Everyone Padded the Schedule. The Job Still Finished Late.
Updated: 6 hours ago
ISSUE 16 · OCTOBER 22, 2026
Nobody pads a schedule to be dishonest. They do it because they've been burned before, and they don't plan to be burned again.
Eliyahu Goldratt built Critical Chain (North River Press, 1997) around that reflex. As a Marris Consulting abstract of the book lays it out: the person doing the work gives a duration with enough slack for roughly an 80% chance of finishing on time. Then a manager adds a margin to protect the team. Then someone above adds another. And still, most tasks finish on their planned date, rarely early. The safety went somewhere. Goldratt's question is where.
His answer is two habits. The first is student syndrome: people start the work shortly before the deadline, the way students do with assignments, so the cushion is spent before the work begins and any surprise lands right on the date. The second is Parkinson's Law: work expands to fill the time allotted. Padding every task protects less than it looks like it does, because the padding gets used up before anyone needs it.
There's a third cost. Longer durations mean more tasks open at once, and under pressure from stakeholders, managers split resources across all of them. The abstract's example: five one-week tasks that could run back to back get spread out so the first one takes three weeks instead of one. Anyone who has watched a crew get pulled onto five fronts to keep everyone happy knows how that goes.
The fix, as the method is usually laid out, is to take the safety out of the tasks and pool it. Plan each task at a 50% best-guess duration. Put the time you removed into a project buffer at the end of the critical chain (the longest string of dependent tasks, counting resource limits, that sets the finish date), and into feeding buffers wherever a side path joins that chain. The end of the project buffer is the date you give the owner. One common rule of thumb sizes the project buffer at half the length of the critical chain. Another uses the difference between a “safe” duration and the 50% one.
Then the buffer becomes a gauge. In the abstract's rule: if less than a third of a buffer has been used, do nothing. Between a third and two-thirds, stay alert and have a recovery plan ready. Past two-thirds, act. That changes the weekly conversation. Instead of arguing about whether one activity is “on schedule,” you ask how much protection is left and how much work is still ahead of it.
A fair warning: this is harder in construction than on paper. Trades work to durations written into contracts, and nobody will agree to a number that makes them look slow or sets them up to miss. Halving a sub's duration in the schedule won't survive the first coordination meeting. What you can take is the question underneath it: how much safety is in this schedule, and who is holding it?
Something to consider Your own project data will show whether student syndrome is real on your jobs. Export the baseline schedule along with actual start and finish dates for completed activities (or pull them from the daily reports), and ask an AI tool to compare planned and actual by trade. How many activities finished early, on the planned date, or late? How often did work start on its planned start date versus later? How much float was gone before the activity even began? A pattern of late starts and on-time finishes is student syndrome in your own numbers. The limits: it only knows what the data holds, actual start dates in daily reports tend to be loose, and it can't tell you why a trade started late. That usually takes a phone call.
So try this. Pick one stretch of your look-ahead that the finish date depends on. Ask each trade two questions: how long if nothing goes wrong, and how long would you commit to? Write down both. The gap is the safety. Then decide where it lives: in one visible buffer that you manage, not spread across a dozen activities where nobody can see it. Check how much of it is gone every week.
Steel Market Snapshot
Mid-October 2026 — beam and tube lists hold. Labor is still open.
WIDE-FLANGE BEAMS $1,750/ton ▬ no change · mill lists Oct 1, RMS incl. | HSS & TUBE +$320/ton ▲ since July 21 · last hike +$100 Sept 29 |
NYS FUEL SURCHARGE 57% ▲ from 47% · effective Oct 1 | MILL UTILIZATION 77.8% ▼ from 78.3% · wk ending Oct 10 |
Wide-flange beams are still $1,750/net ton. Nucor-Yamato's Oct. 1 price list, which replaced the Aug. 3 list, remains the current posting as of Oct. 19: common sections at $87.50/cwt, RMS included. Steel Dynamics' Columbia City list is also still the Oct. 1 revision at $87.50. The rest of the Nucor-Yamato list is unchanged too: C15 channel $88.50, MC12/13/18 $101.25, L8 angles $98.75 and S-shapes $112.25 (about $2,245/ton). There is no November list yet.
Tube is quiet, with a date to watch. Nucor Tubular's announcement index shows nothing since the Sept. 29 increase of $100/ton on HSS, mechanical and piling. Orders already on its books are price-protected for shipment through Oct. 23. The four increases since July (+$60, +$80, +$80, +$100) add up to $320/ton. A bid carrying HSS at a number from before Sept. 29 is at least $100 short.
Freight is the open line. Nucor-Yamato's truck fuel surcharge is 57% for October, and Steel Dynamics' is 53%. Neither mill had posted a November figure when I checked. The raw materials surcharge was last posted at $299/ton on Sept. 15; the October entry had not appeared yet.
