a change to the contract sum at the same unit price.

Check the arithmetic: 3,800 CY × $75.00/CY = $285,000. Both parties know exactly what happens at 3,801 cubic yards. That is a good allowance.

How allowances get abused. By owners, as a place to hide undesigned scope — six allowances totaling $1.8 million on a $12 million job is not a strategy, it is an admission the design is not done. By contractors, by setting the allowance artificially low to make the bid look competitive, knowing reconciliation will make it up later. And by both, by failing to state whether the allowance includes overhead, profit, and general conditions. If it does not, reconciliation adds markup and the owner is surprised. One sentence prevents that argument.

Alternates

An alternate is a priced option the owner may accept or decline, usually at award. An add alternate increases the base scope; a deduct alternate removes something.

Alternates are how an owner with a firm budget and an uncertain program buys optionality. They are also how a public owner accidentally changes who the low bidder is — accept alternates 1 and 3 and a different contractor wins. Most public procurement rules require the owner to state in the bid documents the order in which alternates will be considered, precisely to prevent that. Rules vary by jurisdiction; read the instructions to bidders.

The abuse: pricing a deduct alternate high so the owner never takes it, or pricing an add alternate low to win the award and making it up in the change orders that follow.

Unit prices inside a lump-sum contract

Even a hard-bid lump-sum contract usually carries unit prices for items with genuinely uncertain quantities — rock excavation, unsuitable soil replacement, additional piling, dewatering. The lump sum covers the drawn quantity; the unit price covers variation. Check two things every time: does the unit price include overhead and profit, and does it apply symmetrically to adds and deducts? Most contractors want a higher rate for additions than deletions, which is defensible on mobilization grounds and infuriating to owners. Settle it up front, in daylight, not in month eleven.

Contingency

A contingency is a reserve for identified risk within the defined scope. Three different contingencies live on a typical job, and confusing them is a career-limiting move:

The single most common owner mistake in the industry: assuming the GMP contingency covers scope changes. It does not. That is what the owner's contingency is for, and the owner who carried none is about to have a difficult board meeting. Chapter 6 turns contingency from a percentage into a priced, drawn-down register tied to named risks.

Escalation and commodity price risk

In a stable market nobody thinks about escalation. In a volatile one it is the difference between a good year and a bad one. Northgate carried a $575,200 escalation allowance. Where did it come from?

💰 Money check. Kestrel identified the commodity-exposed portion of the work — the scope where raw-material price movement passes straight through to the subcontract price:

$575,200 ÷ $14,200,000 = 4.05%

What it means: Kestrel carried about four percent escalation on the fourteen million dollars of work whose price moves with commodity markets — not on the whole $47.5 million, which would have been $1.9 million and would have gotten the GMP rejected. Escalation is priced on the exposure, not on the contract.

How escalation clauses are structured — three common shapes, and they are not equal:

Price-hold periods are the practical mechanism underneath all of this. When a supplier quotes, ask how long the price holds. Mill quotes commonly hold for a stated number of days and then float; some commodity items hold for a week. Your entire escalation exposure is the gap between signing the contract and locking the material price — which is why early buyout and early release for fabrication is a risk management activity, not just a procurement activity. See Chapter 16.

🏗️ From the field. I lost $71,000 on a job once because I let a curtain-wall quote go stale. The sub's price held for thirty days, I took forty-eight days to issue the subcontract because I was arguing about a scope gap in the sill flashing, and the aluminum extrusion price moved in between. The scope gap I was arguing about was worth $9,000. I won that argument and lost eight times the money. Close the deal, then argue about the details in a change to the subcontract.

🔄 Check your understanding. A contract carries a $150,000 allowance for landscaping and irrigation and says nothing about markup. The actual landscaping subcontract comes in at $206,000. What is the adjustment to the contract sum, and what is the argument?

4.7 Time is a contract term, not a goal

Everything so far has been about money. Now the other half: the schedule and the budget are the same conversation, and the contract is where the schedule acquires a dollar sign.

Contract time, substantial completion, final completion

Contract time is a duration, usually in calendar days, running from a defined start — notice to proceed, or the date in the agreement. Northgate: 565 calendar days from NTP on March 3, Year 1, producing a contract substantial completion date of September 18, Year 2.

