Case Study 16-1 — Six Days on a Desk

Setup

Project: Northgate Outpatient Pavilion — 132,000 gross square feet, four stories, structural steel frame, 985 tons erected. CM at Risk with a guaranteed maximum price of $47,500,000. Notice to proceed March 3, Year 1; contract substantial completion September 18, Year 2; 565 calendar days; liquidated damages $5,500 per calendar day.

The people. Ray Alvarez, senior project manager for Kestrel Construction Group. Dani Okonkwo, field engineer, eleven days into their first real job. Hank Duffy, project manager for Ironbridge Steel, the structural steel subcontractor. Ruth Caldwell, principal of Caldwell Structural, the structural engineer of record. Wei Chen, Kestrel's project controls manager.

The constraint. Steel erection was scheduled to start August 4, Year 1, and steel erection was on the critical path. Every day of steel slip is a day of substantial completion slip unless somebody buys it back. Northgate's total daily exposure to slipping substantial completion is $10,650 per calendar day — $5,150 in extended general conditions plus $5,500 in liquidated damages.

(Every company, person, and project in this case study is an illustrative composite.)

What Happens

March 7, Year 1 — the subcontract is executed

Ironbridge Steel's subcontract is signed at $3,795,000 against an estimate of $3,840,000 — $45,000 favorable, and the first green number on the buyout log. Ray and Hank had a good scope review call. The exclusions were clean, the connection design responsibility was clearly delegated to Ironbridge's engineer, and Hank's erection crew was confirmed and available.

Ray builds the procurement back-schedule the same afternoon and puts it in the file. Here it is, exactly as written:

Sixty-six calendar days from execution to bolts in the ground. Ray shows it to Wei Chen, who loads the dates into the submittal log in the schedule. Everything is documented. Nothing is missing from the file.

Except one thing, which is not in the file, is not in the schedule, and is not in anyone's head at Kestrel.

The number nobody wrote down

Ironbridge's mill order release is conditioned on approval of the anchor bolt and embed submittal. The reason is mechanical, not bureaucratic: the column layout and base-plate details that generate the bolt-setting plan are the same information that locks column lengths and the advance bill of material. Until the engineer of record signs off on that geometry, Hank will not release material against a mill slot he cannot un-release.

Hank has a reserved rolling slot at the mill for the wide-flange shapes. Release is required by April 18.

Not April 15. April 15 is Kestrel's approval milestone — the date on Ray's back-schedule. April 18 is Ironbridge's production constraint, and it lives on Hank Duffy's shop production schedule, which Kestrel has never asked to see.

Between the two dates sits three days of margin that nobody knows exists and nobody is managing.

March 27 — Ironbridge submits, on time

Hank's detailer transmits the anchor bolt and embed submittal to Kestrel on March 27, exactly per the back-schedule, eighteen calendar days after starting. Ironbridge is not the problem in this story and never becomes it.

March 27 through April 7 — the eleven days

The submittal arrives during the week Kestrel is awarding six packages. Ray is in scope review calls. Dani is chasing insurance certificates. The submittal is logged, stamped received, and placed in the review stack with fourteen other submittals — the mockup schedule, the concrete mix designs, the elevator shop drawings, the temporary power one-line.

It sits for eleven calendar days instead of the planned five.

There is no villain in this paragraph. Nobody decides to delay it. Nobody is lazy. The submittal simply does not look different from the other fourteen, because nothing on it says this document is the gate on a mill slot. It goes out to Caldwell Structural on April 7.

April 7 through April 21 — Caldwell does exactly what it agreed to do

Ruth Caldwell's office takes fourteen calendar days. That is her contractual review period. She uses it, entirely properly, and returns the submittal approved as noted on April 21.

Caldwell is not the problem in this story either. This matters, because when the delay surfaces, the first instinct at Kestrel is to look at the fourteen days rather than the eleven.

April 21 — the slot is gone

Hank Duffy calls Ray the same afternoon.

Hank: "I got the approval. I can't use it."

Ray: "What do you mean you can't use it?"

Hank: "My release date was the eighteenth. The mill rolled somebody else's order. Next opening for these shapes is May twenty-third."

Ray: "That's five weeks."

Hank: "That's five weeks."

The arithmetic of the slip

And the anchor bolts themselves? They arrived on site May 18 instead of May 12 — six days late, absorbed inside the footing sequence without incident. The bolts were never the problem. The approval was, because two entirely different things depended on it and only one of them was on a log.

The decision that followed

Acceleration saves $13,050 against doing nothing — arithmetically a wash. Kestrel accelerated anyway, because Meridian Health System's leased interim clinic space expires October 1, Year 2. Six days late is survivable. Twenty-three days late is a hospital system without a place to see patients.

The second bill arrived later and was not paid in dollars: trade stacking through the fall, a rework event on deck-edge detailing, and a spike in near-misses in weeks 34 through 36 — including the scaffold event on the north elevation that put Milo Serrano on his hands on a level-3 platform at 7:20 in the morning.