What to watch: Labor. Steel Market Update reported on Oct. 7 that the Steelworkers still have no deal with U.S. Steel or Cleveland-Cliffs. On Sept. 30, U.S. Steel put a five-year contract on the table that it said carries a wage increase of nearly 20%; the union called the offer a disappointment, and its own Sept. 30 update lists wages, health care, partnership language and successorship as open. Mills are running a bit softer: AISI put capability utilization at 77.8% for the week ending Oct. 10 (1,796,000 net tons, down from 78.3%, up 4.1% from the same week of 2025; year to date 79.0% versus 77.0%). On demand, the Dodge Momentum Index rose to 288.0 in September, up 2.4% from August's revised reading and 3.7% year over year, and commercial planning excluding data centers is down 14.9% year over year, a narrower gap than August's 18.9%. For precon: carry beams at $1,750 and don't plan on a November drop, price HSS off the Sept. 29 letter and expect a re-quote after Oct. 23, and keep bid validity short until the U.S. Steel and Cliffs talks close. For live data, visit our Market Informer page.
Sources & Further Reading
1. Goldratt, Eliyahu M. Critical Chain. North River Press, 1997. ISBN 0-88427-153-6. (The method itself. The descriptions of safety time, student syndrome, Parkinson's Law, multitasking and buffers in this issue come from the two summaries below, not from the book's text.)
2. Marris Consulting. Critical Chain (book abstract), May 2020. critical-chain-projects.com (Roughly 80% task estimates and stacked management margins; tasks finishing on the planned date and rarely early; student syndrome; Parkinson's Law; the five one-week tasks example; project and feeding buffers; the one-third / two-thirds buffer rule.)
3. Wikipedia. “Critical chain project management.” wikipedia.org (CCPM introduced in Goldratt's 1997 book; tasks planned at 50% best-guess durations; safety pooled into project and feeding buffers; buffer sizing from the gap between “safe” and 50% durations.)
4. Wikipedia. “Critical Chain (novel).” wikipedia.org (Publisher, year and ISBN.)
5. Nucor-Yamato Steel Company. ASTM A992/A572-50 Price List, October 1, 2026 (replaces August 3, 2026). nucoryamato.com (Common wide-flange $87.50/cwt; C15 $88.50; MC12/13/18 $101.25; L8 angles $98.75; S-shapes $112.25. Fetched Oct. 19, 2026.)
6. Nucor-Yamato Steel Company. Raw Material Surcharge and Fuel Surcharge. nucoryamato.com (RMS $14.95/cwt or $299/ton effective Sept. 15, 2026, the latest entry posted; truck fuel surcharge 47% Sept. 1 and 57% Oct. 1. Fetched Oct. 19, 2026.)
7. Steel Dynamics, Structural & Rail Division. Beam Price List, October 1, 2026. steeldynamics.com (Common wide-flange $87.50/cwt; “Pricing increase” revision; October truck FSC 53%. Fetched Oct. 19, 2026.)
8. Nucor Tubular Products. Price Announcement, September 29, 2026. nucortubular.com (“increase pricing on all new orders for HSS, mechanical, and piling products” by US $100.00/ton; orders on the books price protected for shipment through October 23, 2026.)
9. Nucor Tubular Products. Price Announcements index. nucortubular.com (Structural announcements dated July 21, Aug. 7, Sept. 9 and Sept. 29, 2026; none after Sept. 29 as of Oct. 19.)
10. American Iron and Steel Institute weekly data, as reported by ScrapMonster. “U.S. raw steel production report: output up 4.1% YoY.” October 13, 2026. scrapmonster.com (Week ending Oct. 10: 1,796,000 net tons at 77.8%, versus 1,801,000 and 78.3% the prior week and 75.1% a year earlier; YTD 73,763,000 net tons at 79.0%, up 5.3%.)
11. United Steelworkers. “U.S. Steel Bargaining Update – 9/30/2026.” September 30, 2026. usw.org (Wages, health care, partnership language and successorship listed as open issues.)
12. Steel Market Update. Labor coverage, including “USW says no deal yet with U.S. Steel or Cliffs” (Oct. 7, 2026) and “U.S. Steel offers 5-year labor contract to USW, highlights nearly 20% wage hike” (Sept. 30, 2026). steelmarketupdate.com (Headlines and summaries read from the tag page.)
13. Dodge Construction Network. “Dodge Momentum Index Grows 2% in September.” October 7, 2026. construction.com (DMI 288.0, up 2.4% from August's revised 281.3 and up 3.7% year over year; commercial planning excluding data centers down 14.9% year over year.)
14. Dodge Construction Network. “Dodge Momentum Index Flat in August.” September 8, 2026. construction.com (Commercial planning excluding data centers down 18.9% year over year, the comparison point.)
Champion Steel Works — Steel you can trust.