Substantial completion is the milestone that matters. It is not "we're done." It is the point at which the owner can occupy and use the work for its intended purpose — on a healthcare project, generally a certificate of occupancy plus whatever the health authority requires. At substantial completion, liquidated damages stop accruing; risk of loss and responsibility for utilities, security, and maintenance generally shift to the owner; retention is typically reduced and the warranty period starts; and the punch list is issued, defining what is left.

Final completion is everything after: punch list closed, closeout documents delivered, O&M manuals, warranties, as-builts, training, final lien waivers, remaining retention released. Northgate: certificate of occupancy September 24, Year 2; final completion November 17, Year 2 — sixty days of closeout after substantial completion, which is realistic and which surprises every first-time owner. Chapter 40 lives in those sixty days.

Liquidated damages, and why they are not a penalty

Liquidated damages are a stated daily amount the contractor owes for each day of unexcused delay past substantial completion. Northgate: $5,500 per calendar day.

The critical legal idea — genuinely important, not lawyer trivia — is that liquidated damages must be a reasonable pre-estimate of the owner's actual loss, made at the time of contracting, when actual damages would be hard to determine precisely. A provision that operates as a punishment rather than an estimate of loss is, in most U.S. jurisdictions, unenforceable, and the specific test varies by state. That is why a well-drafted LD provision traces back to a calculation. Here is Meridian's:

⚖️ What the contract says. Two things about LDs that new PMs get wrong. First, they are usually the owner's exclusive remedy for delay damages — which protects the contractor, because it caps an otherwise open-ended exposure. Second, an LD provision without a corresponding time-extension mechanism is a trap, and courts in many jurisdictions will not enforce LDs against a contractor that the owner itself delayed. This is the "prevention doctrine," it varies by state, and it is the reason the changes clause and the LD clause have to be read together.

💰 Money check — the number behind every schedule decision on Northgate.

Extended general conditions: $5,150/CD Liquidated damages: $5,500/CD Total daily exposure to slipping substantial completion: $10,650/CD

When the anchor-bolt submittal delay pushed steel erection from August 4 to August 27, Year 1 — twenty-three calendar days on the critical path — the arithmetic was:

23 CD × $10,650/CD = $244,950

What it means: an eleven-day delay in one office, on one submittal, plus a fourteen-day contractual engineering review, cost a quarter of a million dollars. Nobody was negligent. Nobody was lazy. A piece of paper sat on a desk.

Early completion bonuses, and their trap

Some contracts pay a bonus for early substantial completion, often at the same daily rate as the LDs. Symmetry sounds fair, and often is.

The trap is the early-completion schedule. A contractor submits a baseline showing substantial completion forty days ahead of the contract date, suffers a twenty-day owner-caused delay, finishes exactly on the contract date, and then claims twenty days of compensable delay and a lost bonus — arguing it was entitled to its early finish. Whether that claim survives turns on the contract's language about float ownership and baseline approval, and it is one of the most contested questions in construction scheduling. See Chapter 14 on float ownership and Chapter 33 on how these claims get analyzed.

No-damage-for-delay: the clause that changes your risk the most

If you read exactly one clause in your contract before you sign it, read this one.

A no-damage-for-delay clause says: if you are delayed, even by the owner, your sole remedy is an extension of time. No extended general conditions. No unabsorbed overhead. No compensation. Just more days.

Do the arithmetic on Northgate. A thirty-day owner-caused delay:

Same delay. Same fault. A $154,500 difference, decided by one paragraph in the supplementary conditions that you either read or did not.

Courts in many U.S. jurisdictions recognize exceptions — commonly for delay caused by the owner's active interference or bad faith, delay not within the contemplation of the parties, abandonment of the contract, or fraud. Several states limit or prohibit the clause by statute, particularly on public work. Enforceability, exceptions, and statutory limits vary significantly by state and change over time. Do not rely on the exceptions. Price the clause, or negotiate it out, or walk.

