Every number in this track rests on one word.

They do not agree. Take one activity — the core wall reinforcement, twenty-three tonnes, budget $25,838 — on a day when fourteen tonnes are fixed in place. Ask each of the six methods how complete it is.

Forty-two percent to ninety-seven percent. Nobody tied another bar between those two numbers. Somebody just chose a different rule.

This is the most important thing a cost engineer measures and the thing most projects are sloppiest about, so let us take the six methods one at a time and sort the facts from the fiction.

The two that are facts

Units completed. Fourteen tonnes of twenty-three. 60.9%. You can count it, you can weigh it, the bar schedule says how many tonnes the finished wall contains, and the delivery notes say how many arrived. There is no opinion anywhere in it. This is the gold standard, and it works for anything you can measure in a unit: tonnes of steel, cubic metres of concrete, square metres of formwork, linear metres of pipe.

Incremental milestone. For work that comes in stages, you assign value to each completed stage and nothing to a stage in progress. Reinforcement fixed, then inspected, then released for concrete. You earn the fixing milestone only when the whole pour is fixed and signed off — which is why this method reads 42.6% while units-completed reads 60.9%. It is deliberately conservative. It refuses to pay you for a stage until the stage is genuinely closed, and on a well-run job that discipline is worth more than the precision it costs you.

These two are the backbone. Nearly everything on a construction project can and should be measured one of these two ways.

The two that are rough but honest

Start/finish. Zero when it has not started, fifty percent once it has, a hundred when it is done. It reads 50% here. It is crude, but for short activities it is fine — if a task starts and finishes inside one reporting period, the 50% is never actually claimed, because by report day it is either 0 or 100. This is exactly why Week 7 insisted on short work packages. Short packages make this crude method honest.

Weighted milestone. The grown-up version of incremental. You split the activity into weighted steps — steel delivered 30%, fixed 50%, inspected 20% — and claim the weights you have earned. Set honestly, it lands at 60.4%, right next to the truth. Set dishonestly, with 60% of the value loaded onto “materials delivered”, it becomes a machine for claiming money the moment steel arrives on a lorry. The method is only as good as the person who set the weights.

The one that is a lie

Supervisor opinion. You ask the foreman how far along the wall is, and he says ninety percent.

It is not ninety percent. It is 60.9%. But the foreman is not lying, exactly — he is an optimist standing next to something that is mostly built, and mostly built feels like ninety. Everybody who has run a site knows the number that comes back. It is always ninety.

Here is what opinion does to a project. It races to ninety and then it stops, because the last ten percent is the snagging, the tie-off, the inspection, the making-good, the part nobody wants to do and nobody remembers to claim. So the activity sits at ninety percent for four months while it is actually finished, and the earned value you booked early has to be given back — except earned value never visibly gets given back. It just quietly stops growing while the cost keeps climbing, and your CPI unwinds over a quarter for reasons nobody can point at.

Opinion is not a measurement method. It is the absence of one, wearing a percentage.

The one that switches your cost report off

And then the dangerous one, because it looks the most sophisticated. Cost ratio, also called level of effort. You take the cost spent to date and divide it by the budget, and call that your percent complete.

Read that again. You are using what you spent to measure what you earned.

The core wall gang is slow — the same 2.5-tonnes-a-day problem from the whole of this track. They have spent $25,164 to fix fourteen tonnes that should have cost $15,728. Measure progress by units and the CPI is 0.62, and it is screaming at you.

Measure progress by cost ratio, and earned value becomes $25,164, because that is what was spent. Divide it by the actual cost of $25,164. CPI 1.00.

It will always be 1.00. It is 1.00 by construction, on every activity, in every month, no matter what happens on site, because you have defined earned value as equal to actual cost. You have built a cost report that is mathematically incapable of reporting a cost problem.

Level of effort has one legitimate home: work that genuinely has no measurable output. Site management. The safety officer. The tower crane, standing there earning its keep by existing. Preliminaries, which Week 8 told you are earned by the calendar. For those, effort is the output, and level of effort is correct.

Use it on anything you could have counted — on the concrete, the steel, the formwork — and you have quietly turned the lights off. A control account full of level-of-effort activities will report a serene 1.00 while the money burns, and by the time the real measure catches up, the quarter is gone.

The choice is the control

Here is the thing nobody tells you: you choose the method before the work starts, activity by activity, and you write it into the baseline. You do not get to pick it at month end when you can already see which answer you would prefer.

That is the whole discipline. If the method is chosen in advance — units for the steel, milestones for the pours, level of effort for the crane — then the number that comes out at month end is a measurement. If it is chosen at month end, it is a negotiation with yourself, and you will lose.

Practical insight

Take your progress report and, for the five biggest activities, ask one question of each: what did we count?

If the answer is a number of tonnes, metres, or units against a known total — good. That is a measurement.

If the answer is “the foreman reckons about eighty”, or “we've spent about eighty percent of the budget”, you have found an activity that is not being measured at all. And if the activity is anywhere near the critical path, that unmeasured percentage is the one that will move twenty points in the wrong direction the month before handover.

Then check one more thing: how many of your activities are sitting at exactly ninety percent. Count them. That number is the size of the cliff you are going to walk off in the last quarter.

Key takeaways

✔ Percent complete is the rawest input in the cost system. There are six ways to produce it and they disagree by 55 points.
✔ Units completed and incremental milestone are facts — counted or weighed. They are the backbone of construction measurement.
✔ Start/finish is crude but honest if packages are short. Weighted milestone is only as good as the weights.
✔ Supervisor opinion always returns ninety, then sits there for four months while earned value quietly stops.
✔ Cost ratio / level of effort defines earned value as cost spent, so CPI is 1.00 by construction — it cannot report a problem.
✔ Level of effort is correct only for work with no measurable output: management, safety, the crane, preliminaries.
✔ Choose the method in advance, per activity, in the baseline — never at month end when you can see the answer you want.
✔ Count the activities stuck at 90%. That is the size of your end-of-job cliff.

What is coming next

Everything so far has been your own work, measured with your own tonnage sheets.

But on most projects the majority of the cost is not yours to measure. It belongs to a subcontractor, who submits his own valuation, marks his own homework, and would very much like to be paid for materials sitting under a tarpaulin.

Next week: subcontractor cost control — how you value work you did not do, priced by someone whose interests are the opposite of yours.