You have been dividing by your own profit
Two weeks ago the reinforcement package had a budget of $289,388, an actual cost of $189,936, and we called the CPI 0.87.
Look at those two numbers again.
The $289,388 is the bill rate — $1,446.94 a tonne, with twelve percent of preliminaries and fifteen percent of overhead and profit baked into it. It is a price. The $189,936 is invoices.
You cannot divide one by the other. And we did.
Week 3 told you to choose a convention and never let both happen. Week 1 told you that actual cost and earned value have to be measured on the same basis. Then Week 4 quietly broke both rules, and the number that came out was 0.87, and 0.87 looked reasonable enough that nobody stopped.
That is exactly how it happens on a real project.
The correction
Strip the mark-up out. The direct cost of a tonne of reinforcement is $1,123.40 — material, labour, plant. Two hundred tonnes is $224,680.
That is the number the site can be held to. It is the only number in the whole bill that an invoice is allowed to be compared with.
So: 114 tonnes placed, at $1,123.40, is $128,067.60 of earned value. Against $189,936 of actual cost.
CPI 0.67.
Not 0.87. The package is not losing thirteen cents in the dollar. It is losing thirty-three.
And here is the part that should make you check every cost report you have ever signed: the ratio between the two answers is 1.2880. Divide the bill rate by the direct rate and you get 1.2880. It is the mark-up, exactly. When you measure invoices against a selling price, your CPI is flattered by precisely the profit you have not yet made.
This is the whole of this week in one sentence. An estimate and a budget contain the same money and are not the same document.
The estimate was written to be accepted. It is organised by trade, priced at market, marked up, and its reader is a client. It has done its job the moment the contract is signed, and after that it is a historical artefact.
The budget is written to be run. It is organised by who does the work and where, it carries cost and not price, and its reader is a foreman with a problem on a Tuesday.
Nobody hands you the second one. You have to cut it.
Three structures, one address
A budget is not a list. It is a coordinate system, and it has three axes.
The WBS says where: substructure, superstructure, external works. The CBS says what kind of money: labour, plant, material, subcontract, site overhead. And the OBS says who: the concrete foreman, the buyer, the temporary works engineer.
A cost code is the address where those three meet.
Look at what that code is doing. 03-140-10. Superstructure. Reinforcement. Labour.
It is not a filing reference. It is a question with an answer attached: who is responsible for the labour on the reinforcement in the superstructure, and what were they given to do it with? Twenty-four thousand six hundred and fifty-two dollars and eighty cents, and the concrete foreman's name.
Two hundred tonnes of steel, one line in a bill, becomes six codes. They close back to $224,680 exactly, because a budget that does not reconcile to the thing it came from is not a budget — it is an opinion with a spreadsheet.
“An estimate answers what should this cost. A budget answers who do I ask when it doesn't.”
— WHY THE STRUCTURE IS THE POINT
Same money. One is sorted for pricing. The other is sorted for blame — and blame, properly done, is just ownership.
Why it is worth the two days
Go back to the package. It is $61,868 over. That is the whole story you get from one line called “reinforcement”, and it is a useless story, because there is nothing anybody can do with it on Monday.
Now split it by cost code.
The labour code is $29,030 over. Divide it out and it says one thing: the gang is tying 2.5 tonnes a day. The rate from Week 2 assumed five. Half. That is a conversation with the concrete foreman about access, about congestion, about whether the bar schedules are arriving in time to fix from.
The material code is $28,428 over, and it says two things: the steel was bought at $940 instead of $800, and ten percent of it is being wasted instead of five. That is a conversation with the buyer about escalation, and a conversation with the detailer about why bars are being cut on site.
Two failures. Two owners. Two different Monday mornings. Almost exactly half each, which is why the single blended number told you nothing — it averaged two unrelated disasters into one shrug.
And the plant code, at $4,410, is not a problem at all. It is a symptom: the crane is charged by the tonne placed, so a slow gang makes the plant code overrun automatically. Fix the labour and the plant fixes itself. A budget structured by cost type tells you that. A budget structured by bill item never will.
The rules of a code of accounts
Every code has exactly one owner. Not a department. A person, who can be asked a question and is expected to answer it. If two people own a code, nobody does.
The site must be able to book to it without thinking. If a foreman filling in a timesheet on a wet Friday afternoon cannot tell which code a man belongs to in under three seconds, the code structure is wrong — and you will spend the next two years cleaning up miscoded hours that no report can ever unpick.
Cost and progress use the same code. This is the one everybody gets wrong, and it is why Week 14 exists. If the timesheet books to 03-140-10 and the progress measurement books to “Slab L2 rebar”, you have two systems and you will never divide one by the other.
Never more codes than you can actually measure. A four-hundred-code structure on a job with one QS produces four hundred codes of garbage. Fewer codes, honestly filled, beat a beautiful structure nobody maintains.
Practical insight
Open your budget and find one work package. Ask two questions.
First: does this number include our overhead and profit? If nobody knows, or if the answer is “I think so”, then every CPI on your project is wrong by the mark-up — and it is wrong in the flattering direction, which is the direction nobody investigates.
Second: can I see this package split by labour, plant and material? If you cannot, you do not have a cost control system. You have a cost reporting system, and there is a world of difference: one of them tells you the score, and the other one tells you what to do.
Key takeaways
✔ An estimate and a budget hold the same money and are different documents. The estimate is a price. The budget is a cost.
✔ Dividing invoices by a marked-up bill rate flatters CPI by exactly the mark-up: 1.2880 on this job.
✔ The reinforcement package is not at CPI 0.87. On the right basis it is at 0.67.
✔ A cost code is an address: WBS (where) × element (what) × CBS (what kind of money) × OBS (who).
✔ One bill item became six codes and closed back to $224,680 exactly. If it does not close, it is not a budget.
✔ $61,868 of blended overrun is two separate failures: labour $29,030 (a slow gang) and material $28,428 (a bad buy and bad waste).
✔ One owner per code, bookable in three seconds, and the same code for cost and progress.
✔ Fewer codes honestly filled beat a beautiful structure nobody maintains.
What is coming next
Six codes, one package. Now do it for the whole job and you have four hundred of them.
Nobody manages four hundred anything. So they get grouped — into the boxes where a scope of work, a budget, a schedule and a single human being all line up, and where a variance can finally be handed to somebody who can be expected to fix it.
Next week: control accounts and work packages — the level at which a project is actually managed, and how to choose it without either drowning or going blind.
Enjoyed this lesson?
Join with Google to get each new lesson the moment it's published — and help me see which topics matter most to you. No spam, one email a week, unsubscribe anytime.
Already following on LinkedIn works too — this is just for the weekly email.