⚠️ Safety alert. Contract terms are a safety input, not just a money input. Stack an aggressive contract time, $5,500/day in liquidated damages, a no-damage-for-delay clause, and no float ownership, and you have built a machine that converts every schedule slip into pressure on a crew that had nothing to do with it. Remember the scaffold near-miss on the north elevation in week 34 — a mason tender stepping onto an unsecured plank at 7:20 in the morning. The investigation found three failures, and the third was a crew running behind after the steel acceleration, with an unwritten "make it up" pressure. That pressure did not originate on the scaffold. It originated in a contract, travelled through a schedule, and arrived as an unsecured plank. Chapter 24 treats safety as a property of the production system. The contract is part of that system.

4.8 The clauses that decide who pays: a working tour

Here is a plain-English tour of the clauses that will actually decide arguments on your job. Each one gets a full treatment in Appendix G; this is the field version.

🔍 Why this works. Notice requirements feel like bureaucratic cruelty until you see what they are for. A 7-day notice is not there to trap you. It is there to let the owner mitigate — redirect the crew, expedite a decision, change the sequence — while the problem is still cheap. A claim delivered eight months later, when the concrete is poured and mitigation is impossible, asks the owner to pay for a decision it was never given the chance to make. That is why courts enforce these clauses. Understanding the purpose is also how you argue your way out when you missed the deadline: show the owner had actual knowledge and was not prejudiced. It is a real argument. It is a much worse argument than having noticed on time.

4.9 The line: aggressive versus fraudulent

Two techniques in this chapter can be used well or criminally, and the difference is not obvious from the outside. See the line clearly, because the people who cross it usually did not decide to — they drifted.

Unbalanced bidding

On a unit-price contract you bid a price for each item, and you are allowed to have a view about which of the owner's estimated quantities are wrong. Everybody does. Watch what a view is worth.

This is the Cottonwood Creek Bridge Replacement — $18.7 million of heavy civil for a state DOT, unit price, 210 working days. Four of its major items.

The balanced bid — unit prices reflecting the contractor's actual cost view:

Now: project engineer Ingrid Sørensen walks the site, reads the borings, and concludes two things. The DOT's structural excavation quantity is overstated — the existing abutment footings are shallower than the plans suggest, and actual will run closer to 11,100 CY. And the drilled shaft quantity is understated — the rock line is deeper on the east bank, so shafts will run closer to 3,910 LF.

The unbalanced bid — same total, money moved:

The unbalanced bid is $1,020 lower. On a bid tab, it wins. Nobody notices.

Now pay both bids at actual quantities — excavation 11,100 CY, concrete 2,880 CY, rebar 494,000 LB, shaft 3,910 LF:

$4,810,920 − $4,712,550 = $98,370

What it means: the unbalanced bidder submitted a bid that was $1,020 lower and collected $98,370 more. That is the entire mechanism, and it is why every DOT in the country reviews unit prices for material unbalancing.

Where the line is. Three positions on one spectrum:

1. Legitimate. Your unit prices reflect your genuine cost and risk view on each item — your judgment about mobilization, sequencing, and where your production risk sits. Your drilled-shaft number is high because your cost is high. You can defend every price with a takeoff and a production rate.
2. Aggressive, and generally treated as unacceptable. Prices deliberately distorted away from your cost to exploit quantities you believe are wrong. Most public agencies reserve the right to reject a materially unbalanced bid — one where the distortion creates reasonable doubt that the agency will pay the lowest ultimate cost. Rejection costs you the job and your standing with that agency, which is worth more than one job.
3. Fraud. You have actual knowledge the quantity is wrong — from a prior contract, from information the owner does not have, from a conversation you should not have had — and you price to exploit it while representing your bid as responsive. That is misrepresentation, and on a federally funded project, false-claims territory: debarment, civil liability, and in serious cases criminal exposure. These rules vary by jurisdiction and funding source. The point is that this is a difference in category, not in degree.

The practical test I use: could I explain this unit price to the agency's chief engineer, out loud, with my takeoff in front of me, and have it sound like arithmetic instead of a scheme? If the honest answer is no, do not submit it.

Front-loading the schedule of values

The schedule of values (SOV) is the breakdown of your contract sum into line items, used to measure progress and generate monthly payment applications. Front-loading means assigning inflated value to early activities so your billings run ahead of your actual costs.

