Case Study 2 — Twenty-Two Percent Under: Three Weeks of Lost Production at Cottonwood Creek

Every person, company, and project in this book is a Tier-3 illustrative composite. The numbers are internally consistent and realistic; they are not a real project.

Setup

The project. The Cottonwood Creek Bridge Replacement — $18,700,000 of heavy civil for a state department of transportation, a unit-price contract, 210 working days. Superintendent Del Ferraro; project engineer Ingrid Sørensen.

The pay item in trouble. Roadway excavation and embankment for the two bridge approaches: 86,000 BCY of cut, hauled 2.6 miles along an on-site haul road to the embankment fill. The material is a granular sandy soil with a 22 percent swell factor, so the trucks carry 105,000 LCY.

The unit-price fact that makes this hurt. On a unit-price contract, the DOT pays for cubic yards measured, not for hours worked. Quantity risk sits with the owner; productivity risk sits entirely with the contractor. Every hour of lost production is a dollar the contractor never gets back, and there is no change order to write.

The plan, as bid.

The planned cycle:

Trucks required = 18.3 ÷ 2.4 = 7.6 → 8 trucks

Planned fleet rate = $172 + (8 × $88 = $704) + $305 = $1,181/hr

Planned unit cost = $1,181 ÷ 300 = $3.94 per LCY

Planned duration = 105,000 ÷ 300 = 350 hr ÷ 9 = 38.9 → 39 work days

What Happens

Three weeks of "we're a little behind"

Fifteen work days into the operation, Ingrid Sørensen runs the quantities against the schedule for the monthly report and brings Del a number he does not like.

Ingrid: "We should have moved 40,500 loose yards by now. We've moved 31,604. We're at 234 an hour against 300. That's twenty-two percent under."

Del: "The machine's running fine. Nobody's broken down. We haven't lost a day to weather."

That sentence is the entire case study. The machine was fine. The excavator was loading trucks at exactly the rate it was bid to load them. Everything that went wrong happened somewhere else in the cycle, which is why three weeks went by with nobody able to name it.

Del had been managing the operation the way most superintendents do: by looking at the loading tool. The loading tool looked good.

The stopwatch

Del spends two mornings at the cut and two afternoons at the fill with a stopwatch and a clipboard, timing forty complete truck cycles element by element. Here is what he brings back.

Three separate small failures, none of which anybody would have escalated on its own:

1. The fill was choking. The embankment section had narrowed as it rose, and one dozer was spreading and one compactor was working a single lift with a single access ramp. Trucks were backing down that ramp one at a time and waiting for the dozer to clear the previous load. Dump and maneuver went from 2.0 minutes to 5.0.
2. The haul road had degraded. A 2,400-foot stretch had rutted after the spring rain. Average loaded speed over the whole route dropped from 22 mph to about 18, and empty return from 26 to 22.
3. The cut face had been worked back to a point where trucks could only approach single-file. Spotting went from 0.8 minutes to 1.4.

The fleet-matching error

Here is the part Del had not done, and it is the error that made three small problems into a twenty-two percent shortfall.

Trucks required = cycle time ÷ load time = 24.6 ÷ 2.4 = 10.25 → 11 trucks

He still had eight. The fleet had been matched to an 18.3-minute cycle that no longer existed, and nobody had recomputed it since day one.

With eight trucks on a 24.6-minute cycle, the trucks govern, not the excavator:

Fleet production = (8 trucks × 12 LCY) ÷ 24.6 min × 60 = 5,760 ÷ 24.6 = 234 LCY/hr

234 ÷ 300 = 78 percent of plan — 22 percent under

The excavator, meanwhile, was capable of 300 LCY/hr the entire time and was simply standing still for a quarter of every hour with nothing under the bucket. The machine was fine. The machine was idle.

What three weeks cost

Projected over the full 105,000 LCY if nothing changes:

105,000 × $1.11 = $116,550 of overrun

Duration = 105,000 ÷ 234 = 448.5 hr ÷ 9 = 49.8 → 50 work days instead of 39 — eleven work days late

On a unit-price contract, all $116,550 of that is the contractor's. There is no entitlement, no change order, and no argument to be had. The DOT pays $/CY and the cubic yards are the cubic yards.

Two ways to fix it

Del prices both, and the comparison is more interesting than it looks.

Fix A — throw trucks at it. Add three trucks for a total of eleven, matched to the broken 24.6-minute cycle. Production returns to the excavator-governed 300 LCY/hr.

New fleet rate = $172 + (11 × $88 = $968) + $305 = $1,445/hr

New unit cost = $1,445 ÷ 300 = $4.82/LCY — still $0.88 over the bid

Remaining 73,396 LCY ÷ 300 = 244.7 hr × $1,445 = $353,590

Fix B — fix the cycle and keep eight trucks.

Corrected cycle:

Trucks required = 18.8 ÷ 2.4 = 7.83 → 8 trucks. He already has eight.

New fleet rate = $1,181 + $170 = $1,351/hr

New unit cost = $1,351 ÷ 300 = $4.50/LCY

Remaining 73,396 LCY ÷ 300 = 244.7 hr × $1,351 = $330,590 + $18,600 one-time = $349,190

And doing nothing: 73,396 ÷ 234 = 313.7 hr × $1,181 = $370,479, plus eleven work days.