Analysis

Six calendar days of desk time produced a twenty-three-day critical-path delay and $168,000 of acceleration. That is $28,000 per day that a document sat in a stack. No crew, no crane, no material on the Northgate project burned money at that rate.

The mechanism deserves precision, because the wrong lesson here is "review submittals faster."

The real failure is that the back-schedule had no float and tracked the wrong milestone. A chain of six sequential steps with zero slack will fail, because at least one step in six always takes longer than planned. Ray built a plan that required perfection from Kestrel, from Ironbridge, and from Caldwell simultaneously, and then was surprised when one of the three was ordinary.

The second failure is subtler and more valuable: Kestrel managed its own milestone instead of the subcontractor's constraint. April 15 was a Kestrel date. April 18 was the date that actually mattered, and it belonged to Hank. Nobody asked for it. This is the most transferable lesson in the chapter — your approval date is not the constraint; the fabricator's release date is — and the only way to learn it is to ask every long-lead subcontractor a specific question at buyout: "What is your no-later-than release date, and what happens if we miss it?"

The four control points

Each of these was available in March, at essentially no cost, and each one alone would have prevented all twenty-three days.

1. Ten days of procurement float in the back-schedule. The subcontract was executed March 7 because that is when the paperwork cleared. Ironbridge's price and scope were settled during preconstruction, weeks earlier. A letter of intent issued February 26 — authorizing submittal preparation and capping Kestrel's exposure at, say, $50,000 if the job did not proceed — would have started the chain nine days earlier and absorbed the entire slip with room to spare. Cost: one page and a signature.

2. A gating-submittal list with a 48-hour internal clock. Of the fifteen submittals in Kestrel's stack that week, exactly four were gates on long-lead releases. Those four should have been on a separate one-page list, taped to the wall, with a named owner (Dani), a two-business-day internal commitment, and a daily standup that asked only one question: where are the four? The other eleven could have taken two weeks and nothing would have happened.

3. Tracking the release-to-mill date instead of the approval date. If April 18 had been a line on the procurement log with Hank's name next to it, the eleven-day hold would have generated an alarm on April 4, not a phone call on April 21. Fourteen days of warning is enough time to hand-carry a submittal to a structural engineer's office.

4. A pre-negotiated expedited review for gating submittals. Ruth Caldwell's fourteen days are contractual, not physical. At the design-team kickoff in March, Kestrel could have identified the four gating submittals by name, asked Caldwell and Trellis Engineering for a seven-day turnaround on those four specifically, and paid for it out of the design-assist line if necessary. Engineers say yes to this request more often than contractors ask it, because a named short list is a manageable promise and "review everything faster" is not. Seven days instead of fourteen would have delivered approval on April 14 even with Kestrel's eleven-day hold.

Notice what none of the four are. None of them is "work harder." None is "hire more staff." All four are things you do in the first three weeks of a job, on paper, for free — which is Theme 3 of this book, stated as plainly as it can be stated: the project is built twice, and the first build determines the second.

What Kestrel changed

Kestrel's procurement log now carries three columns it did not have before Northgate: required-on-site date, supplier's no-later-than release date, and days of float remaining. The third column is calculated and it is colored. When float on a gating item drops below five days, it appears on the Monday operations report whether or not anyone thinks there is a problem.

Discussion Questions

1. Ray's back-schedule was complete, accurate, documented, and shared with the scheduler. It still failed. What is the difference between a schedule that is correct and a schedule that is managed, and which of the four control points addresses that difference most directly?

2. Caldwell Structural used exactly the review period its contract allowed. Ironbridge submitted exactly on time. Kestrel's eleven days were the only deviation from plan. Is it fair to say Kestrel caused 100% of the delay? Argue the other side: what responsibility, if any, does a fabricator have to tell the contractor about a constraint the contractor never asked about?

3. The acceleration decision saved $13,050 on paper — essentially nothing — and Kestrel did it anyway. Explain the decision in terms of the owner's business rather than the contractor's cost report. Then explain why a purely arithmetic decision framework would have gotten this wrong.

4. Control point 4 (pre-negotiated expedited review) requires asking the architect and engineers for something before you need it. What is in it for them? Draft the two sentences you would say at the design-team kickoff to make that request land as collaboration rather than as pressure.

5. The acceleration produced a near-miss spike in weeks 34–36. Trace the causal chain from an April submittal review to a November scaffold plank, naming each link. At which link would you have intervened, and with what?

Your Turn

Take the switchgear back-schedule from §16.8 of the chapter — the one that ends at the February 10, Year 2 rigging window — and do to it what the four control points would do.

1. Add procurement float. How many days, and where in the chain do you put them? (Hint: float at the front is cheaper than float at the back. Explain why.)
2. Identify the equivalent of the "release-to-mill" date. For manufactured equipment, what is the supplier's real internal constraint, and what question do you ask Devlin Achebe at Halcyon Electric to find it?
3. Write the one-line entry for this item as it would appear on Kestrel's revised procurement log, including all three new columns.
4. State the date, in Year 1, on which this item would first appear on the Monday operations report if the float column were set to alarm at five days.