The 240 Centimetre Wardrobe Was Priced as a 180
· 7 min read · Written by Faceela Research & Editorial Team
Reviewed by Ahmed Hassan Algammal — Founder and Enterprise Systems Consultant
The wardrobe was quoted from a drawing and a memory. The estimator had built one like it before, he remembered roughly what it came to, and he adjusted for the finish. That is not a lazy method. In a workshop that has never once been told what its own jobs actually cost, it is the only method available.
The wardrobe on the drawing was 240 centimetres to the ceiling. The one in the memory was 180. Everything above 180 is another shelf, another metre of edging, another pair of hinges, another partial sheet, and a second man for the lift on site.
None of that is a mistake in arithmetic. It is a mistake in what was being counted.
That is the whole argument here: an estimate that is never compared with the finished job is a guess that gets more confident each time it is used. An estimate that is compared, job after job, becomes a rate — and the comparison is worth more than the estimate ever was.
Priced by the unit, built by the metre
Almost every joinery quotation in this country is written per unit or per square metre of elevation, because that is what the client asked for and what the tender comparison needs.
The workshop does not consume anything per unit. It consumes board by the sheet, edging by the metre, hardware by the hinge and the runner and the lift-up, finish by the litre and the spray booth hour, and labour by the man-hour at the machine, on the bench and on the site.
So the price and the cost are denominated in different units, and the conversion between them lives in one person's head. It is usually a good head. It is also the only copy, it is not written down anywhere, and it does not update itself when the board price moves or when the workshop starts using a different runner.
The consequence is not that estimates are wrong. It is that nobody can say by how much, in which direction, or on which kind of job.
The loop
Three legs exist in every workshop. The fourth is what makes them worth anything
1. The estimate
Built from a drawing and a memory. Per unit, mostly, because that is how the client asked — while the cost is per metre, per sheet, per hinge and per hour.
What comes out: A price, and a build-up nobody opens again
In most workshops: Built
2. The job
Boards issued, edge banding, hardware, the finish, the hours on the machines, the subcontract that went out for powder coat, the truck, the install crew, the two men who went back with a van.
What comes out: Material issued, hours booked, subcontract invoices, delivery notes
In most workshops: Built
3. The actual cost
Closed on the job, usually late, usually by finance, usually as one number. It answers whether this job made money, which is the smaller of the two questions it could answer.
What comes out: A margin, for one job that is already finished
In most workshops: Built
4. The return leg
The actual rate per metre, per sheet, per hinge and per hour goes back and becomes what the next quotation is priced from. Not the whole job — the rates inside it.
What comes out: A rate the estimator did not have to remember
In most workshops: Not built
Feeds: The estimate
Without the fourth leg it is a line, not a loop
The estimate is never compared to the actual, so the comparison never becomes a rate, so the next estimate is priced from the same memory that priced the last one. An estimating system that is not fed actuals is a calculator with a logo on it.
The one word that changes everything
The specification says veneer. It does not say from how many batches, or whether the run has to match across a wall.
It says handleless. It does not say whether that is a J-profile routed into the door, a gola channel that has to be extruded and delivered, or a push-to-open mechanism that adds a piece of hardware to every door in the job.
It says duco. It does not say whether that happens in the booth on a Tuesday or on site after the marble is in, with masking, with a second visit, and with an argument about who protects it afterwards.
Every one of those is a single word in the specification and a different cost structure in the workshop. An estimator working from a drawing reads them at reading speed, because they read at reading speed. The workshop discovers what they meant eleven weeks later.
Two jobs that look identical and are not
Two clients order the same wardrobe. The drawings are near enough the same drawing.
One is a villa with a store room, a driveway and a client who signs. The other is a fitted apartment on the eleventh floor of an occupied tower, with a two-hour lift booking and no storage anywhere on site, and a designer who has views about grain direction.
The board cost is identical. Nothing else is. And the estimate that priced the second one from the memory of the first has already lost the difference before a sheet was cut, which is why what the site does to the installation cost is not a separate subject from estimating — it is a column the estimate did not have.
What actually drives the cost, line by line
The table below is the translation nobody writes down. It is short on purpose: these six lines cover most of what separates a quoted job from a delivered one.
| The line in the specification | The driver it really is | What a per-unit price misses |
|---|---|---|
| Height to ceiling rather than standard | Sheets, edging metres, shelf count | The top box is a second unit wearing one number |
| Veneer rather than laminate | Batch, matching, spare, sequence | Waste rises and the offcut cannot be used elsewhere |
| Handleless | Profile, channel or mechanism, per door | Hardware count multiplies by the number of doors |
| Soft-close throughout | A named hardware item per hinge and runner | It is a specification line, not a finish preference |
| Duco applied on site | Visits, masking, protection, rework | Booth cost was priced; site work was not priced at all |
| Curved or angled fronts | Machine time, jigs, scrap rate | Per-metre pricing assumes straight cuts |
The last column is where the money is. Each of those rows is something the estimator knew and could not price, because a per-unit rate has nowhere to put it.
The return leg
Here is the part that changes the business, and it is not the estimating.
When the job is finished, somebody has to put the actual consumed board, the actual hardware, the actual finish and the actual hours next to the estimate for the same job, at the same level of detail, and look at the difference. Not the total margin — the totals hide compensating errors and always have. Line by line, in the same categories the estimate was built in.
Do that for ten jobs and the workshop stops estimating from memory. It starts estimating from its own history, which is the only data about this workshop that is true anywhere.
Skip it, and the estimate keeps being a memory, and the memory keeps being of the jobs that went well, because those are the ones people remember. That is why estimates drift optimistic in a direction nobody chose.
An estimating system that is never fed actuals is a calculator with a logo on it.
What to make a vendor show you
On a live system, on a real job, not on slides.
- Build an estimate from a bill of materials and show the sheets, metres, pieces and hours it is made of, not just the total.
- Change one specification line — laminate to veneer — and show every cost line that moves as a result.
- Show the estimate carrying the hardware count per door rather than a lump for hardware.
- Take a finished job and produce estimated against actual, in the same categories, on one screen.
- Show the variance at line level and name which line lost the money.
- Take the actual from that job and use it to price the next quotation, and show where the new rate came from.
- Produce the same estimated-against-actual comparison across every job of one type, for the last year.
Item 4 is the one that decides it. Every quotation tool can build a number. A system that can put the delivered job next to the number that sold it, in the categories the estimator used, is the only one capable of ever making the next estimate better — and item 6 is the proof that the loop is actually closed rather than merely reported on.
Item 7 is what turns it from a post-mortem into a pricing policy.
The short version
Joinery is quoted per unit and built per sheet, per metre, per hinge and per hour, and the conversion between the two lives in one estimator's head where nobody can audit it or inherit it.
A single word in a specification — veneer, handleless, duco, curved — changes the cost structure rather than the cost, and a per-unit rate has nowhere to record that.
Two jobs that look like the same drawing are not the same job once the address is known, so the site conditions belong in the estimate and not in a surprise at installation.
And the estimate is not the deliverable. The comparison is. A workshop that puts actual next to estimated on every finished job, line by line, is pricing from its own history within a year; one that does not is pricing from a memory of the jobs that went well.
The system that carries the estimate through to the actual — one bill of materials, one job, one comparison — is Flow, built for joinery, steel and fabrication. Nearby: what the approved drawing commits you to, the shortage created eight weeks before it was found, what the sheets do on the machine, and job costing for fabrication workshops. If the numbers on your screen already disagree with the numbers in the workshop, start with why your dashboard is lying to you.