The Best Parts Don’t Make the Best Project: Ackoff on Managing the Seams

ISSUE 10 · JULY 23, 2026

You bought the job out right. The steel erector has a résumé you’d frame. The concrete sub is one you’ve run four projects with. The MEP contractor came in under budget with the best references in the market. On paper, it’s the strongest team you’ve put on a project in years.

Six months later you’re five weeks late, and nobody can tell you exactly where it went.

Here’s the uncomfortable part: nothing on that list was a mistake. Every trade is good. Every trade, by their own numbers, is performing. And the job is still slipping.

Russell Ackoff spent his career explaining why that happens. He was a Wharton professor and a pioneer of operations research — the study of how complicated systems actually behave — and he had a thought experiment he used to open his talks. Suppose you want to build the best car in the world. Bring every car ever made into one hangar, hire the best engineers, and have them pick the best part from each one: the Rolls-Royce engine, the best transmission off another, the best brakes off a third. Now assemble those world-class parts into a world-class car.

You can’t. The parts don’t fit — they were never designed to work with each other. What you end up with isn’t the best car in the world. It isn’t a car at all. It’s a very expensive pile.

Ackoff’s conclusion, from a 2003 lecture: “A system is never the sum of its parts; it’s the product of their interaction.” And the warning that follows it: “When you improve the performance of a part of a system taken separately, you can destroy the system.”

We buy construction in parts. The whole industry is organized that way — scopes, cost codes, and sub schedules are all drawn up trade by trade. But the building doesn’t get built by trades working separately. It gets built in the seams between them: the slab that has to be poured and cured before steel can set, the layout that has to be right before MEP rough-in, the RFI answer that has to land before fabrication releases.

Every scope has an owner. Every seam has two owners, which in practice is the same as none.

And it gets worse when each part optimizes itself. The erector sequences the steel the way that minimizes their crane time, which saves them money. But that sequence opens the deck to the trades behind them three weeks later than a different sequence would have. The concrete sub pours the big open slab first because it’s the best production day they’ll have all month, not because it’s the area the follow-on work needs. Nobody did anything wrong. Both decisions would pass an audit. Together, they push the job right.

That’s not a sub problem. It’s a system problem — and it belongs to the PM, because the PM is the only person on the project whose actual job is the whole thing. Everybody else is responsible for a part. You’re responsible for how the parts interact.

So manage the seams like they’re scope, because they are. At your next pull plan or buyout meeting, ask every trade two questions: what do you need from the crew ahead of you, and what are you handing off to the crew behind you? Write the answers down somewhere they’ll be looked at again. That list — the handoffs, not the scopes — is a truer picture of your project than the schedule is. And when something slips, make the first question “which interaction broke?” instead of “whose fault was it?” The second question feels productive and almost never changes anything.

Watch your reporting for the same trap. Per-trade productivity, per-package cost variance, per-sub PPC — all useful, all measuring parts. It is entirely possible to run every part green and lose the job.

Something to consider Project data arrives pre-siloed: daily reports by trade, RFIs by package, cost by code. That structure makes it easy to see how each part is doing and nearly impossible to see how they’re interacting. This is a decent use for AI. Dump the daily reports, RFI log, and schedule updates in together — all trades, not one — and ask cross-cutting questions. Which handoffs have slipped more than once? Which pairs of trades keep showing up in the same delay notes? Did that RFI in the steel package show up as a delay in the MEP package a month later? It’s aggregation work: turning trade-level data into a project-level view, which is exactly the job nobody has time to do by hand. What it can’t do is sit in your coordination meeting and notice that two foremen have quietly stopped talking to each other. The interactions that are written down, AI can find. The ones that matter most usually aren’t.

Pick one handoff this week — the one that’s burned you before — and manage it as its own scope for a month. That’s a systems view you can start on Monday.

