Case Study 25.1 — Eleven Days: The Anchor-Bolt Submittal, Traced

All people, companies, and projects in this book are illustrative composites. The numbers are internally consistent and typical of the situation described; they are not drawn from a specific real project.

Setup

Project: Northgate Outpatient Pavilion — 132,000 SF, four stories, CM at Risk with a $47,500,000 Guaranteed Maximum Price (GMP), 565 calendar days, notice to proceed March 3, Year 1.

The item: one submittal. Anchor bolts and embeds for 985 tons of structural steel. Specification Section 05 12 00. Approximately forty sheets — a bolt-setting plan, base-plate details, embed plate details, templates, and a bill of material.

The parties: Ironbridge Steel (project manager Hank Duffy), Kestrel Construction Group (project manager Ray Alvarez, field engineer Dani Okonkwo), Caldwell Structural (principal Ruth Caldwell, engineer of record).

The stakes: steel erection is on the critical path. Planned start August 4, Year 1. Northgate's combined daily exposure to slipping substantial completion is $10,650 per calendar day — $5,150 in extended general conditions plus $5,500 in liquidated damages.

What everyone believed on March 7: that this was under control. There was a back-schedule. It was in the file. Ray had written it himself.

Sixty-six calendar days from execution to bolts in the ground. Zero float anywhere in the chain, and no resubmittal allowance in it at all.

What Happens

Control point 1 — February 20, Year 1: the date nobody computed

Re-run that back-schedule the way Chapter 25 §25.3 teaches it, with a realistic allowance for one resubmittal cycle: 7 calendar days for the fabricator to revise, 10 for a second engineer review. Seventeen days.

The submittal due date moves from March 27 to March 10. The subcontract execution date moves from March 7 to February 20.

The chain was fifteen days infeasible on the day the subcontract was signed. Nobody knew, because nobody had run it that way. The plan assumed a first-round approval, which is an assumption, not a plan.

Cost of the intervention: re-running one column of arithmetic before award. Free.

Control point 2 — the week of March 10: the question nobody asked

Ironbridge's mill order release is conditioned on approval of this submittal. The reason is mechanical, not bureaucratic: the column layout and base-plate geometry that generate the bolt-setting plan are the same information that locks column lengths and the advance bill of material.

Hank Duffy holds 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. April 18 is Ironbridge's production constraint, and on March 10 it exists in exactly two places: Hank Duffy's shop production schedule, and Hank Duffy's head.

Cost of the intervention: one question, asked once, in a phone call that was already happening. "Hank, what is the last date you can release to the mill, and what happens if you miss it?" Free.

March 27, Year 1 — a Friday, hour by hour

This is the day the submittal arrives, and it is worth walking slowly, because nothing goes wrong on it. That is the point.

Read that table again and find the villain. There is not one. Nobody is lazy. Nobody makes a bad decision. The single defect in the entire day is at 8:06 a.m., and it is not a decision at all — it is a sort order.

March 27 through April 7 — the eleven days

The planned Kestrel review is five calendar days: transmit to Caldwell by April 1.

Buyout continues. The mechanical and electrical packages are awarded. A subcontractor's bond comes back with the wrong obligee. The mass excavation contractor hits unsuitable material and there is a two-day argument about a differing site condition. The stack on the corner of the desk gets taller and then shorter and then taller.

Submittal 019 goes out to Caldwell Structural on April 7. Eleven calendar days instead of five.

Control point 3 — Monday, March 30: the report nobody ran. A ball-in-court report generated that Monday would have shown eleven items in GC review, sorted by days aging, with a column showing what each one gates. Submittal 019 would have appeared with three days remaining against a five-day internal target and a zero-float flag on the chain.

It takes about eleven minutes to produce that report and read it aloud. Cost of the intervention: free. It was not run, because nobody had built a log that could produce it.

Control point 4 — April 7: the email nobody sent. On the day the submittal was transmitted, Kestrel could have sent Ruth Caldwell a short, factual, entirely ordinary request:

Submittal 019 (anchor bolts and embeds, Section 05 12 00) is transmitted today. This package releases Ironbridge Steel's mill order, which must be released by April 18 to hold a reserved rolling slot. The contractual review period would return it April 21. Is an expedited return by April 16 possible? Please advise so we can plan.

Ruth Caldwell is not an adversary. She is an engineer of record with a contractual fourteen-day period and no information suggesting this package is different from any other. Asked in advance, with a reason, on a forty-sheet bolt plan she has effectively already designed, an eight- or ten-day return is a normal accommodation.

Cost of the intervention: one email, four minutes. Free. It was not sent, because nobody at Kestrel knew April 18 existed.

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 stamped approved as noted on April 21.

Note the stamp. This is the best available outcome from the review — no resubmittal, immediate release, comments to incorporate in fabrication. On the day the review returns, Caldwell Structural has performed flawlessly and Ironbridge Steel has performed flawlessly.

The job is still going to lose twenty-three days.

April 21 — the slot is gone

Hank Duffy takes the call. Release was required April 18. It is April 21.

May 12 through May 18 — the slip nobody noticed

There is a second, smaller consequence that gets forgotten in the telling, and it is instructive.

Fabrication of the bolts could not start until April 22 rather than April 16. Twenty-one days of fabrication and galvanizing plus five days of transit put the anchor bolts on site May 18 instead of May 12 — six days late to the first footing forms.