The total is identical. Every number is defensible in isolation. And by the end of foundations, the front-loaded contractor has billed roughly $400,000 more than it has earned — an interest-free loan from the owner.

Why people do it: cash flow. Construction is a working-capital business and the first four months are the hardest. Chapter 34 explains exactly why profitable contractors go broke.

Why it is a bad idea anyway. You sign a certification on every payment application — a representation that the work covered is complete per the contract documents and the amount is due; on a federally funded job, a knowingly false one carries false-claims exposure. You starve at the end, because the $400,000 you pulled forward is not there in month fourteen when the finish trades, the commissioning agent, and the punch crew need paying, and jobs die in the last ten percent. It shows up in your WIP schedule, which your surety and your bank read, and an overbilled contractor looks profitable right up until it doesn't. And owners' reps have seen it — a $340,000 mobilization line on a $6.8 million job gets a phone call, and now the relationship starts with you explaining yourself.

Where the line is: an SOV where each line reflects the actual value of that work, including a reasonable and disclosed allocation of general conditions and mobilization, is honest. An SOV built to accelerate cash without regard to value is a false statement in monthly installments.

🏗️ From the field. Curtis Boone front-loaded Rivermont Elementary School #12. He is not a bad person; he is a good builder who believed cash flow was a game everybody played. In month eleven his drywall sub walked because Curtis could not fund a payment, the district withheld against a schedule he could not defend, and the claim he eventually filed was undermined by his own pay applications — which showed him 78% billed on work an independent observer put at 61% complete. The district's lawyer put those two numbers on one slide. That slide was the case.

📋 Try it: price the Willow Street site package three ways

You are the estimator. Here is the site package for the Willow Street Community Center — 2.1 acres, flat, with one existing 8-inch water main to relocate.

Field general conditions for the site package: $31,000.

Price it three ways:

(a) Lump sum — carry a 4% risk contingency on cost, then 7% overhead and profit. (b) GMP — 3% construction contingency, 4% fee. (c) Cost-plus with a 5% percentage fee.

Then answer the question that matters: the actual quantities come in 12% high across all six lines. For each structure, what does the City of Rivermont pay, what does the contractor earn or lose, and who absorbed the overrun?

Spaced Review

Answer these before you read the restatements. Say them out loud if you can — retrieval beats re-reading, every time.

1. From Chapter 3. You were told that delivery method and contract type are one decision. What does that mean mechanically — why can't you mix and match freely?

Because the delivery method determines when the contractor is engaged relative to design, and that timing determines what can be priced. You cannot get a lump sum from a contractor who joined at 40% design — there is nothing complete enough to bind a price to. You cannot get the preconstruction value of CM at Risk from a hard-bid process that selects on price after the drawings are finished. The delivery method sets what information exists at pricing; the pricing structure allocates the risk of what is still unknown. One decision, two names.

2. From Chapter 2. Recall roughly what net margin looks like in general contracting. Now use it: a project manager makes a $200,000 mistake. How much new revenue does the company have to win to erase it?

Net margins in general contracting commonly run in the low single digits — often on the order of one to three percent, varying widely by market, sector, and year. At 2%: $200,000 ÷ 0.02 = $10,000,000 of additional revenue to earn back one mistake. Kestrel does about $410 million a year. That mistake just consumed the profit on 2.4% of the entire company's annual volume. This is why a $57,510 contingency line matters, and why the difference between a 4% fee and a 3.5% fee is not a rounding error.

3. Deep callback, from Chapter 1. What is $10,650, and what are its two components?

It is Northgate's total daily exposure to slipping substantial completion: $5,150 per calendar day of extended general conditions plus $5,500 per calendar day of liquidated damages. You now know where both halves come from — the first from $2,900,000 ÷ 565 CD, the second from Meridian's build-up of leased interim clinic space, duplicated staffing, deferred clinical margin, and financing carry. In Chapter 1 it was a number. It is now a derivation you could defend in a negotiation, which is the difference between knowing a fact and owning it.