---

Steel Market Snapshot

Mid-July 2026 — Nucor resumes HRC hikes at $1,135 as plate pushes to $1,335

HOT-ROLLED COIL $1,135/ton ▲ +$5 (Jul 13) after 3-week hold	PLATE $1,335/ton ▲ +$45 over four weeks (Jul 7)
WIDE-FLANGE BEAMS $1,100–1,400/ton ▬ Flat · last hike ~$90/ton late May	REBAR +$60/ton ▲ SDI/CMC/Gerdau (early Jun)

Nucor resumed raising hot-rolled coil on July 13 after holding flat for three weeks, and plate mills are stacking increases into the fall order books. The pause was a breath, not a top. Supply is still the story: spot availability is thin, mills are prioritizing contract customers, and service center inventories keep drawing down.

Hot-rolled coil is at $1,135/ton (Nucor CSP, July 13), up $5 after a three-week hold at $1,130. That pause followed 23 consecutive weekly increases totaling $380/ton through June 22. SMU’s market average was about $1,160/ton as of July 7 — roughly 36% above the same week last year. Lead times hold at three to five weeks, and West Coast CSI coil sits at $1,185/ton.

Plate keeps climbing and looks the tightest of the bunch. SMU’s assessment averaged $1,335/ton on July 7 — up $15 week over week and $45 over four weeks. SSAB Americas announced another $40/ton increase July 8 on spot orders shipping on or after August 30, and Nucor is seeking $60/ton more with its September order book while holding a $10/ton fuel surcharge from August 1. Plate sources are telling SMU they see no peak this year.

Wide-flange beams run roughly $1,100–1,400/ton through service centers, with no new July mill announcement on top of Nucor-Yamato and Gerdau’s ~$90/ton beam hike in late May and Gerdau’s $40–80/ton long-product increase effective June 8.

Rebar holds the early-June ground: base prices up $60/ton from Steel Dynamics, CMC, and Gerdau, plus roughly $40/ton on 20-foot bar.

What to watch: Mill capability utilization eased to 79.7% the week ending July 11 (1.84 million net tons), down from 80.4% the week prior but up 3.9% year over year, with year-to-date utilization at 78.8% versus 77.0% last year. Two things cut against the run: imports climbed back above 2 million tons in both May and June, and service center flat-rolled supply fell for a sixth straight month to 44.5 shipping days — thin inventories that will have to be refilled. For precon, nothing changes yet: bid firm-to-higher through Q3, and get plate-heavy packages locked before those September books open. For live data, visit our Market Informer page.

---

Sources & Further Reading

1. Ackoff, Russell L. “Systems Thinking and Management” (2003 lecture), quoted in Hunter, John. “Ackoff on Systems Thinking and Management.” The W. Edwards Deming Institute, September 2, 2019. deming.org

2. Ackoff, Russell L. Re-Creating the Corporation: A Design of Organizations for the 21st Century. Oxford University Press, 1999. ISBN 0195123875.

3. “Ackoff on Systems Thinking” (the best-car thought experiment). Lateralworks, January 17, 2019. lateralworks.com

4. Steel Market Update. “Nucor ups spot HR price by $5/ton.” July 13, 2026. steelmarketupdate.com

5. Steel Market Update. “Nucor holds spot HR price at $1,130/ton, again.” July 6, 2026. steelmarketupdate.com

6. Steel Market Update. “SSAB ups plate prices by $40/ton, again.” July 8, 2026. steelmarketupdate.com

7. Steel Market Update. “Nucor aims for a $60/ton bump on plate prices, holds fuel surcharge.” July 16, 2026. steelmarketupdate.com

8. Steel Market Update. “Plate sources predict higher spot prices, no peak in sight.” July 16, 2026. steelmarketupdate.com

9. IndexBox. “U.S. Raw Steel Production Week Ending July 11, 2026: AISI Reports 1.84 Million Net Tons.” July 2026. indexbox.io

10. Steel Market Update. “Steel imports recovered further in May and June.” July 16, 2026. steelmarketupdate.com

11. Steel Market Update. “June service center shipments and inventories report.” July 16, 2026. steelmarketupdate.com

12. GMK Center. “Nucor-Yamato and Gerdau are raising beam prices in the U.S.” May 2026. gmk.center

13. GMK Center. “Three major US steelmakers raise rebar prices by $60/t.” June 2026. gmk.center

---

Champion Steel Works — Steel you can trust.