Jamal Foster absorbed it without drama. He resequenced the footing pours, placing the forty-one footings that carry no column anchorage first and holding the column footings for the following week. Foundations still completed July 3, exactly as scheduled.

Why did one six-day slip vanish and another six-day slip cost twenty-three days? Because a footing sequence is continuous — you can reorder it — and a mill rolling slot is discrete. A six-day slip against a continuous activity costs whatever the resequencing costs, which was nothing. A six-day slip against a discrete scheduled event costs the interval to the next opening, which was five weeks.

Lead-time chains do not add. They snap.

The decision

The arithmetic says acceleration saves $13,050 — nearly a wash, and well inside the error bars of any of those estimates. The real driver was that Meridian Health System's leased interim clinic space expires October 1, Year 2: six days late is survivable, twenty-three days late is not.

Kestrel accelerated. Nadia Haddad approved the $168,000.

And then the second lesson landed, months later and in a different currency: acceleration produced trade stacking, 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, when a plank that had been lifted overnight by another trade shifted under a mason tender at 7:20 in the morning.

A submittal-log failure in April became a safety investigation in November. Nobody traded safety for schedule. The pressure moved through the organization on its own.

Analysis

What failed, precisely

Not competence. Not diligence. Not any individual's judgment. The failure was in a sort order and a missing column, and it produced a $168,000 acceleration, a six-day residual slip, and a near-miss.

Here is the whole causal chain in one table.

Any one of the four would have saved it. Control point 1 gives you fifteen days of margin. Control point 2 makes the real gate visible so the whole team manages to April 18 instead of April 15. Control point 3 catches the eleven days on day three of eleven. Control point 4 recovers the six lost days inside the review.

All four were free. All four were missed.

The one that hurts most is control point 3, because it was the closest to being caught and the cheapest to institutionalize. Six days of desk time produced a 23-day critical-path delay and a $168,000 acceleration — $28,000 per day that a document sat in a stack.

The mechanism, stated generally

Three things about this story generalize to any project you will ever run.

Nothing on a submittal announces its own importance. Submittal 019 looked exactly like the mockup schedule and the temporary power one-line, because a submittal is a stack of paper and stacks of paper are visually identical. The only thing that can distinguish the gating submittal from the routine one is a column in a log that says what it gates. If that column does not exist, your team is managing by memory, and memory fails during buyout.

The document that matters is the one people touch daily, not the one that is technically correct. Kestrel had a correct back-schedule. It lived in a procurement file that opened monthly. The submittal register — the document the team touched every day — had no required-approval date in it at all. Producing a correct analysis and storing it where nobody works is functionally identical to not producing it.

Your own column is the one to look at first. When the delay surfaced, the first instinct at Kestrel — and it is everyone's first instinct — was to look at Caldwell's fourteen days rather than Kestrel's eleven. Caldwell used its contractual period and returned the best possible stamp. The eleven days were ours.

What Kestrel changed

Three things, on April 16 of Year 1, and all three carried to every Kestrel job afterward.

1. One log, sorted by submittal due date, with a required-approval date and a "gates what" column on every row.
2. A gating-submittal list — the six items whose approval releases a long-lead purchase — on one page, with a 48-hour internal turnaround commitment and a daily standup that asks only where are they.
3. A Monday ball-in-court report, published to the entire team including the architect and the owner's representative, without editorial comment.

Total cost of the three changes: about ninety minutes a week of one field engineer's time.

Discussion Questions

1. Four control points were available and free, and all four were missed. Rank them by how easy each would be to institutionalize as a habit rather than a heroic individual act — and explain why the easiest one to institutionalize is not necessarily the one that would have saved the most days.

2. Caldwell Structural took its full fourteen days and returned the best possible stamp. Ironbridge Steel submitted exactly on schedule. Kestrel's project coordinator logged the submittal accurately in every field. Every individual performed. Explain, in your own words, how a system composed entirely of people performing correctly produced a twenty-three-day failure.

3. A six-day slip on the anchor bolts cost nothing (Jamal Foster resequenced the footings). A six-day slip on the mill release cost thirty-five days. Identify three other discrete, scheduled events on a typical commercial building where a small slip triggers a large one, and say how you would flag them on a submittal log.

4. Kestrel's acceleration saved $13,050 on the arithmetic and was still the right decision. Explain why — and then explain what would have to be true about Meridian's situation for "absorb the 23 days" to have been the right answer instead.

5. The acceleration produced a spike in near-misses eleven weeks later. Argue both sides: is it fair to attribute a safety event on the north elevation to a submittal that sat on a desk in April? What would you write, and not write, about that connection in a lessons-learned report that Nadia Haddad will read?

Your Turn

Take your own project — or the Willow Street Community Center package in Appendix K — and build the gating-submittal list. Not the whole log. The four to eight submittals whose approval releases a long-lead purchase or gates a discrete, scheduled event.

For each one, produce four things on a single page:

• The required-on-site date and where it came from in the schedule.
• The release gate — and, critically, whether that gate is your approval date or the fabricator's release date, which you will find out by asking.
• The submittal due date, back-scheduled with one resubmittal cycle included.
• The name of the one person who is responsible for it.

Then answer the question that would have changed this case study: is any of those dates already in the past?