Project Checkpoint: The Willow Street Contract-Type Risk Map

In Chapter 3 you wrote a delivery-method comparison memo explaining why the City of Rivermont Parks & Recreation chose design-bid-build for Willow Street. That memo was about procurement. This one is about exposure: the City chose DBB and lump sum, which means you own more risk than under any other structure in this chapter. Now know exactly which risks, and what each is worth.

Recall your project. $6.8 million, 24,000 SF, two stories, wood-framed second floor over structural steel and CMU first floor. 425 calendar days. LDs $1,200/CD. Retention 5%. 100% payment and performance bonds. Prevailing wage. A 2.1-acre flat site with one existing 8-inch water main to relocate. Full package in Appendix K.

Your deliverable: a contract-type risk map — four columns, ten rows: the clause; plain English (one sentence, for somebody who does not read contracts); dollar exposure (a defensible estimate with the arithmetic shown); negotiate or accept (and if negotiate, what you offer in exchange).

Derive your daily rates first — three of the ten rows depend on them. Willow Street's general conditions are $680,000, exactly 10% of the contract sum:

$680,000 ÷ 425 CD = $1,600 per calendar day of extended general conditions $1,600 + $1,200 LDs = $2,800 per calendar day of total exposure to a slipped completion

Three rows, worked, so you can see the standard:

Build the remaining seven. Strong candidates: differing site conditions (and whether it survived the supplementary conditions — the water main relocation is your exposure); notice requirements; indemnity; termination for convenience; flow-down to your subs; prevailing wage and certified payroll compliance; the changes clause and its markup limits; consequential damages; dispute resolution and venue; order of precedence.

Then flag the three you would negotiate, and — this is what separates a professional from a wish list — say what you would give up for each. Concessions are not free. Offer a tighter notice period, a lower change-order markup, an earlier milestone. Something the City actually wants.

Next checkpoint: Chapter 5 builds the legal-framework checklist behind this map — bonds, insurance certificates, lien-notice deadlines, the AHJ list. Several exposures you just priced are managed by a policy or a bond rather than a clause, and you need to know which.

Chapter Summary

The one-sentence version: a construction contract is a risk-allocation instrument with a price attached, and the price is what the market charges to absorb the risks you assigned.

Seven things to check before you sign anything:

1. What is the basis of payment, and what specifically is inside the guarantee?
2. Is there a differing site conditions clause — and was it edited in the supplementary conditions?
3. Is there a no-damage-for-delay clause? Price it or negotiate it out. On Northgate one 30-day delay is a $154,500 difference.
4. What are the notice periods, and did you calendar every one on day one?
5. Are liquidated damages tied to a time-extension mechanism, and are they an exclusive remedy?
6. Is the consequential damages waiver mutual and intact?
7. What is the extended general-conditions daily rate, and is it agreed in the contract rather than left to be proven later? Northgate's is $5,150/CD, and it was worth negotiating for.

Three numbers to carry from memory: $5,150/CD (extended GC, from $2,900,000 ÷ 565 CD); $10,650/CD (total daily exposure, adding $5,500/CD in LDs); and $47,500,000 → $49,594,200 — what a guarantee actually did on a well-run job, 4.4% above the cap, with nobody doing anything wrong.

And the threshold concept once more, because you will have to say it to an owner someday: a guaranteed maximum price caps what the contractor can charge for the scope defined on the day it was set. It caps the contractor's exposure. The owner's cost is capped by the owner's discipline, the completeness of the design, and the ground. Say it at the GMP meeting, before anyone asks, and you will spend the next eighteen months solving problems instead of relitigating a misunderstanding.

What's Next

Chapter 5 takes the machinery underneath these clauses: mechanic's liens and the notice deadlines that vary brutally by state, payment and performance bonds and what a surety actually does when a contractor defaults, the insurance program that has to cover what your indemnity clause promised, and how disputes resolve when negotiation fails. Then Chapter 6 turns this chapter into a working tool — a risk register with owners, probabilities, and dollar impacts, and a contingency drawn down against named risks instead of guessed at as a percentage. After that, Chapter 7 teaches you to read the documents these contracts keep referring to, and to find the ten discrepancies that will cost somebody money.