Case Study 1 — Three Findings on the North Elevation

Every person, company, and project in this book is a Tier-3 illustrative composite. The events below are constructed to teach; they are not a real accident.

Setup

Project: Northgate Outpatient Pavilion — 132,000 SF, four stories, $47,500,000 GMP, 565 calendar days, Kestrel Construction Group for Meridian Health System.

When: Week 34 of the job. The steel had started 23 calendar days late after the anchor-bolt submittal sat in Kestrel's own office for 11 days. Kestrel accelerated: a second erection crew, premium Saturday time, and a resequenced enclosure by area, at a cost of $168,000 to recover 17 of the 23 days. That decision was made in week 31, in the trailer, by Ray Alvarez and Nadia Haddad, and it was — on the arithmetic and on the interim-clinic-lease deadline — the right call.

Where: The north elevation, level 3. Ashcroft Masonry is laying the CMU backup wall behind the architectural precast. Their frame scaffold has been standing nine days. It was erected by Ashcroft's own crew, on mudsills, tied to the structure, guardrails on, planked out.

The people: Emiliano "Milo" Serrano, mason tender, four years in the trade. Beatriz "Bea" Salgado, Kestrel's corporate safety director, formerly an ironworker. Marguerite "Margo" Deacon, general superintendent, 31 years. Ray Alvarez, senior project manager and the narrator of this book. Dani Okonkwo, field engineer, in their eighth month.

What Happens

The timeline

The three findings

Finding 1 — no competent-person inspection for that shift. The tag was two days stale. Ashcroft's competent person was on site Tuesday morning but had gone directly to the level-2 face, where the crew was starting. There was no procedure making the tag a precondition of access; it was a document that got updated when somebody remembered.

Finding 2 — a scaffold modified by a trade that did not erect it, with no re-inspection. The Halcyon electricians altered a structure that belonged, operationally, to Ashcroft. There was no rule in the subcontracts, in the orientation, or on the after-hours form that said anything about it, and no mechanism by which anybody would have found out.

Finding 3 — a crew running behind after the acceleration, with an unwritten pressure to make it up. Ashcroft was eleven working days behind on the north elevation. The resequenced enclosure had pushed their start. Nobody at Kestrel had ever told Ashcroft in writing that they were expected to recover. Nobody had to. Ashcroft's superintendent had heard "north elevation" in every coordination meeting for three weeks, and he had done what a good subcontractor does, which is find hours: earlier starts, tighter breaks, and a crew that went up the ladder at 7:20 without looking at anything on the way.

The conversation that produced the report

The first draft of the report had two findings. Bea brought it to Ray's office.

Bea: "It's missing the reason."

Ray: "It has the tag and it has the modification. Those are the causes."

Bea: "Those are the mechanisms. The reason is on page one of the schedule update. Ashcroft is eleven days behind because we accelerated the steel, and every meeting for three weeks has ended with somebody saying 'north elevation' in a particular tone of voice. That crew went up a ladder at 7:20 in the morning without looking at the tag because they were behind and they knew it and they knew we knew it."

Ray: "If I write that down, it goes in the file. It's discoverable. And it says that a decision I made in week thirty-one produced this."

Bea: "Yes."

Ray: "Was the acceleration wrong?"

Bea: "No. I'd have made the same call. That's exactly why it has to be in there. If accelerating is always right and it never carries a cost we write down, then we'll accelerate every job and be surprised every time."

It went in the report.

The corrective actions

Two categories, and the difference between them is the point of this case.

Corrective actions for the specific findings:

Corrective actions for the system — the ones that made this investigation worth having:

1. The temporary-structures register. Every temporary structure on the job gets a row: designer, design document, erector, competent person and backup, inspection frequency, removal authority. Read aloud at the weekly coordination meeting until everyone can name their own cells.

2. The acceleration protocol. This is the one that came out of Finding 3, and it is the one that would not exist if the finding had been left out. Kestrel's rule, adopted company-wide after Northgate:

Any decision to accelerate, resequence, or compress work is accompanied by a written statement of the hazard it creates and the specific controls added because of it — before the acceleration is directed. The statement names the trades whose sequence changes, the areas where trades will now be stacked, the crews expected to recover time, and what is being added to offset it: additional supervision, additional planning time, an added safety walk, or a deliberate decision to give a crew the days rather than the pressure.