Fix B wins by $4,400 on this pay item alone — and by considerably more once you count the four remaining months of aggregate base, riprap, and structure hauling that will use the same haul road. It also removes a real hazard: a rutted grade is where loaded trucks roll over, and the DOT's erosion-control inspector had already written one deficiency for material tracked onto the state route from that stretch.

Del does Fix B and releases nothing — he keeps eight trucks, which is what the corrected cycle needs, and puts one on standby availability so a flat tire does not immediately drop him to seven.

Net recovery: $21,289 against doing nothing on this item, roughly 7.7 work days, and an unquantified but real reduction in the chance that somebody puts a loaded tandem on its side.

Analysis

1. The failure was invisible because it was distributed. No single element of the cycle grew enough to trigger anybody's alarm. Half a minute of spotting, a minute and a half of haul, three minutes at the fill. Each one, on its own, is a shrug. Together they are 34 percent of the cycle and 22 percent of the production. The only way to see a distributed failure is to measure the elements separately, which is what a stopwatch study is for and why it takes two mornings, not five minutes.

2. Del was watching the wrong machine. He watched the loading tool because the loading tool is the machine that looks like it is doing the work. It was performing exactly as bid. The failure was in the system the loading tool feeds, and the system's binding constraint had migrated from the excavator to the trucks without anybody noticing the handoff. When the constraint moves and nobody re-measures, you optimize the wrong thing for weeks.

3. Fleet matching is not a one-time calculation. The formula — trucks required equals cycle time divided by load time — was applied correctly on day one and never applied again. Any change to haul distance, haul road condition, dump geometry, or cut access changes the cycle, and any change to the cycle changes the required fleet. Recompute it every time something changes, and put "cycle-time check" on the weekly superintendent's routine.

4. Adding trucks is a treatment, not a cure — and sometimes it is the right treatment anyway. Fix A restores production immediately and costs less to start. It also locks in $0.88 per yard of permanent overrun and leaves a bad haul road in place for the rest of the job. Fix B costs more up front and returns the operation to something close to the bid. On a short remaining quantity, Fix A can genuinely be right; on a long one, it almost never is. Do the arithmetic instead of following a rule.

5. Unit price concentrates productivity risk. Chapter 4 taught that a unit-price contract puts quantity risk on the owner and productivity and price risk on the contractor. This is what that sentence means on a Tuesday: 105,000 loose yards is 105,000 loose yards no matter how long it takes you, and the DOT's check does not change. On a cost-plus job the owner would have absorbed most of this; on a lump-sum job it would look the same as it does here. Know which risk your contract gave you, and put your measurement effort there.

6. The safety finding was free and nobody was looking for it. The rutted haul road was a production problem, an erosion-control compliance problem, and a rollover hazard at the same time. That is not a coincidence. Degraded production conditions and degraded safety conditions are usually the same conditions, which is theme 4 stated as a diagnostic rather than a slogan: when production drops for no obvious reason, walk the operation, because whatever you find will usually also be a hazard.

Discussion Questions

1. Ingrid caught this in the monthly quantity reconciliation, on work day 15. Design a reporting routine that would have caught it on work day 3. Be specific: what gets measured, by whom, how often, and what number triggers a stopwatch study?

2. Rework Fix A and Fix B assuming only 18,000 LCY remained instead of 73,396. Which is cheaper, and what does that tell you about when "throw trucks at it" is the professionally correct answer?

3. Del kept eight trucks after the fix rather than releasing any, and put one truck on standby availability. Using the 92-percent-availability argument from §21.4.3 of the chapter, evaluate whether eight is actually enough for a corrected 18.8-minute cycle.

4. The corrected dump-and-maneuver time is 2.4 minutes, not the original 2.0. Del deliberately did not plan on getting back to 2.0. Why is that the right call, and what would it have cost him to be wrong?

5. Suppose this had been Northgate — a GMP contract with a construction contingency — instead of a unit-price DOT job. Would the $116,550 have been recoverable? Answer carefully, and name the specific document you would go looking for first.

Your Turn

You are running an earthmoving operation with the following measured cycle. The haul is 4.6 miles each way, loaded at an average 22 mph and empty at an average 28 mph.

Your excavator produces 265 LCY/hr. Trucks hold 14 LCY. You have eight trucks. The excavator plus operator costs $196/hr, each truck costs $92/hr, and support costs $280/hr. There are 62,000 LCY left to move at 9 productive hours per day.

Compute:

(a) The cycle time and the required fleet. (b) Your current production rate, and how far under the excavator's capability you are running. (c) The cost per LCY as you are running today, and the cost per LCY with a matched fleet. (d) The change in total cost and in work days from matching the fleet on the remaining quantity. State whether matching is a saving or a cost, and if it costs money, say what makes it worth doing anyway. (e) Then the judgment question, which is the one that matters: your trucking broker can give you three more trucks starting Thursday, or you can shorten the haul by 1.4 miles each way by opening a temporary crossing that costs $26,000 and takes four days to build. Work both, choose one, and name the condition that would flip your answer.

Write the numbers down before you go looking for a formula. If your first instinct was to check whether a bigger excavator would help, reread paragraph two of the Analysis.