3. The recovery expectation gets written down or it does not exist. If Kestrel expects a subcontractor to make up eleven days, that expectation goes in a letter with a revised sequence, a means of recovery, and a discussion of what it costs. Unwritten pressure is the most dangerous kind, because nobody can push back on something nobody said.

4. Near-misses get investigated like injuries. Same form, same interviews, same corrective-action tracking, same closeout review. Northgate had eleven near-misses that month and investigated one. After this, the criterion became potential severity, not actual outcome.

Analysis

Why the first two findings were easy and the third was hard. Findings 1 and 2 blame procedures. Nobody in the room owns a procedure personally, so writing them down costs nothing. Finding 3 blames a decision, and the decision had a name on it. Investigations tend to stop at the last mechanical link in the chain — the plank, the tag, the electrician — because the mechanical links are unarguable and the management links are uncomfortable. The predictable result is that organizations fix the same class of incident repeatedly and never fix the thing generating it.

Why the corrective actions for Findings 1 and 2 were necessary but not sufficient. Suppose Kestrel had done only the tag system and the modification rule. Both are genuinely good controls, and both would probably have caught this specific plank. But the pressure would still have been there, and pressure does not have one outlet. The same crew, eleven days behind, would have found the next shortcut: an unballasted ladder, a platform loaded past its class with block, a guardrail section left off for a hoisting operation and not replaced. A control removes a hazard. It does not remove the reason people are looking for shortcuts.

Why the acceleration itself was still correct. This is the part that gets lost. The acceleration cost $168,000 and recovered 17 days against an exposure of $10,650 per calendar day and a hard external deadline — Meridian's interim clinic lease expiring October 1 of Year 2. Six days late was survivable; twenty-three was not. Choosing not to accelerate would not have been the safe decision; it would have been a different decision with a different set of consequences. The error was not the acceleration. The error was accelerating without pricing the hazard it created and funding the controls to offset it. Acceleration has a line item. Everybody knows the overtime line. Almost nobody writes the supervision-and-planning line next to it.

Why "nobody was hurt" is the most dangerous sentence in an incident report. The physics of that morning came down to whether a top rail held. It held. Had it not, the same three findings would have appeared in a report with a fatality attached, and the corrective actions would have been identical — arrived at through a very different process, at a very different cost, with lawyers in the room. The information content of the event was the same either way. Only the price of the tuition changed.

Discussion Questions

1. Ray's objection to Finding 3 was that writing it down creates a discoverable record implicating his own decision. That is a real concern, not a fake one. How do you answer it? What is the argument for writing it anyway, and what is the professional line between honest root-cause analysis and creating an admission?

2. The acceleration protocol requires a written hazard statement before the acceleration is directed. Write one for the week-31 decision on Northgate: name the trades whose sequence changes, the areas where stacking will occur, and three specific controls you would fund.

3. Corrective action 5 — the eight-second morning question — costs nothing and was arguably the most effective item on the list. Why do controls like that get proposed rarely and adopted even more rarely? What does that tell you about how organizations choose safety investments?

4. Kestrel had eleven near-misses that month and investigated one. Write the criterion you would use to decide which near-misses get a full investigation. Then estimate what your criterion costs per month in hours, and defend it.

5. The scaffold was, in every physical sense, a good scaffold — correctly designed, correctly erected, correctly tied, correctly planked. Argue both sides of this: is a shared temporary structure that four trades will use over eight weeks a scaffold problem or a contract problem? What does your answer change about how you write the next subcontract?

Your Turn

Take a scaffold on a job you know — or the masonry scaffold you specified for the Willow Street Community Center in this chapter's Project Checkpoint — and write the modification-control clause you would insert into every subcontract on that job. Two hundred words, maximum. It must answer, unambiguously: who owns the scaffold, who may alter it, what a trade does when it needs a modification, what happens on discovery of an unauthorized modification, who re-inspects, and who pays for the delay if a modification request is not answered promptly.

Then answer the harder question in one paragraph: what would make a trade actually follow it at 11:40 on a Monday night when the alternative is a ten-minute wait and everybody has gone home? If your answer is "training" or "culture," you have not finished. Find the answer that changes the situation rather than the person.